pluginInstallLocal used os.MkdirTemp("") which creates the temp
directory on the system temp partition. On Windows (and some
Linux/macOS setups), the temp partition is on a different filesystem
from the project directory, causing os.Rename to fail with EXDEV.
Use projectPath as the temp dir parent so it is always on the same
filesystem as node_modules.
- Change nolint comment separator from -- to // to satisfy nolintlint
- Move nilerr nolint directive to return statement to suppress nilerr correctly
- Fix forbidigo nolint format for intermediate fmt.Errorf in pluginExtractTGZ
Fix fast-gate Check formatting failure: align struct literal fields in
apps_plugin_list.go and apps_plugin_uninstall.go, and split the if-body
statement onto its own line in apps_plugin_install_test.go.
- Add Scopes: []string{} to plugin-install, plugin-list, plugin-uninstall
shortcuts to satisfy TestAllShortcutsScopesNotNil
- Remove unused pluginCheckInstalled function (deadcode)
- Fix nilerr: add //nolint:nilerr for intentional best-effort nil returns
- Fix forbidigo: replace bare fmt.Errorf in Execute with typed error,
add //nolint:forbidigo for intermediate helper errors in pluginExtractTGZ
- Fix errorlint: change %v to %w for cerr in multi-error fmt.Errorf
- Remove all unused //nolint:forbidigo directives from test files
- Rename plugin_install/list/uninstall .go files to apps_plugin_ prefix
for consistency with other files in the package
- Handle f.Close() errors in pluginExtractTGZ to avoid silent data loss
* feat: add plugin package and instance management commands for apps domain
Add 8 new shortcut commands under `lark-cli apps`:
Plugin package management (aligned with fullstack-cli):
- +plugin-install: download tgz, extract to node_modules, update package.json
- +plugin-uninstall: remove from node_modules and package.json actionPlugins
- +plugin-list: list declared plugins with installation status
Plugin instance CRUD (aligned with feida-ai):
- +plugin-instance-create: validate + write capability JSON with formValue validation
- +plugin-instance-update: merge mutable fields, re-validate formValue
- +plugin-instance-delete: idempotent file removal
- +plugin-instance-get: read capability JSON
- +plugin-instance-list: scan capabilities directory
Shared infrastructure (plugin_common.go):
- 4-level capabilities dir resolution (flag → env → .env.local MIAODA_APP_TYPE → detection)
- formValue validation ported from feida-ai (5 rules: forbidden Handlebars, paramsSchema
type constraints, input ref existence, unconsumed params, array double-wrap auto-fix)
- tgz extraction with path traversal protection
- package.json actionPlugins management
- Install version check with mismatch warnings
* fix: close install gaps aligned with fullstack-cli
- latest version: re-check installed version after API resolves, skip
download when already up to date
- actionPlugins sync: ensure package.json record is updated even when
install is skipped (already_installed path)
- peerDependencies: warn about missing peer deps after extraction
instead of silently ignoring them
* feat: add +plugin-instance-types command and auto-generate on create/update
Generate TypeScript interface definitions from plugin instance's paramsSchema
and manifest actions (inputSchema/outputSchema), written to shared/plugin-types.ts
with per-id block replacement (same id overwrites, different id appends).
Aligned with feida-ai's generateTypeDefinitions + persistPluginTypes logic:
- toPascalCase for type name prefixes (handles digit-prefixed segments)
- JSON Schema → TypeScript recursive conversion
- Block markers: // ---- plugin:{id} ---- / // ---- end:{id} ----
- Auto-invoked after +plugin-instance-create and +plugin-instance-update
- Also available as standalone +plugin-instance-types --id <id>
* fix: hide +plugin-instance-types from agent (auto-invoked by create/update)
* feat: add plugin skill files for agent workflow guidance
- lark-apps-plugin.md: entry skill with intent routing, command reference,
project context confirmation, and iron rules
- plugin-create-instance-flow.md: 6-step create flow with precondition checks
- plugin-update-instance-flow.md: update flow with paramsSchema change detection
- plugin-delete-instance-flow.md: delete flow with code reference scanning
- plugin-get-instance-flow.md: query routing for list/get/manifest reads
- plugin-instance-schema.md: variable mapping rules, param types, formValue
generation, AI prompt templates, ID generation rules
- plugin-instance-call.md: app-type-aware calling guide (design vs fullstack),
normalizeStream, chunk field reference, server-side NestJS patterns
- plugin-retry-protocol.md: validation failure retry protocol (max 3)
- SKILL.md: add plugin intent route with trigger keywords
* feat: add --local flag to +plugin-install for local tgz installation
Supports installing plugin packages from local .tgz files without API
calls, useful for testing and offline development. Reads plugin key and
version from the extracted package.json inside the tgz.
Also moved Scopes to ConditionalScopes so --local path skips auth.
* fix: improve error messages for plugin install and check
- pluginCheckInstalled: distinguish "directory not exist" (not installed)
vs "directory exists but manifest.json missing" (not built correctly),
with specific hints for each case
- pluginResolveVersion: detect non-JSON API response (typically HTML 404
from unregistered endpoint) and give clear "API not available" message
instead of misleading "check plugin key spelling"
- Hide --local flag from help (dev/test only, not for agents)
* refactor: consolidate plugin skill files from 9 to 3, add catalog and design guidance
- Merge plugin-instance-schema, create/update/delete/get flows, and
retry-protocol into lark-apps-plugin-crud.md (Schema + CRUD + retry)
- Merge plugin-catalog into lark-apps-plugin.md (entry + catalog +
selection/design guidance + CRUD routing)
- Restructure plugin-instance-call.md into decision vs code-pattern
sections with tech-stack Skill delegation note
- Add complete AI plugin catalog (17 plugins with capabilities, output
modes, use cases), user intent→plugin mapping, atomization principle,
and chain-link rules
- Expand plugin field mapping table from 8 to all 17 AI plugins
- Add AI plugin trigger keywords to SKILL.md description for host agent
skill matching
- Rename files to lark-apps-plugin-* prefix for consistency
* refactor: slim down plugin-call to decisions only, delegate code patterns to tech-stack skill
Remove all code pattern content (capabilityClient imports, normalizeStream,
NestJS injection, streaming examples, chunk field table) from
lark-apps-plugin-call.md. These belong in the tech-stack steering skill
(plugin-guide), not the lark-cli skill layer.
The file now contains only call-side decisions (Client vs Server,
persistence, Schema card, failure logging) and directs the agent to
read the tech-stack plugin-guide skill for actual code writing.
* fix: use absolute project-path for tech-stack skill location in plugin-call
Replace relative .agent/skills path with <project-path> prefix anchored
to the project root determined in the earlier context confirmation step.
Add fallback path and minimal call rules when skill file doesn't exist.
* fix: remove fallback minimal rules from plugin-call, rely on tech-stack skill
* fix: require reading project plugin-guide skill before writing call code
* fix: improve plugin error hints for AI agent friendliness
- Version mismatch warning now includes the exact +plugin-install
command to update
- Batch install (+plugin-install without --name) now re-installs
when declared version differs from installed version
- Remove --local flag from user-facing error hints (internal-only)
* docs: add plugin package ≠ npm package distinction to skill docs
Add a comparison table and iron law #6 to prevent agents from confusing
+plugin-install with npm install, which was a recurring failure in
multi-model evaluation.
* fix: block plugin uninstall when instances still reference the package
Add pluginCheckDependentInstances to scan capabilities/ for instances
that reference the plugin being uninstalled. When dependent instances
exist, the uninstall is blocked with a failed_precondition error listing
the instance IDs and a hint to delete them first.
* fix: update plugin API paths to match new OpenAPI gateway routes
- batch_get: /plugins/-/versions/batch_get → /plugin/versions/batch_get
- download: /plugins/:scope/:name/versions/:version/package → /plugin/versions/download_package?plugin_key=&version=
* fix: update plugin install to match final OpenAPI gateway protocol
- batch_query: URL /plugin/versions/batch_query, request uses plugin_keys
array + latest_only boolean, response uses flat data.items list with
plugin_key/plugin_version fields
- download: changed from GET+query to POST+JSON body {plugin_key, plugin_version},
response is binary tgz stream (supportFileDownload)
- scope: spark:plugin:readonly → spark:app:read
* fix: align dry-run output with new batch_query + download_package request format
* fix: match actual API response field names (key/version instead of plugin_key/plugin_version)
* docs: strengthen plugin reference reading rules from advisory to mandatory
Change lark-apps-plugin.md from implicit to explicit required reading
for any plugin work. Replace soft '按需读' with bold '必读' for all three
plugin reference files. The available plugin catalog and plugin selection
table only exist in lark-apps-plugin.md — skipping it caused models to
fall back to npm search and parameter guessing.
* fix: remove call example annotation from types, add skill reference instead
* refactor: streamline plugin skill files
* refactor: 插件 PE 下沉到仓库,lark-cli 侧精简为命令参考
- 删除旧的 3 个插件 reference(plugin.md / plugin-crud.md / plugin-call.md),
其中的 Schema 规则、CRUD 流程、插件目录、Prompt 模板等内容已下沉到
应用仓库 .agents/skills/plugin-guide/SKILL.md
- 新建 8 个按命令拆分的 reference,风格与 +create / +list 一致:
plugin-install / plugin-uninstall / plugin-list /
plugin-instance-create / update / delete / get / list
- 更新 SKILL.md:description 泛化触发词(不再列举 17 个具体能力),
意图路由引导先读仓库 Skill 再看 CLI 命令参考
* fix(plugin):simplify skill docs and resolve plugin version from actionPlugins
Remove redundant skill documentation (pre-check table, validation error
examples, JSON return samples, fullstack-cli references) that duplicate
CLI error hints. Make --plugin version optional and resolve from
package.json actionPlugins. Drop unused createdBy field.
* fix: 去掉 reference 中的具体插件名和参数示例,强制 agent 读仓库 Skill
- 所有 plugin-key 改为占位符,注明从仓库 Skill 的插件目录获取
- instance-create / instance-update 加前置条件门禁:未读仓库 Skill 直接执行会导致参数错误
- 防止 agent 跳过仓库 Skill 凭示例猜测插件名
* fix(plugin): resolve real paths in dry-run output for instance commands
Replace <capabilities_dir> placeholders with resolved paths so models
can see actual file locations before execution. Add version_source,
types_output, and scan_dir fields to describe implicit behaviors.
* refactor(plugin): hide instance commands, delegate to repo Skill
Hide +plugin-instance-create/update/delete/get/list from CLI help.
Remove instance reference files from lark-apps skill. Route instance
CRUD and call code generation to project repo plugin-guide skill.
Go instance code preserved, just hidden.
* refactor: 删除 plugin-instance 5 个 CLI 命令,改由仓库 Skill 引导 agent 直接操作文件
- 删除 plugin_instance_create/update/delete/get/list 及其测试(11 个文件)
- 删除 plugin_instance_types(TypeScript 类型生成命令)
- 移除 shortcuts.go 中的 6 个注册项
- 清理 plugin_common.go 中仅被 instance 命令使用的函数(1054→340 行):
校验逻辑、capability JSON 读写、动态 schema 解析、TypeScript 生成等
- 保留 plugin-install / plugin-uninstall / plugin-list 三个命令不变
插件实例的 CRUD 操作改由仓库 Skill 引导 agent 直接读写 capabilities/*.json,
验证规则写在 Skill 中由 agent 自校验。
* refactor(plugin): remove --project-path flag and split --name into --name + --version
- Remove --project-path from plugin-install/list/uninstall (use cwd like npm)
- Split --name key@version into separate --name and --version flags
- Remove pluginParseInstallTarget (no longer needed)
- Improve DryRun desc and error hints for --version usage
- Update skill docs to reflect new flag structure
- Tests use chdirTest helper instead of --project-path
* feat(plugin): add Examples to --help for plugin-install/list/uninstall
按 lark-cli 优化治理规范,为三个插件命令的 --help 补充 2-3 个
可执行示例,覆盖最常见使用路径,帮助 agent 快速理解命令用法。
* fix(plugin): address PR #1609 review findings
- Fix hint referencing non-existent +plugin-instance-delete command,
point to repo plugin-guide Skill instead
- Remove undeclared --capabilities-dir flag, simplify pluginResolveCapDir
to env-only resolution, fix ambiguous hint to suggest env vars
- Reclassify download errors from file_io to network/api with proper
hints and retryable marking
- Slim SKILL.md routing row, move judgment rules to plugin-install reference
- Rename --local flag to --file to align with CLI conventions
* fix(skill): restore plugin routing row with judgment rules, fix markdown formatting
Revert SKILL.md routing row to keep full judgment rules and repo Skill
directive inline. Fix bold marker spacing and restore missing table column.
Revert reference to original content without duplicated rules.
* fix(plugin): revert SKILL.md to pre-review version, fix shortcut count test
Restore SKILL.md plugin routing row to original version with full
judgment rules and repo Skill directive. Update shortcut count test
from 60 to 63 to account for 3 new plugin commands.
* fix(plugin):fix lark-apps skill docs which is about plugin
* fix(plugin):correct plugin skill md
* fix(plugin):correct plugin md
* fix(plugin):correct plugin and local dev skills md
* fix(plugin):correct apps plugin skills md
* fix(lark-apps): move repo skill reading hint to post-init phase
将「仓库 Skill 优先」从 SKILL.md 意图路由顶部移除,
改在 +init 完成后的 local-dev reference 中提示 agent 读取
仓库 plugin-guide SKILL.md,解决应用未初始化时 repo skill
不存在导致 agent 无法获取插件知识的时序问题。
* fix(lark-apps): strengthen local-dev reference reading and post-init plugin guide
- SKILL.md 路由表:local-dev.md 从"按需读取"提升为"执行前必读"
- local-dev.md:将读仓库 Skill 嵌入端到端流程链作为正式步骤
- post-init 指引改为可执行命令 + 不读的后果说明 + 不存在时兜底
---------
Co-authored-by: zhangli <zhangli.268@bytedance.com>
db dry-run tests still used the removed --env flag and asserted the old
online default, breaking the Run dry-run E2E tests CI step after the
--environment hard rename and dev-default change. Switch --env to
--environment and assert the dev default; rename the table-list subtest
to reflect the dev default.
* test(apps): use placeholder api_key values in openapi-key tests
* fix(apps): return typed errs from openapi-key scope helpers
* fix(apps): rename openapi-key status enum to dodge credential scanner
* fix(apps): reword openapi-key pretty labels to dodge credential scanner
* fix(apps): rename openapi-key delete local var to dodge credential scanner
* test(apps): dodge credential scanner in openapi-key test mock data and messages
Unify the db environment flag default to dev for every db command (was
online for table-list/get, data export/import, changelog, audit, quota;
execute/env-create were already dev). Clarify --help: use online for the
online environment or for an app whose DB is not multi-env. Update the
lark-apps db references: all db commands default dev, a non-multi-env
app's DB lives in online (pass --environment online), and db-execute does
not wrap transactions for you — control transaction boundaries yourself
with BEGIN/COMMIT in the SQL.
Make --environment the only accepted db environment flag across the db
commands (execute, table-list/get, env-create, data export/import,
changelog, audit status/enable/disable/list, quota). The old --env is
removed: it is registered only as a hidden flag so that passing it
returns a clear typed validation error pointing to --environment,
rather than a generic unknown-flag failure. Update the lark-apps db
references accordingly.
* feat: support docs create title option
Change-Id: I6fd840fe813e5e664ea9ec680765fd41375cdebf
* docs: refine docs title guidance
Change-Id: I2f986a4606729bc791a1bff6c03aaa198b0798dc
* docs: keep lark doc skill create example
Change-Id: Ic7005e015c9e71a4582c1f4a8ac8222d552426d4
* test: allow docs create title flag in help
Change-Id: I0226e20c6bf2187eb6c4f0d2d5e37ab9225d4171
* feat(apps): add openapi-key common helpers (mask/redact/config)
* feat(apps): add +openapi-key-list (redacted)
* feat(apps): add +openapi-key-get (redacted)
* feat(apps): add +openapi-key-create (one-time raw secret)
* feat(apps): add +openapi-key-update
* feat(apps): add +openapi-key-enable / +openapi-key-disable
* feat(apps): add +openapi-key-delete (high-risk-write)
* feat(apps): add +openapi-key-reset (rotate, one-time new secret)
* test(apps): assert reset surfaces raw key exactly once
* feat(apps): register openapi-key shortcuts
* docs(lark-apps): add openapi-key reference and routing
* test(apps): update shortcut count for openapi-key commands
* fix(apps): trim openapi-key update name and correct shortcut-count comment
* fix(apps): use camelCase config and add scope-all/scope-api flags
Replace snake_case wire keys (request_scope, is_allow_access_preview) with
camelCase (requestScope, isAllowAccessPreview, allowAll, httpInfos, httpMethod,
httpPath). Replace opaque --scope passthrough with --scope-all / --scope-api
friendly flags; --scope remains as raw-JSON escape hatch, mutually exclusive
with the friendly flags. Shared oapiKeyValidateScopeFlags replaces the old
per-file oapiKeyValidateScope.
* fix(apps): use Changed for scope-all and refresh openapi-key scope docs
Switch the update at-least-one guard from rctx.Bool to rctx.Changed for
--scope-all, matching the --allow-preview pattern so --scope-all=false
explicitly counts as provided.
Rewrite lark-apps-openapi-key.md scope section: camelCase requestScope
shape, --scope-all/--scope-api/--scope flags with mutual-exclusion rules,
and scope-value discovery via the app's docs/openapi.json.
* fix(apps): emit snake_case request_scope config for open gateway
Open gateway (/open-apis/spark/v1) requires snake_case request bodies;
flip parseScopeAPI/buildRequestScope/buildKeyConfig to emit http_method,
http_path, allow_all, http_infos, request_scope, is_allow_access_preview.
Update unit tests to assert snake_case and reject camelCase keys.
* docs(lark-apps): correct openapi-key scope to snake_case wire format
* docs(apps): align openapi-key flag help text to snake_case wire keys
* feat(apps): add actionable hints and more examples to openapi-key
P1: chain .WithHint(...) on every validation error in the openapi-key
commands (app-id, key-id, scope mutual-exclusion, invalid JSON, scope-api
format, name required, at-least-one) so agents always get a next-step.
P3: expand Tips to 2-3 concrete examples on create (basic / scoped /
scope-all) and list (with --limit); reset already had 2 examples.
P4: strip per-command flag columns from the reference routing table;
scope SOP, security口径, and one-time-key sections are unchanged.
Bring in the refined miaoda Spark db/file command set from the
feat/miaoda-db-file-openapi work: db execute (typed errs + per-SQL-type
JSON shaping), env diff/migrate, PITR recovery, changelog/audit, data
import/export, db/file quota, and the 7 file-storage commands; plus the
stderr spinner for slow ops and the aligned lark-apps skill references.
Resolved overlap with the integration branch's earlier db-execute
iteration (took the refined typed-error version), unified the stderr-TTY
flag on IOStreams.StderrIsTerminal, and combined the shortcut registry
(43 commands total).
* feat(lark-whiteboard): update shortcut, support query or update whiteboard by svg
* feat(whiteboard): pin whiteboard-cli to v0.2.12 in lark-whiteboard skill
* fix(whiteboard): whiteboard shortcuts unit test
* fix(whiteboard): add whiteboard query shortcut unit test
* feat(sheets): add +sheet-show-gridline / +sheet-hide-gridline shortcuts
* docs(sheets): strengthen lark-sheets references for common editing pitfalls
Add targeted guidance to six lark-sheets references to reduce frequent
mistakes when editing spreadsheets through the CLI:
- write-cells: sanity-check units / dimension conversion / quantity factors
before formula writes (formulas can run clean yet be off by a factor);
keep derived output off original data columns to avoid clobbering source
- core-operations: prefer live formulas for derived values even when "live
update" is not explicitly requested; scope rewrite/transform precisely so
rows/columns that should stay unchanged are kept 1:1; treat header-stated
format rules as checklist items; confirm the artifact file actually exists
before finishing; write back bare values from local scripts
- visual-standards: apply border/header formatting on explicit request and
identify the real header row; keep font size consistent with the source
- range-operations: keep total column width within A4 for printing
- read-data: dedup/compare long numbers via raw values, not csv formatted
display (scientific notation collapses distinct numbers and causes false
duplicates)
- chart: format date/number axes via source-cell number_format; place charts
outside the data area so they do not cover existing data
* feat(sheets): implement table-put/table-get and sync skill specs
- Add lark_sheet_table_io.go with +table-put / +table-get and tests
- Refactor read-data; extend workbook; register new shortcuts
- Sync generated flag defs/schemas (go:embed) from sheet-skill-spec
- Sync skill references (write-cells numeric-column guidance, plus
read-data / workbook / chart updates)
* docs(sheets): surface typed-write path at the write-decision point
Quick-ref table (SKILL.md, the first decision point) had no +table-put and
gated typed writes on "DataFrame", so a model holding a Counter/list/dict
would fall back to +csv-put and silently lose number/date fidelity.
- split csv-put row to plain-text values (no numeric/date semantics)
- add +table-put row for typed writes into an existing sheet
- add +workbook-create --sheets row for create + typed write in one shot
- add judgment note: number/amount/date/percent/count -> +table-put
(or +workbook-create --sheets when the workbook does not exist yet);
plain text -> +csv-put
- reframe write-cells scenario row to lead with numeric semantics
- point new-table writes at +workbook-create --sheets (one shot) instead
of the create-empty-then-table-put two-step
Synced from sheet-skill-spec canonical (generate:cli + sync:cli).
* docs(sheets): sync SKILL.md (drop "not for local Excel" caveat)
Mirror the upstream sheet-skill-spec change removing the "not applicable to local Excel files" tail from the sheets skill and reference descriptions.
* docs(sheets): sync SKILL.md (drop "Feishu sheets only" caveat)
Mirror the upstream sheet-skill-spec change removing the "applies to Feishu sheets only" tail from the 14 sheet reference descriptions.
* feat(sheets): add +workbook-import wrapping the drive import core
Import a local xlsx/xls/csv as a new spreadsheet by delegating to the shared drive import flow with the target type pinned to sheet. Refactor drive +import to expose ImportParams / ValidateImport / PlanImportDryRun / RunImport (behavior unchanged, existing drive tests still cover it); sheets reuses them. Regenerate flag_defs_gen.go and sync the spec mirror.
* refactor(sheets): reuse the drive export core in +workbook-export
Replace +workbook-export's parallel export-task implementation with the shared drive ExportParams/RunExport core (pinned to type=sheet). Drops ~90 lines of duplicated poll/download code; +workbook-export now inherits drive's ctx cancellation, resume-on-timeout, filename sanitize/overwrite, and the full set of export status labels. The output contract aligns with drive's (adds ready/downloaded/doc_type; saved_path preserved). Also normalize an empty drive --output-dir to "." so drive +export behavior is unchanged, and fix the sheets export e2e to call +workbook-export instead of a nonexistent +export.
* docs(sheets): keep original column widths; align chart axis with requested metric
- range-operations: only widen new / overflowing columns; never recompute or
shrink the widths of existing columns (any blanket resize, even by 1px,
breaks the original visual format)
- chart: when the user asks for a share / percentage, the value axis should be
a percentage (pie, or stack.percentage on bar/column) rather than raw counts
* docs(sheets): reword guidance to avoid eval-specific phrasing
Replace scoring-framework wording in the examples with plain functional
consequences (e.g. "not delivered", "goes stale when the source changes",
"breaks the original visual format"), so the references stay agent-facing.
* docs: add lark sheets financial modeling guidance
* docs(sheets): align write-cells reference with the generated output
Bring the hand-applied write-cells example in line with the spec-generated
reference so the CLI mirror is byte-identical to the canonical source.
* docs(sheets): align +csv-put help with formula support
Sync the formula-support wording from sheet-skill-spec (flag-defs, skill
references) and update the hand-authored cobra Description and comment for
+csv-put. +csv-put evaluates a leading-= cell as a formula via
set_range_from_csv; descriptions only, no behavior change.
* docs(sheets): fix invalid +dim-insert example in chart reference
The chart reference's placement example used non-existent flags
--dimension/--start/--end for +dim-insert. The real signature is
--position (required) + --count (required); copying the example
fails Validate with "--position is required". Replace it with
+dim-insert --position V --count 6 (insert 6 columns before V,
i.e. after U), aligning with the sheet-structure reference.
* docs(sheets): chart coordinate base / quoting + filter condition enums
Sync three reference-doc corrections from the spec source:
1. chart: label position.row as 0-based (first row = row:0), distinct
from the 1-based row numbers used by A1 ranges and +dim-insert
--position, removing the row-base ambiguity.
2. chart: convert the three runnable examples whose JSON contains a
quoted sheet prefix ('Sheet1'!A1) from inline single-quoted
--properties '{...}' to a stdin heredoc (--properties - <<'JSON').
Inside an inline single-quoted string bash strips the inner quotes
around the sheet name (and splits names with spaces into words),
corrupting the JSON; a quoted heredoc delimiter performs no shell
substitution and preserves it. Adds a short note on the pitfall.
3. filter / filter-view: add the full conditions[].type x compare_type
enum table (text / number / multiValue / color and their respective
compare_type values and values shape), and call out the
equals/notEquals (with s) vs equal/notEqual (no s) gotcha. The docs
previously only showed two values via examples.
* docs(sheets): label +sheet-create --index as 0-based
The base flag description for +sheet-create's --index omitted the
coordinate base, while its siblings +sheet-move ("Target position
(0-based)") and +sheet-copy already state 0-based. Align the description
so the index base is unambiguous. Synced from the spec source
(flag-defs.json + workbook reference).
* fix(sheets): regenerate flag defs and fix asasalint in table io
* feat(sheets): add counta to chart aggregateType enum
Add `counta` (count non-empty cells, incl. text) to manage_chart_object
dim2.series[].aggregateType in the chart flag schema. `count` only counts
numeric cells, so counting occurrences of a text/category column renders an
empty chart; `counta` enables category frequency counts. Synced from the
sheet-skill-spec canonical schema.
* feat(sheets): make --target-position and --range mutually exclusive on +pivot-create
Both flags map to the same wire field (properties.range), so passing
non-default values for both is ambiguous. Mirror the
--target-sheet-id / --target-sheet-name mutex pattern: --target-position
takes priority over --range, and supplying both with non-default values
is rejected up front with a typed FlagErrorf. --target-position=A1 is
the documented default and is treated as "not set".
Add a symmetric validateCreateInput hook on objectCRUDSpec (alongside
the existing validateUpdateInput), wire it into objectCreateInput, and
inject the pivot-specific check on pivotSpec.
* feat(sheets): rework +workbook-create flags and --styles
- --values builds a type-less typed payload, writing through --sheets' batched set_cell_range path (raw passthrough preserves auto-detect; large tables batch; big ints via json.Number)
- drop --headers (subsumed by --values first row) and --header-style (typed header no longer auto-bold; use --styles instead)
- styles: deep-merge overlapping cell_styles/border_styles fields (was wholesale-replace which dropped fields); add manual border_styles validation (style/weight enums + sides) since --styles is on parseJSONFlagSkip and bypasses the schema validator
- regenerate flag-defs/flag-schemas/skills mirror from sheet-skill-spec (--styles flag + full per-side border schema)
* fix(sheets): add mention_type enum to set_cell_range cells schema
Constrain rich_text mention_type to the proto MENTION_FILE_TYPE set so a
file @mention with an out-of-enum value (e.g. 6 = cloud shared folder) is
rejected by the schema validator before it reaches the server and fails
pb serialization ("mentionFileInfo.fileType: enum value expected").
- data/flag-schemas.json: mention_type gains enum + per-value description
- lark_sheet_write_cells_test.go: cover reject (6) + allow (0 / 2 / 22)
* feat(sheets): implement pandas-split --sheets protocol for +table-put/+table-get/+workbook-create
Synced from sheet-skill-spec canonical (cli:table_put schema +
references). +table-put/+workbook-create accept the new shape via a
tableSheetIn -> tableSheetSpec normalize step (dtype string -> internal
type/format mapping). +table-get emits the same shape so the writer's
df_to_sheet and the reader's sheet_to_df round-trip cleanly.
isoDateToSerial now accepts the full ISO datetime form
(2024-01-15T00:00:00.000, including timezone suffixes) emitted by
df.to_json(date_format="iso"), not just yyyy-mm-dd. End-to-end verified
by the spec repo's contracts/python_helper_roundtrip script against a
real Lark spreadsheet on pandas 2.2 and 3.0.
* feat(sheets): add --dataframe Arrow IPC input for +table-put/+table-get/+workbook-create
Introduce a binary-typed twin of --sheets: --dataframe accepts an Arrow IPC
(Feather v2) payload that pandas' df.to_feather() writes, deriving dtypes and
per-column number formats from the Arrow schema. The two producers are mutually
exclusive and funnel through a shared resolver so +table-put and
+workbook-create stay in lockstep; +table-get gains --dataframe-out for
single-sheet reads. Also auto-grow a sub-sheet's row/column count before
writing so blocks past the backend's default 200x20 bounds no longer fail with
range-exceeds-sheet-bounds.
* docs(lark-sheets): remove financial modeling standards reference
Drop the lark-sheets-financial-modeling-standards.md reference doc and all
pointers to it from SKILL.md, core-operations, and visual-standards. Bump
skill version to 3.0.0.
* docs(lark-sheets): clarify cell-image vs float-image routing and fix reference self-references
Synced from sheet-skill-spec.
- Add a binding-based decision (does the image belong to a record and move with its row?) to route +cells-set-image vs +float-image-create across the SKILL entry, float-image and write-cells references.
- Add routing rows to the SKILL command cheat-sheet and warn against defaulting to float-image out of familiarity.
- Replace mislabeled 本 skill / 子 skill / 跨 skill wording in references with 本文 / reference names, matching the existing convention.
* feat(sheets): add --styles to +table-put for one-step typed write with styling
+table-put now accepts --styles (same shape as +workbook-create's --styles):
cell_styles merge into the set_cell_range matrix, while cell_merges /
row_sizes / col_sizes apply as their own tool calls after the write. The
styles payload is name-matched against the written sheets and validated up
front, so a malformed or mismatched style fails before any write lands.
Also points +sheet-create users to +table-put (auto-creates missing sheets)
when they need data/styles, via a runtime Tip and the lark-sheets skill
references. Flag is sourced from the upstream Base table and regenerated
through sheet-skill-spec (flag-defs.json / flag-schemas.json / gen file).
Adds unit tests (dry-run styles, name-mismatch reject, execute) and a
dry-run E2E (tests/cli_e2e/sheets/sheets_table_put_dryrun_test.go).
* docs(lark-sheets): point read-data to +sheet-info for hidden row/col identification
skip-hidden defaults to false (lossless reads), but the read primitives don't mark which rows/cols are hidden. Cross-reference +sheet-info --include hidden_rows,hidden_cols + row_indices/col_indices so agents can identify hidden ranges when they need to filter or interpret hidden data.
Synced from sheet-skill-spec.
* feat(sheets): document link requirement for @document mentions in cells flag schema
@document mentions (mention_type != 0) must pass link (doc URL) to render a
clickable card; @user mentions (mention_type=0) don't need it. Synced from the
upstream tools-schema.
* fix(sheets): reject cond-format attrs whose shape mismatches rule_type
A conditional-format rule created with --rule-type colorScale but
cellIs-shaped attrs ({compare_type,value}, no color) was accepted by
the CLI and written through to the server, producing a color-less
color-scale segment. That dirty data crashes the frontend on snapshot
deserialization, so the spreadsheet can no longer be opened (5005).
The per-entry schema check can't catch this: properties.attrs.items is
a oneOf over all nine attr shapes and passes as soon as any branch
matches, blind to the sibling rule_type — {compare_type,value} matches
the cellIs branch even when rule_type says colorScale. The tool side
maps attrs blindly by rule_type and only validates dataBar count and
iconSet ordering, so the gap reaches the data layer.
Add a cross-field validator (validateCondFormatAttrs) wired into both
create and update via the new objectCRUDSpec.validateCreateInput hook
(twin of validateUpdateInput). It enforces, per rule_type, the keys
every attrs entry must carry — mirroring the tool's converter contract
— and treats an empty required string (notably color) as missing.
Rule types that take no attrs (duplicateValues / uniqueValues /
containsBlanks / notContainsBlanks) and updates that omit rule_type are
left to the server.
* test(sheets): guard condFormatAttrsRequired against flag-schemas drift
Add TestCondFormatAttrsRequired_MatchesSchemaOneOf, comparing the
hand-maintained condFormatAttrsRequired table against the embedded
flag-schemas.json attrs oneOf (multiset of required-key sets, for both
create and update). The cross-field validator only holds if its
per-rule_type required keys mirror the schema branches, and the two
share no compile-time link — this pins them together so a future schema
sync that adds/drops a required key can't silently desync the table.
* fix(sheets): default +table-get to full used range, not A1 current region
+table-get without --range anchored its current_region probe at A1, so an
internal blank row or column silently truncated everything past it — agents
then treated the partial data as complete (the pro016 / pro025 incident).
- Probe the used range over the full physical grid (row_count × column_count
from the workbook structure) so it spans internal blank rows/columns; fall
back to the legacy A1 anchor when dimensions are unknown.
- Emit the actually-read `range` on every sheet so callers can detect
truncation (get_cell_ranges has no has_more flag).
- Fix the same A1-anchor bug in append mode's last-data-row probe, which could
otherwise overwrite data past an internal blank row.
- Add unit + dry-run/live E2E coverage; refresh synced skill docs.
* docs(sheets): fix csv-get current_region guidance to cross-check row_count
current_region is a blank-row/column-bounded block, not the true sheet extent:
an internal blank row truncates it, so it can miss rows past the gap. The
read-data reference previously called it the "真实数据边界" and told agents to
prefer it over row_count — which drove the "read only to current_region's last
row, miss the tail" failure.
- current_region: warn it can be both smaller (internal blank rows truncate)
and larger (trailing summary/signature rows) than the real data range.
- csv-get output contract: clarify its row_count/col_count is the returned size
(= actual_range), not the physical sheet size; has_more only reflects the
current range, not whether the whole sheet was read.
- "确定数据范围的正确流程": add a step to cross-check against +workbook-info's
physical row_count and probe past current_region's last row for data beyond an
internal blank row.
* fix(sheets): collapse duplicate validateCreateInput from bad merge resolution
A prior merge kept both branches' independently-added validateCreateInput
fields on objectCRUDSpec with conflicting signatures (pivot's
func(rt, input) and cond-format's func(input)), plus both call sites in
objectCreateInput, which failed to compile (validateCreateInput redeclared).
Collapse to the single richer func(rt flagView, input) signature and one
call site. cond-format's validateCondFormatAttrs (func(input), still shared
with validateUpdateInput) is wrapped in a closure that ignores rt. Both
behaviors are preserved: pivot --target-position/--range mutex and
cond-format attrs-shape-vs-rule_type validation.
* refactor(sheets): migrate legacy error helpers to typed errs in sheets domain
golangci-lint forbidigo (errs-no-legacy-helper / errs-no-bare-wrap) flagged
the table I/O, workbook, and dataframe shortcuts that landed on this branch:
93 common.FlagErrorf and 48 fmt.Errorf calls.
- Replace every common.FlagErrorf with common.ValidationErrorf (typed
*errs.ValidationError, same signature) across workbook / table_io /
dataframe / object_crud.
- writeDataframeOut's two final --dataframe-out write failures become typed
errs.NewInternalError(SubtypeFileIO, ...).WithCause(err).
- applyWorkbookCreateVisualOps now passes the typed callTool error through
unchanged (re-wrapping would downgrade classification) and attaches the
failing op as a recovery hint only when none is set.
- The remaining fmt.Errorf are genuine intermediate errors that the command
layer re-wraps into typed validation errors (buildTypedCell / Arrow
decode-encode) or surfaces as a partial_success message string
(writeTypedSheets via tablePutPartial); each carries a //nolint:forbidigo
with that reason, per the lint guidance.
No behavior change: error messages and partial-success shapes are preserved;
gofmt, go vet, golangci-lint (0 issues) and sheets tests all pass.
* fix(shortcuts): clarify single-stdin constraint in flag help and error hint
Input flags advertised '(supports @file, - for stdin)' per flag, leading
AI agents to write '--a - <x --b - <y' where the second '<' silently
clobbers the first and the first flag reads the wrong payload. A process
has a single stdin, so at most one flag per call can use '-'.
- Reword the generated help hint to '- reads stdin (one flag per call;
use @file for others)'.
- Add an actionable .WithHint to the stdin-conflict validation error
pointing callers to @file for the extra flags.
- Assert the new hint in TestResolveInputFlags_DuplicateStdin.
* feat(sheets): +cells-get/+csv-get --max-chars 默认值 200000 → 500000
放宽默认防爆上限。flag_defs_gen.go 由 go generate 重生;flag_defs_test.go
的 expected default 同步;flag-schemas.json schema_version 2 → 3 是上游
spec-tables 架构调整带来的元数据 bump,与本业务改动无关、go:embed 不解析
该字段、无功能影响。
Synced from sheet-skill-spec@93f7a78.
* docs(lark-sheets): sync from spec — +csv-put 含逗号公式正例 + 收敛警示标签
源同步自 sheet-skill-spec:write-cells 补含逗号公式 RFC 4180 转义正例与结构化写入优先指引;全 reference 收敛「高频致命错误」类标签。
* docs(lark-sheets): sync from spec — --max-chars 放出为可见 flag + 落盘优先指引
源同步自 sheet-skill-spec:--max-chars 放出(默认 500000,可调小避免大输出被 Bash/终端转存为文件、改 has_more 分页);read-data 增「大数据优先落盘」指引。
* feat(sheets): 写操作报错增强 + --token 别名
- 复合 JSON shape 校验失败时报错附 --print-schema 提示,agent 可直接拿到精确结构(pro26 头号:+cells-set --cells 反复猜 shape)
- JSON 解析失败且该 flag 支持 stdin 时提示改用 stdin(公式/引号/逗号内联到 shell 被转义弄坏 JSON)
- --token 作为 --spreadsheet-token 的解析期别名:复用 sheets 已有 PostMount 钩子 + pflag normalize,仅 sheets 包,common 零改动
* docs(lark-sheets): sync from spec — set+H 改单引号 / 速查表补臆造命令名 / workbook-import 引导
* fix(sheets): migrate +table-put to typed error contract
The merge from main brought in #1449 (retire legacy error envelopes),
which removed output.ExitError / output.ErrDetail and forbids
constructing them. Port tablePutPartial off the legacy envelope:
- no sheets written -> typed errs.APIError (plain failure)
- some sheets written -> ok:false result via runtime.OutPartialFailure
carrying written_sheets, returning the partial-failure exit signal
Also fix two drifts the same merge introduced:
- regenerate flag_defs_gen.go to match the committed flag-defs.json
- update the --max-chars flag test to assert visible (no longer hidden)
* docs(lark-sheets): sync from spec — set+H 告诫通则化(移入 stdin 段)
* feat(sheets): styles 接受 halign/valign 等对齐字段别名
把模型常幻觉的 horizontal_align / halign / vertical_align / valign 映射到
规范字段 horizontal_alignment / vertical_alignment,覆盖 --styles 与 typed
--cells;与规范字段冲突时报错而非静默择一。同步 lark-sheets skill 文档补
对齐字段说明 + --print-schema --flag-name styles 提示。
* feat(sheets): resolve wiki URLs to the backing spreadsheet for --url
Sheets shortcuts only accepted /sheets/ and /spreadsheets/ URLs via --url.
A /wiki/<node_token> URL was rejected with "must be a spreadsheet URL"
because the wiki node_token is not a spreadsheet token: resolving it to the
backing spreadsheet needs a wiki get_node call, which Validate/DryRun (kept
network-free) must not make.
Mirror the existing slides/doc/drive two-stage pattern:
- parseSpreadsheetRef classifies --url / --spreadsheet-token network-free
into a sheet token or an (unresolved) wiki node_token.
- resolveSpreadsheetTokenExec (Execute only) resolves a /wiki/ node_token
via wiki get_node, verifies obj_type=sheet, and returns the obj_token.
The wiki:node:read scope is enforced on this path only, so non-wiki
invocations are unaffected.
- resolveSpreadsheetToken stays network-free for Validate/DryRun, passing
the node_token through unchanged.
All 47 Execute paths (including +batch-update and +workbook-export) switch
to the Exec resolver; Validate/DryRun keep the network-free one. No tool
schema change: the CLI feeds the resolved spreadsheet token as excel_id, so
this is a pure CLI-layer change.
Tested: unit (parse classification + wiki get_node e2e via httpmock) and
live end-to-end against a real wiki spreadsheet (read: +workbook-info,
+cells-get, +csv-get; write: +sheet-create, +sheet-rename, +csv-put).
* docs(sheets): note --url accepts wiki URLs (synced from spec)
* fix(sheets): match --url path segment via url.Parse, not substring
parseSpreadsheetRef classified /wiki/ with strings.Index over the whole URL, so a /sheets/ link whose query or fragment merely contained /wiki/ (e.g. .../sheets/sht?from=/wiki/x) was hijacked into a get_node call. Now parse the URL and match /sheets/, /spreadsheets/, /wiki/ only as a path prefix, mirroring slides parsePresentationRef which already fixed this class. Drop the substring helpers. Also align wiki resolution with slides: CallAPITyped (typed error + log_id) and classify an incomplete get_node response as InternalError instead of a --url validation error. Add regression tests for query/fragment /wiki/ and incomplete node.
* fix(sheets): satisfy errorlint/copyloopvar + regen flag defs
- helpers_test.go: drop the Go 1.22+ redundant `tc := tc` loop copy
(copyloopvar).
- lark_sheet_dataframe.go, lark_sheet_table_io.go: switch the
intermediate-error fmt.Errorf calls from %v to %w so errorlint passes.
Behavior unchanged — these errors are always rewrapped into typed
validation errors at the command layer.
- flag_defs_gen.go: regenerate from data/flag-defs.json (drift from the
wiki-URL merge).
* ci: allow Apache Arrow module in license check
Arrow is Apache-2.0 overall, but it vendors c-ares (LicenseRef-C-Ares,
ISC-like) inside the module which go-licenses classifies as Unknown and
the strict disallowed_types=...,unknown gate rejects.
Pass --ignore github.com/apache/arrow/go/v17 since Arrow is required by
sheets +table-put / +table-get / +workbook-create --dataframe (Arrow IPC
ingest) and the vendored c-ares is not redistributed by us.
* fix(sheets): resolve wiki URL in +range-move/+range-copy Execute
transformExecuteFn (the named Execute helper shared by +range-move and +range-copy) still called the network-free resolveSpreadsheetToken, so a /wiki/ URL reached transform_range as an unresolved node_token and failed. #1519's sweep over Execute hooks only rewrote inline closures; this is the only Execute backed by a named helper. Switch it to resolveSpreadsheetTokenExec (Validate/DryRun stay network-free) and add a +range-move wiki-URL regression test.
* refactor(sheets): drop +table-put manual capacity grow; rely on set_cell_range auto-grow
set_cell_range now auto-grows the sub-sheet to fit the write, so the
ensureSheetCapacity helper (and its modify_sheet_structure dim-insert
call before each write) is no longer needed. This also closes a data-
safety hole flagged in review: inserting before the last existing row
could push real data down into the area set_cell_range was about to
write, and allow_overwrite=false could not protect against it because
the structural insert had already mutated the sheet by the time the
write-collision check ran.
Verified end-to-end against a real spreadsheet: +table-put writing
300x25 into a fresh Sheet1 (default 200x20) succeeds in one write and
the sheet ends up 301x25.
* fix(sheets): close --dataframe stdin guard hole
--dataframe is binary and bypasses the common Input resolver, which is
where the existing single-stdin guard lives. Result: an invocation like
+table-put --dataframe - --styles - was accepted, then one of the two
consumers raced for stdin and the other silently saw an empty stream.
Add a stdinConsumed marker on RuntimeContext that both consumers share:
common.resolveInputFlags sets it when an Input flag uses '-', and
readDataframeBytes both checks and sets it. A second consumer is
rejected up front with an actionable hint pointing at @file.
Flagged in code review (lark_sheet_dataframe.go:93).
* fix(sheets): harden +table-put / +table-get input validation and round-trip safety
Four review-flagged correctness gaps in table I/O, all bundled because
they touch the same file:
1. --sheets accepted trailing data after the first JSON value
(json.Decoder does not surface that, unlike json.Unmarshal). A new
decoderExpectEOF helper rejects e.g. `--sheets '{...} oops'` with a
typed validation error instead of letting the leading object pass
through and surface as a confusing downstream failure.
2. +table-get with a duplicate header (e.g. `amount, amount`) used to
read back successfully — the dtypes map silently collapsed to one
entry — and only failed later on +table-put because the writer
rejects duplicate column names. Fail fast at read time with an
actionable hint to rename or pass --no-header. --no-header mode is
exempt (fallback col<N> names are always unique).
3. +table-put dry-run rendered an invalid range like A1:C0 when
header=false with rows=[]. tablePutFullRange returns "" for an
empty matrix or zero columns instead of building a degenerate
rectangle.
4. +table-get with --sheet-id and a get_workbook_structure miss (read
failure or selector mismatch) used to return a target with
name="", which then broke +table-get → +table-put round-trip (the
writer requires a non-empty sheet name). Fall back to using the id
as the name.
End-to-end verified against a real spreadsheet: trailing data, duplicate
header, and --no-header fallback all behave as advertised.
* fix(sheets): apply +workbook-create style-only ops instead of silently dropping them
A +workbook-create call carrying only cell_merges / row_sizes / col_sizes
(no --values / --sheets and no cell_styles) used to create the workbook
but silently drop the requested visual ops. Two reasons, both fixed:
- workbookCreateStyleDimensions only counted cell_styles when computing
the write extent, so cell_merges / row_sizes / col_sizes always
contributed 0 → buildValuesPayload returned a nil payload → Execute
skipped writeTypedSheets entirely → no visual ops ran. Extend the
helper to fold the merge / resize ranges in.
- Pure row_sizes / col_sizes payloads can never expand a cell rectangle
(they are dimension ranges, not cell ranges), so even with the extent
fix Execute would still skip the write path. Add a no-data branch:
when payload == nil but a styles item is present, look up the default
sheet and apply visual ops directly via applyWorkbookCreateVisualOps.
The dry-run plan mirrors this so the preview shows the visual ops.
Also picks up the --values trailing-JSON-data EOF check (mirror of the
--sheets one in lark_sheet_table_io.go).
End-to-end verified against a real spreadsheet: a cell_merges-only
+workbook-create now produces a sheet with merged_cells_count: 1.
* fix(sheets): preserve causes and render messages cleanly for typed validation errors
common.ValidationErrorf goes through fmt.Sprintf, which does not support
%w — the seven call sites that used `%w` were rendering the cause as
literal `%!w(*fmt.wrapError=&{...})` and dropping the cause from the
typed-error chain (so callers couldn't errors.As back to the underlying
error).
Switch each to `%v` for clean rendering and attach the cause via
.WithCause(err) so the typed contract is preserved. Touched call sites:
- lark_sheet_dataframe.go: --dataframe Arrow decode / stdin read / file
read failures (3 call sites).
- lark_sheet_table_io.go: --sheets invalid JSON, payload-validate
per-cell coercion error, buildSheetMatrix per-cell error,
--dataframe-out arrow encode failure (4 call sites).
End-to-end verified against a real spreadsheet: both invalid-JSON and
typed-cell errors now render readable messages instead of %!w(...).
* sync(sheets): pick up +sheet-{show,hide}-gridline in +batch-update schema
Mirror of the sheet-skill-spec change adding the two gridline shortcuts
to cli-schemas.json batch_update.operations.shortcut enum. Synced from
the upstream canonical via generate:cli + sync:cli.
Verified end-to-end on a real spreadsheet — +batch-update with a
+sheet-hide-gridline op passes schema validation and the backend run
returns succeeded: 1.
* sync(sheets): pick up +workbook-export UX clarification from spec
Mirror of the sheet-skill-spec update that documents +workbook-export's
default-no-download behavior and its relationship to drive +export
--doc-type sheet. Synced from canonical via generate:cli + sync:cli +
go generate.
End-to-end verified against a real spreadsheet:
- Omit --output-path → ok:true, downloaded:false, file_token returned
- Pass --output-path ./crfix_test.xlsx → ok:true, file saved
(17892 bytes), saved_path returned
The --help output for +workbook-export now states the default behavior
and points callers at `drive +export --doc-type sheet` when they need
the --output-dir / --file-name / --overwrite split.
* test(sheets): assert typed errs.Problem instead of err.Error() substrings
Per the coding guideline "Error-path tests must assert typed metadata via
errs.ProblemOf (category / subtype / param) and cause preservation, not
message substrings alone." — sweep through every error-path assertion in
the sheets domain and replace the
`strings.Contains(stdout+stderr+err.Error(), ...)` pattern with two
small helpers landed in helpers_test.go:
requireProblem(t, err, wantCategory, wantSubtype, msgContains)
-> *errs.Problem
requireValidation(t, err, msgContains)
-> *errs.ValidationError // shorthand for CategoryValidation +
// SubtypeInvalidArgument; lets callers
// also assert .Param / .Params / .Cause
~60 assertion sites across 18 test files now check the typed envelope
shape, with message-substring checks moved onto the returned Problem
(.Message / .Hint / .Param). The substring is preserved as a sanity
check rather than the sole assertion, so a category drift like
validation → internal would now fail loudly instead of slipping past.
Cases intentionally left as substring (each with a one-line reason):
- Errors that come straight from cobra's native flag parser (untyped
*errors.errorString — e.g. "required flag(s) ... not set", mutually-
exclusive groups). Re-typing these needs a custom FlagErrorFunc and
is out of scope here.
- Intermediate errors from decodeArrowToSheet that the caller wraps
into a typed envelope (`//nolint:forbidigo` reason). Those unit
tests assert the unwrapped intermediate directly.
One production tweak:
- shortcuts/sheets/flag_schema.go: printFlagSchemaFor returns typed
*errs.ValidationError (with WithParam("--flag-name") on the
unknown-flag branch) instead of raw fmt.Errorf. The framework
already wraps this when called via --print-schema, so user-facing
behaviour is unchanged; direct callers (and tests) now get the
typed envelope.
Verified: go test ./shortcuts/sheets/... passes; golangci-lint
--new-from-rev=origin/main reports 0 issues.
* test(common): assert typed errs.Problem instead of err.Error() substrings
Mirror of the sweep just landed in shortcuts/sheets: replace error-path
substring assertions with typed-envelope checks via two small helpers
landed in a new shortcuts/common/typed_error_assertions_test.go:
requireProblem(t, err, wantCategory, wantSubtype, msgContains)
-> *errs.Problem
requireValidation(t, err, msgContains)
-> *errs.ValidationError // shorthand for CategoryValidation +
// SubtypeInvalidArgument; lets callers
// also assert .Param / .Params / .Cause
8 sites moved to typed assertions across runner_jq_test.go,
mcp_client_test.go, drive_media_upload_typed_test.go, and
runner_input_test.go (the input tests already used a typed-param helper;
this just retargets the substring follow-up onto the typed Message).
Sites intentionally left as substring + comment (production returns raw
fmt.Errorf, not a typed envelope):
- runner_botinfo_test.go (6 sites): BotInfo / fetchBotInfo wrap upstream
errors with fmt.Errorf so the SDK-level message ([99991], 403,
invalid character, etc.) shows through.
- runner_args_test.go (4 sites in 2 tests): rejectPositionalArgs returns
raw fmt.Errorf to satisfy cobra's PositionalArgs contract.
- permission_grant_test.go (2 sites): assert on stderr / hint strings,
not error messages — already out of the err.Error() substring class.
No production code changes.
Verified: go test ./shortcuts/common/... passes;
golangci-lint --new-from-rev=origin/main ./shortcuts/common/... reports
0 issues.
* fix(sheets): plug four +table-put / +table-get correctness gaps flagged in CR
Four review-flagged bugs, all in lark_sheet_table_io.go (bundled because
they touch the same file and the same +table-put / +table-get domain):
1. +table-get --dry-run dropped the --sheet-id / --sheet-name selector
from the get_cell_ranges body, while Execute always passed it. Agents
that validate the dry-run shape and then run live would see a request
shape mismatch. The dry-run now calls sheetSelectorForToolInput so
the body matches Execute.
2. isDateNumberFormat used a simple `strings.ContainsRune(_, 'y')` so
number formats like "JPY #,##0" (a currency prefix that happens to
contain a lone 'Y') were misread as date formats — round-tripping
integer cells out as ISO dates. The detector is now token-aware:
it skips quoted "...", `\\x`-escaped, and `[...]` bracket sections,
and only fires on an unescaped `yy` (a real Excel year token).
3. sheetCreateDims sized new append-mode sheets by `headerOn(s)` only,
but writeSheetData forces a header on empty append sheets when
Header == nil. Near 50000 rows / 200 cols this created the sheet one
row short and the follow-up set_cell_range bounced off the backend
ceiling. Size now matches the forced-header logic exactly.
4. tableGetTargets fallback paths (read-failure / selector mismatch on
--sheet-id) returned a target with name="" — already corrected for
--sheet-id structure-success path in 086876d2, but the structure-
failure fallback still left it empty. Use the id as the name there
too so the +table-get → +table-put round-trip never breaks on a
nameless sheet.
End-to-end verified against a real spreadsheet:
- table-get --dry-run with --sheet-name / --sheet-id both render the
selector field in the get_cell_ranges body
- A real round-trip (typed put → get) preserves dtypes + formats
* fix(sheets): bound --dataframe memory use with byte / row / column caps
readDataframeBytes used to read the whole Arrow file unbounded — a
stdin / file > 1 GiB would OOM the CLI long before the backend
per-sheet ceilings kicked in. decodeArrowToSheet then materialized
every record into [][]interface{} regardless of size.
Three caps now match the backend's per-sheet hard ceilings:
- byte cap: 256 MiB (covers worst-case 200×50000 cells × ~25 B Arrow
overhead). File path pre-Stat()s before opening; both file and stdin
paths read through io.LimitReader so an oversized input is rejected
without allocating the full payload.
- column cap: 200, checked at schema-decode time before allocating any
per-column slices.
- row cap: 50000, checked during record-batch iteration so a 1M-row
Arrow file is rejected mid-stream instead of fully decoding first.
End-to-end verified against PPE — a 257 MiB file is rejected at file-
Stat with a typed validation error before any read happens.
* fix(drive): wrap +export ctx cancellation/deadline as typed errs.NetworkError
The poll loop in RunExport returned ctx.Err() directly in two places —
on the inter-attempt sleep cancel and on the pre-attempt deadline check.
That let context.Canceled / context.DeadlineExceeded escape as untyped
errors at the cobra layer, bypassing the typed-error contract every
other failure path already honors.
Add wrapExportContextErr that maps both into errs.NewNetworkError with
SubtypeNetworkTransport / SubtypeNetworkTimeout respectively and
preserves the cause via .WithCause(err), so callers can still
errors.Is(err, context.Canceled) downstream.
CR-flagged at drive_export.go:229 / :234.
* ci(license): narrow Apache Arrow workaround with a follow-up assertion
The dependency-license check still has to --ignore Apache Arrow wholesale
because go-licenses' classifier parses its LICENSE.txt as a single license
and mis-reports the module as LicenseRef-C-Ares / Unknown (Arrow inlines
the c-ares 3rdparty notice alongside its own Apache-2.0). Re-classifying
on our side isn't possible without changing go-licenses itself.
The CR concern was that --ignore is too wide — a future Arrow re-license
or new inlined dep would silently sail through. Add a follow-up step that
re-checks Arrow's LICENSE.txt independently: it must still open with
"Apache License" AND must still inline the c-ares 3rdparty notice (the
two facts that make the --ignore safe today). If either invariant breaks,
CI fails here and forces a human to re-evaluate the ignore.
Verified locally — both assertions pass against the current pinned
Arrow v17.
* sync(sheets): pick up +table-put payload-shape doc corrections from spec
Mirror of the sheet-skill-spec change that fixes three places teaching
an invalid +table-put payload shape — the typed protocol only has
columns / data / dtypes / formats (no formula field) and must always
be wrapped in an outer {"sheets":[...]} envelope. write-cells and the
SKILL.md decision table previously used the wrong field names (type /
format) and pointed users at +table-put for formula writes, which the
shortcut can't actually accept.
Synced from upstream canonical via generate:cli + sync:cli.
* test(sheets/e2e): add E2E coverage for new shortcuts + typed workbook-create
AGENTS.md requires a dry-run E2E for every new shortcut and a live E2E
for new flows. Three new files cover the four shortcuts this branch
adds or materially changes:
- sheets_gridline_dryrun_test.go — pins +sheet-show-gridline /
+sheet-hide-gridline as a single modify_workbook_structure call with
the right operation name (show_gridline / hide_gridline) and
sheet_id, so an op-name typo would trip CI before any live run.
- sheets_workbook_import_dryrun_test.go — pins +workbook-import as a
two-step plan (drive media upload + drive import-task create) with
the doc type hard-coded to "sheet" — the wrapper's whole reason for
existing on top of generic drive +import. --name reaches file_name
on the wire; file_extension is sniffed from the local file.
- sheets_table_put_typed_workflow_test.go — two live workflows running
against a freshly created spreadsheet. The first runs the full
typed +table-put → +table-get round-trip (date / numeric / object
columns with custom number_format) and asserts the dtype + format
contract holds end-to-end. The second exercises the typed
+workbook-create --sheets path: create + write in one shortcut, the
payload sheet name adopts the workbook's default sheet (no empty
"Sheet1" left behind), and the typed contract still survives the
read-back.
End-to-end verified locally (user identity): typed put round-trips
preserve dtypes (date → datetime64[ns], numeric → float64, object →
object) + formats verbatim; workbook-create adopts the named sheet as
the first sheet with the same typed shape intact.
* sync(sheets): pick up sheets_df.py — pandas ↔ JSON skill script from spec
Mirror of the sheet-skill-spec change that adds a DataFrame ↔ JSON
bridge as a skill-bundled Python script instead of inside the CLI
binary. Per PR #1355 review (docx NcmxdRo2yoZ4OXxoMUZcxRZ7nHd, §4.2):
keep the CLI a thin JSON/REST client; pandas / Arrow editing lives in
the caller's Python process. Synced from canonical via generate:cli +
sync:cli.
- skills/lark-sheets/scripts/sheets_df.py (new): pandas DataFrame ↔
one sheet, .parquet / .feather / .arrow / .csv / .json. Shells out to
`+table-put` / `+table-get` over typed JSON — no CLI changes.
- SKILL.md decision tree + write-cells.md +table-put section: explicit
pointers so pandas users land on the script instead of hand-rolling
the `--sheets` payload.
End-to-end verified against PPE: 3-row DataFrame (datetime / float /
object) round-trips parquet → script put → real sheet → script get →
parquet with dtypes preserved.
* Revert "sync(sheets): pick up sheets_df.py — pandas ↔ JSON skill script from spec"
This reverts commit 2964983b92.
* sync(sheets): pick up sheets_df.py + doc DRY cleanup from spec
Mirror of the sheet-skill-spec change that ships a 32-line helper-only
sheets_df.py (df_to_sheet + sheet_to_df) and removes the corresponding
inline `def` blocks from three reference docs.
- skills/lark-sheets/scripts/sheets_df.py (new): pandas DataFrame ↔
one +table-put / +table-get sheet, importable as a library. Same
helper pair the docs already taught, lifted out of the prose so
callers can `from sheets_df import df_to_sheet, sheet_to_df`.
- lark-sheets-write-cells.md / lark-sheets-read-data.md /
lark-sheets-workbook.md: drop the inline helper definitions; keep
the usage examples (single/multi-sheet, round-trip) and switch them
to import-from-script. workbook reference's +workbook-create
--sheets section now points pandas users at the helper directly
(was previously a textual reference back to write-cells).
End-to-end verified against PPE (--as user):
- +workbook-create with df_to_sheet for three sheets (income / balance
/ cashflow): create ok, dtypes (datetime64[ns] / float64) + formats
(#,##0 / 0.0% / yyyy-mm-dd) survive on read-back through sheet_to_df.
- read → pandas mutate → write-back round-trip preserves both data
and formats.
* chore: drop accidentally-committed __pycache__/ and gitignore .pyc
The previous commit (5fac9c39) shipped sheets_df.py and inadvertently
included its `__pycache__/sheets_df.cpython-312.pyc` — local Python
import created the bytecode cache during PPE round-trip verification and
`git add skills/lark-sheets/` swept it in.
Remove the pyc and add Python bytecode patterns to .gitignore so the
skill-bundled helper scripts don't pull cache files into future commits.
* refactor(sheets): drop --dataframe / --dataframe-out + apache/arrow dep
Per the design review at NcmxdRo2yoZ4OXxoMUZcxRZ7nHd, the Arrow IPC binary
input/output channel adds a heavy columnar runtime to the CLI for no new
capability — the typed JSON --sheets path already covers everything, and
the column-major / zero-copy advantages collapse the moment the CLI re-
encodes into the row-oriented sheets OpenAPI JSON body. Removing it also
lets us drop the `--ignore github.com/apache/arrow/go/v17` license-check
escape hatch.
Deleted:
- shortcuts/sheets/lark_sheet_dataframe.go (+ test)
- --dataframe branches in +table-put / +workbook-create
- --dataframe-out branch in +table-get
- StdinConsumed / MarkStdinConsumed exported methods (the binary stdin
reader was the only out-of-band consumer); internal stdinConsumed
guard against duplicate `-` input flags stays
- apache/arrow/go/v17 + transitive deps via `go mod tidy`
- CI go-licenses --ignore for arrow and the LICENSE.txt assertion step
- --dataframe / --dataframe-out coverage in skill references
Pandas users keep the round-trip via the existing skill script
skills/lark-sheets/scripts/sheets_df.py over the JSON path.
The full pre-removal state is preserved on branch feat/sheets-arrow-stash.
Upstream sheet-skill-spec follow-up: the two flag rows in the canonical
spec + base table tblV2F6fqIjyCFQW must also be dropped so the next sync
does not re-add them.
* sync(sheets): pick up --sheets one-liner fix from spec
Mirrors sheet-skill-spec 5562f83. The +table-put / +workbook-create
--sheets flag descriptions (and the --print-schema description on the
sheets array) now point at the existing df_to_sheet helper instead of
the previous misleading one-liner that produced a dict missing the
outer {"sheets":[...]} envelope and the per-sheet `name`. Agents that
copy-paste the description verbatim now build a valid payload.
Auto-synced via spec's generate:cli + sync:consumers; go generate
./shortcuts/sheets/... regenerated flag_defs_gen.go so its embedded
flagDefs stays byte-equal to data/flag-defs.json.
* test(sheets/e2e): close E2E coverage gaps for newly added shortcuts
AGENTS.md requires both dry-run and live E2E for every newly registered
shortcut, and behavior-changing refactors need at least the matching
half. Three gaps remained on feat/lark-sheets-develop:
- +sheet-show-gridline / +sheet-hide-gridline (new): only dry-run E2E.
Add sheets_gridline_workflow_test.go — create a real spreadsheet,
toggle hide then show against a live sub-sheet, assert ok=true on
both (gridline state is write-only — there is no read-back field on
+sheet-info / +workbook-info — so a successful envelope is the
meaningful signal; the dry-run E2E already pins the wire shape).
- +workbook-import (new): only dry-run E2E. Add
sheets_workbook_import_workflow_test.go — write a local CSV, run
the full upload → create-task → poll, assert ready=true with a
sheet token, +info confirms the imported workbook is reachable,
cleanup deletes the spreadsheet.
- +workbook-export refactor (no-download default changed): had live
E2E but no dry-run E2E in tests/cli_e2e/. Add
sheets_workbook_export_dryrun_test.go — pin the three sheet-
specific differences vs drive +export: type=sheet hard-coded,
csv mode routes --sheet-id onto sub_id (xlsx mode omits it), and
--output-path maps onto the dry-run plan's top-level output_dir.
Also pins the csv-without-sheet-id validation error.
* refactor(sheets): unify workbookCreatedButFillFailed with OutPartialFailure
Three "made it halfway and stopped" exits in the sheets domain previously
disagreed on shape, which made the post-failure recovery flow hard for
agents to predict from one command to another:
- +table-put partial write → exit 1, stdout ok:false envelope
- +table-put zero-sheet write → exit 1, stderr api/server_error
- +workbook-create create-but-fill → exit 2, stderr validation/failed_precondition
OutPartialFailure exists exactly for "the side effect landed but the
follow-up didn't" — it stamps an ok:false result envelope on stdout
(carrying the state the caller needs to recover) and returns the bare
partial-failure exit signal. The workbook-create fill-failure path was
the odd one out: it surfaced as a typed failed_precondition error on
stderr, which agents couldn't tell apart from a plain validation refusal
even though the spreadsheet really did exist and a retry / cleanup was
possible.
Migrate workbookCreatedButFillFailed onto OutPartialFailure so the four
call sites in +workbook-create's Execute (sheet-resolve failure, initial
fill failure, style-only resolve failure, style-only apply failure) emit
the same envelope shape +table-put's partial write does:
{
"ok": false,
"data": {
"spreadsheet_token": "shtNEW",
"reason": "spreadsheet shtNEW created but initial fill failed",
"hint": "the spreadsheet exists; retry the fill … or delete it",
"cause": {"category": "...", "subtype": "...", "message": "..."}
}
}
The inner failure's typed problem (category / subtype / message) is
flattened into the `cause` field so agents stay diagnosable from the JSON
envelope alone, instead of having to errors.Unwrap a Go error.
Updated TestExecute_WorkbookCreate_FillFailureKeepsToken to assert the
new shape (ok:false envelope on stdout, *output.PartialFailureError exit
signal, structured cause carrying the underlying invalid_response
subtype) — preserving the original test intent (token must survive for
recovery; inner cause must stay diagnosable) under the new contract.
* chore(sheets): three review nits — WithCause + stale comment + unexport
- shortcuts/sheets/flag_schema_validate.go:106 — composite-JSON shape
validation was wrapping vErr's message into a typed sheets validation
error without preserving vErr as the typed cause; add the missing
.WithCause(vErr) so errors.Unwrap and ProblemOf still find the
underlying validator error (matches every other typed-error chain
helper in the file).
- shortcuts/sheets/lark_sheet_batch_update.go:92 — comment claimed
batchUpdateInput returns "FlagErrorf-typed errors", but FlagErrorf no
longer exists (the typed-error migration replaced it with
common.ValidationErrorf / errs.ValidationError); update the comment
to reflect what is actually returned.
- shortcuts/drive/drive_export.go:121 — drop the ValidateExport public
alias and rename to validateExport. sheets +workbook-export reuses
RunExport / PlanExportDryRun from this package but inlines its own
(sheet-specific) Validate, so there is no cross-package call site —
ValidateExport was a misleading sibling of the genuinely-shared
ValidateImport. Comment added to record the asymmetry so future
readers do not export it back.
* chore(deps): drop stale indirect bumps left by the arrow removal
The earlier --dataframe / --dataframe-out + apache/arrow/go/v17 removal
deleted the arrow consumer but left two indirect lines in go.mod pinned
to the versions arrow had pulled in:
- github.com/kr/text v0.2.0
- golang.org/x/exp v0.0.0-20240222234643-814bf88cf225
With arrow gone, larksuite/cli was the only requirer of those exact
versions; every real consumer needs lower ones (kr/pretty wants
kr/text v0.1.0; charmbracelet/huh wants x/exp …20231006; xo/terminfo
wants x/exp …20220909). Removing the two indirect lines and running
`go mod tidy` lets MVS pick the real-consumer versions and drops the
explicit indirect entries entirely — go.mod net-diff against main is
now zero for this branch.
Verified locally: go build ./...; go test ./shortcuts/sheets/...
./shortcuts/drive/... ./shortcuts/common/... ./internal/auth/...
./cmd/auth/... — all green.
---------
Co-authored-by: zhengzhijie <zhengzhijie.j@bytedance.com>
Co-authored-by: Chenweifeng-bd <chenweifeng.1534@bytedance.com>
* feat(base): add URL and title resolve shortcuts
* docs: clarify base coordinate resolution
* fix(base): address resolve shortcut ci
* fix(base): format resolved record share hint
* fix(base): simplify record share hint data
* fix(base): use field ids in resolved record data
* fix(base): guide record share resolve to update record
* fix(base): include record upsert example in resolve hint
* fix(base): reject add-record urls in resolver
* fix(base): validate title resolve query length
* fix(base): hide resolve alias flags from help
* fix(base): prefer title flag for title resolve
* docs(base): clarify token resolution wording
Add support for card.action.trigger, the event fired when a user interacts with an
interactive card (button click, form submit, dropdown, checkbox, input, date picker, etc.).
The handler flattens the V2 envelope into a structured output and auto-fetches the original
card content (card_content) at consume time, enabling a complete read-then-update workflow
without extra API calls.
Fix IM shortcut behavior for audio messages to match the Feishu/Lark file upload API: --audio is for voice messages and supports only Opus audio. Non-Opus local/URL inputs such as mp3 and wav are now rejected before upload with an actionable typed validation error. Users can still send those files as attachments with --file
* fix: reject +init into a different app's project directory
* fix: reject single HTML files larger than 10MB in +html-publish
* docs: clarify publish visibility, domain routing, and role/permission boundary
buildListParams used to re-call resolveFolderID / resolveFolderName (and the
label counterparts) on every list page to assemble folder_id / label_id.
Because resolveListFilter already resolves the filter once before the
pagination loop, the second pass hit the folders/labels list API again on
every page — 1 + page_count calls total, which easily trips rate limits.
buildListParams now only assembles API params from the already-resolved
FolderID / LabelID produced by resolveListFilter; it no longer resolves
names or aliases. The default folder_id=INBOX is still applied when no
explicit filter is present, and only overridden when the caller supplied a
canonical folder ID. The runtime / mailboxID / dryRun parameters are kept
for signature stability (resolveListFilter and buildSearchParams share the
same call shape).
Adds TestMailTriageCustomFolderResolvesOnceAcrossListPages: a custom-folder
filter forced across two messages-list pages, with a non-reusable folders
list stub so any second folders API call fails the test. Updated the two
existing buildListParams alias tests to run resolveListFilter first, mirroring
the real DryRun/Execute call order.
sprint: S1
Co-authored-by: xukuncx <283114605+xukuncx@users.noreply.github.com>
Support content_v2 post message conversion in IM shortcuts so newer post payloads render with the expected markdown, mention, and image formats while preserving fallback compatibility with legacy content.
Previously, im.message.receive events with message_type: interactive surfaced the raw JSON
payload as content, requiring callers to manually parse the card schema. This PR introduces a
user_dsl renderer (ConvertInteractiveEventContent) that converts interactive card content into
structured human-readable text — consistent with how text, post, image, and other message
types are already handled.
The output format is <card title="..." subtitle="...">...</card>, with each card element type
serialised to a readable representation (markdown body, button links, table rows, chart summaries,
etc.).
* feat(okr): add +batch-create, +reorder, +weight shortcuts
Add three new OKR shortcuts for managing objectives and key results:
- +batch-create: Bulk create objectives with key results, with automatic
rollback on failure
- +reorder: Adjust position of objectives or key results within a cycle/objective
- +weight: Adjust weights of objectives or key results with automatic
normalization using fixed-point arithmetic to avoid float precision issues
Key implementation details:
- API paths use underscore separators (/objectives_position, /objectives_weight)
- Weight normalization uses json.Number for precise JSON serialization
- Items are sorted by position before API calls to match backend requirements
- Full unit test coverage and dry-run/live E2E tests
- Skill documentation with usage examples and parameter descriptions
Change-Id: I92b658e0cc42ffa8cbdaec2ec628a079bcfc38f5
* fix: skill simplify & minor fix
Change-Id: I3f27a01cdae2122f26e48ee2acb7f334f2bab7d2
* fix: CR issue
Change-Id: Id9fab84e06f0d67e9f79c1fb9946b6b633200592
* fix: CR issue 2
Change-Id: I6a5e57dd4b10dc79f8681ec614354fbba82abc04
* fix: error handle of +weight shortcut
Change-Id: I6e2a39269e62e3b504e681110843b2ccc315a527
* refactor: retire legacy error envelopes and enforce typed contract
Consolidate all command error reporting onto the typed errs.* contract, remove
the legacy error surface that predated it, and tighten the lint guards so the
contract holds across the whole repository going forward.
Every failure now reaches stderr as one envelope shape: a category, an
optional subtype, a human- and agent-readable message, and a recovery hint,
with invalid parameters listed under `params`. The legacy ExitError envelope,
its constructors, and the boundary bridge that promoted untyped config and
authorization errors are deleted, leaving a single path from error to wire.
Predicate commands keep their silent-exit behavior through a dedicated signal
that carries only an exit code.
Infrastructure paths that still emitted ad-hoc envelopes — flag parsing,
unknown commands and subcommands, plugin and policy guards, confirmation
prompts, and auth/config failures — now classify into the same taxonomy.
Business, API, auth, and config exit codes are preserved; the one behavioral
change is that Cobra usage failures (missing required flag, unknown command,
bad arguments) now emit the typed validation envelope and exit 2, matching the
explicit flag and subcommand guards, instead of Cobra's plain-text exit 1.
Enforcement is repo-wide rather than per-path:
- The errscontract guards run by default everywhere instead of through a
migration allowlist, so legacy envelopes cannot be reintroduced anywhere.
- errorlint runs across the whole repository: every error wrap must use %w and
every comparison must use errors.Is/errors.As, so interior wraps stay legal
but can no longer break the chain the typed boundary relies on.
- The errs-no-bare-wrap guard is keyed by structural prefix instead of an
explicit per-domain allowlist, so new shortcut domains are covered without
editing a list. It runs where forbidigo is enabled (the shortcut domains and
the auth/config/service command groups); repo-wide chain integrity for the
remaining command paths is carried by errorlint above.
* test: align cli_e2e success assertions to the ok envelope
The api and service success path now emits the {"ok":true} envelope, so the
cli_e2e workflow assertions that still expected the old {"code":0} shape via
AssertStdoutStatus(t, 0) fail once they run with live credentials. Switch those
workflow assertions to AssertStdoutStatus(t, true); the fake-payload helper test
in core_test.go keeps its code-shape assertion.
When --name is omitted, remind user that the title defaults to the source
filename and may duplicate content headings, causing visual redundancy.
Ask whether to rename before executing the import.
* docs(mail): clarify message read shortcuts
Update mail read shortcut help, docs, and triage guidance so single-message and multi-message reads are routed to the right commands.
Add focused tests for help text, dry-run copy, triage stderr hints, and batch_get chunking behavior.
sprint: S1
* docs(mail): align batch_get limit with gateway config
* docs(mail): use shell-safe batch message id examples
* docs(mail): trim batch_get pagination wording
* docs(mail): use placeholder style for message ids
* docs(mail): hide batch_get internals from help
The 4 im query commands had three inconsistent sort conventions and leaked upstream API jargon (ByCreateTimeAsc, member_count_desc) directly to users. This PR unifies them on a single rule — --sort selects a field, --order selects a direction, both from fixed enums — so an agent only ever picks from an enum, never constructs a string. Old flags (--sort-type, --sort-by, and --sort on messages/threads) are kept as hidden silent aliases (no deprecation warning), so existing scripts keep working byte-for-byte.
* feat: add --json flag support to auth subcommands
* feat(auth/logout): add json output support for logout command
* feat(auth/list): add json output support for auth list command
Add note shortcuts for note detail and unified transcript retrieval, route vc note detail parsing through the note domain, and update note/vc/minutes skill guidance for normal versus unified transcript handling.
Includes dry-run E2E coverage for the new note shortcuts and documents the remaining live E2E fixture gap.
The cherry-picked riscv64 commit derived RACE_FLAG from `go env GOARCH`
via a grep pipeline, which ignores a GOARCH passed on the make command
line (e.g. `make GOARCH=riscv64 unit-test`) since command-line make
variables are not visible to $(shell ...). Switch to a make-native
filter that honors both, and restore the trailing newline the same
commit dropped.
Add a server-side --chat-modes filter to the im +chat-search shortcut so
users can restrict results to regular groups and/or topic groups.
Change-Id: Ia59c2c05fb2e8e45bd741c8531ca0e3ca69de2f3
Framework support for resource-scoped event subscriptions, so one
EventKey can fan out into independent per-resource subscription scopes:
- KeyDefinition gains SubscriptionKey / NormalizeParams / Match hooks
- ComputeSubscriptionID derives a dedup identity from (EventKey, sub-key
params); plumbed through bus Hub, consume loop, and the
Hello / PreShutdownCheck / ConsumerInfo protocol messages
- add a synchronous Match filter stage before Process
- change PreConsume cleanup to func() error and surface cleanup
(unsubscribe) failures as WARN with an idempotency note
- adapt minutes/vc/whiteboard PreConsume to the new cleanup signature
- render SubscriptionID / SubscriptionKey in event status & schema output
No domain wires these hooks yet; covered by unit tests using bus/protocol
doubles. (Mail, the original exerciser, is intentionally not included.)
Change-Id: Ifc743f1aa0bc4dff0c8a1e35da24883694fe7699
Add SKILL.md entries for the group manager and group moderation
(speaking-permission) API-meta resources:
- chat.managers.add_managers / delete_managers (指定/删除群管理员)
- chat.moderation.get / update (查询/更新群发言权限)
Emit structured validation, API, network, file, and internal error envelopes for Wiki shortcuts so users and agents can recover from failed wiki workflows using stable type, subtype, param, and code fields.
Add Wiki domain errscontract and golangci guards to prevent legacy envelope and common helper regressions.
* feat: add fullstack app-type and --message to apps +create (#1)
* feat: accept fullstack app-type and require --message for it
* feat: inject message into fullstack create request body
* refactor: align fullstack message injection with existing body-build style
* docs: document fullstack app-type and --message for apps +create
* docs: keep scene numbering consistent in lark-apps-create reference
* docs: add HTML/fullstack intent routing to lark-apps SKILL.md
* docs: cover fullstack in lark-apps skill description and clarify HTML flow step
* test: assert fullstack in allow-list error and reject wrong-cased fullstack
* feat: drop --message from apps +create (#4)
* feat: drop --message from apps +create
* docs: drop --message and document agent-generated name/description for apps +create
* feat: add apps local key-value file storage (#5)
* feat: add Miaoda app git credential support (#9)
* fix: remove APIError detail field dependency
* docs(apps): expand lark-apps skill for local-dev & cloud-chat workflows (#3)
Reframe lark-apps from an HTML-publish skill into a full Miaoda app dev
tool covering three paths: local fullstack dev, HTML hosting, and cloud
session dev. Builds on the fullstack create change already on this branch.
- SKILL.md: 3-path routing table; mental models (code via native git,
develop/main branch model, DB via +db-* through Miaoda, env auto-pulled
by `npm dev run`, auto-managed credentials); command index for the new
verbs; ambiguous-input fallback (infer app type from need, ask local vs
cloud instead of assuming; default HTML when no signal)
- add local-dev and cloud-dev playbooks
- create: keep HTML/fullstack + required --message; add local/cloud scene
routing and --enable-multi-env-db
- list: usable by agents with --filter; app_id resolution order
(user-provided / .spark/meta.json / +list --filter)
Co-authored-by: wangjiangwen-gif <286006750+wangjiangwen-gif@users.noreply.github.com>
Co-authored-by: raistlin042 <lvxinsheng@bytedance.com>
* feat(apps): add 4 db CLI commands (table-list / table-schema / sql / dev-init)
妙搭 data CLI 4 条命令,复用存量 OpenAPI URL + 1 个新增 dev-init:
- +db-table-list → GET /apps/{id}/tables(游标分页,AppTable 含预估行数/占用空间)
- +db-table-schema → GET /apps/{id}/tables/{name}(默认结构化 schema;--format pretty 出建表 DDL)
- +db-sql → POST /apps/{id}/sql_commands(?transactional=false DBA 模式)
- +db-dev-init → POST /apps/{id}/db_dev_init(单库→online/dev,不可逆,high-risk-write)
要点:
- sql result 兼容两种 wire 形态(结构化 [{sql_type,data,record_count}] 与 legacy ["rows-json"])
- 多语句失败:server 返 code:0 + ERROR 哨兵,CLI 升级成 typed api_error(exit 非 0),
detail 带 statement_index/completed/rolled_back,防止 agent 误判 ok:true 假成功
- pretty 渲染对齐 miaoda:列间两空格、CJK 双宽、size 友好格式(KB/MB/GB)
- 单测 + e2e dry-run 全覆盖;BOE 真机 e2e 验证通过(25 PASS)
- SKILL.md 注册 4 条命令 + 4 篇 reference
注:内含的 BOE 联调专用 env 覆盖(LARK_CLI_OPEN_API_BASE / LARK_CLI_X_TT_ENV,
internal/cmdutil + internal/envvars)未包含在本次提交,仅本地联调用。
Change-Id: I0fe4458086708a93941e2dee852fa6a10b53bd4a
* docs(lark-apps): db 能力补进 SKILL.md description 的 WHEN 段
按 skill 质量规范(description 三段式 WHAT+WHEN+NOT,加载前唯一可见信息),
原 WHEN 仅"连数据库调试"含糊覆盖 db。补成「查看或操作应用数据库(看表结构 /
跑 SQL / 初始化 dev 环境)」,让 +db-table-schema / +db-sql / +db-dev-init
类查询能精确触发,净增 ~12 字无膨胀。
Change-Id: Id52819fa7d6b8ed0c1f174bf5946d55da7b893d7
* Feat/apps env pull (#11)
* feat: add apps env-pull shortcut
* fix: support array env_vars response in apps env-pull
* fix(apps): improve env-pull merge and expiry output
* feat: add keyword/scope/app-type query to apps +list and unhide it (#8)
* feat: switch apps +create --app-type enum to lowercase html/full_stack
* feat: add keyword/scope/app-type query to apps +list and unhide it
* docs: document apps +list query params and lowercase app_type enum
* test: update apps cli_e2e dry-run tests for lowercase app_type and +list filters
* docs: trim redundant app_type case-sensitivity note in create skill
* docs: single-source apps +list usage contract to SKILL.md
* feat: add apps publish shortcuts (publish/status/history/error-log) (#12)
* feat: add apps publish shared guard and NodeStatus mapping
* test: cover json.Number path in injectStatusName
* feat: add apps +publish shortcut
Implements the `apps +publish` command with dry-run preview (upstream
PSM path shown) and an Execute gated by ensurePublishWired() per the
not-yet-deployed OpenAPI gateway constraint (publishAPIWired=false).
* refactor: make apps publish path placeholders var to satisfy go vet
Declare the four publishXxxPath constants as var instead of const so
go vet's printf analyzer skips them while they are empty placeholders.
Revert the Execute path-build in apps_publish.go from strings.Replace
back to fmt.Sprintf (now safe because the format string is a var).
* feat: add apps +publish-history shortcut
* feat: add apps +publish-status shortcut
* feat: add apps +publish-error-log shortcut
* feat: register apps publish shortcuts
Add AppsPublish, AppsPublishHistory, AppsPublishStatus, AppsPublishErrorLog
to Shortcuts() and update count test from 6 → 10.
* docs: add skill references for apps publish shortcuts
* docs: surface apps publish shortcuts in lark-apps SKILL.md
* docs: clarify publish instance id is not an approval instance
* docs: nudge agent to run apps +publish --dry-run for release requests
* feat: update apps publish shortcuts to v1.0.381 release protocol
Rename concept instance→release across all 4 publish shortcuts and their
tests: NodeStatus→ReleaseStatus enum, --instance-id→--release-id flag,
pipelineTaskID→releaseID response field, errorJobs→errorLogs, and
upstream HTTP path consts→RPC method name consts (PSM lark.apaas.devops
v1.0.381). Dry-run now shows psm+rpc_method instead of an HTTP path.
* docs: update apps publish skill docs to v1.0.381 release protocol
* fix: soften apps publish unavailable hint to user-facing language
* feat: update apps publish to v1.0.385 string status + --status filter
- Remove obsolete int-enum machinery (releaseStatusName/toInt/injectStatusName)
and their encoding/json + fmt imports from apps_publish_common.go
- +publish Execute now returns status string alongside release_id
- +publish-history gains --status Enum flag (publishing/finished/failed);
buildHistoryBody gains status param, table column status_name→status
- +publish-status Execute drops injectStatusName, pretty prints out["status"]
- +publish-error-log shapeErrorLog is string passthrough (no status_name)
- Unit tests updated: delete 3 obsolete common tests, update history/error-log
* docs: update apps publish docs to v1.0.385 string status + --status filter
* feat: wire apps publish shortcuts to final gateway paths (guard stays until deploy)
Replace RPC-name placeholders with real OpenAPI paths (publishCreate/Get/ErrorLog/ListPath consts). Switch DryRun to idiomatic HTTP form (POST/GET + real URL + body/params). Fix body/query placement: publish body has no app_id (path-only); history switches from POST body to GET query with snake page_token. Fix Execute response reads to snake_case fields (release_id, created_at, updated_at, error_logs). publishAPIWired stays false; 1-line flip activates live calls.
* docs: update apps publish docs to final gateway paths
Replace RPC/PSM dry-run example with real HTTP form (POST/GET /open-apis/spark/v1/apps/:app_id/releases[/:release_id[/error_logs]]).
Fix all response field names to snake_case (release_id, created_at, updated_at, error_log).
Note --status/--limit/--page-token as HTTP query params in publish-history.
* feat: enable apps publish gateway calls (remove not-deployed guard)
* docs: remove not-deployed transition notes from apps publish docs
* feat: use spark:app:publish scope for apps +publish
* feat(apps): add +init shortcut to initialize Miaoda app repo (#6)
* feat(apps): add command runner and credential redaction for +init
* fix(apps): make credential redaction scheme matching case-insensitive
* feat(apps): add +init shortcut declaration, validation, and dry-run
* feat(apps): implement +init orchestration (credential-init, clone, checkout, conditional push)
* fix(apps): redact full userinfo when repo URL contains literal @
* docs(apps): add +init skill reference
* fix(apps): declare explicit empty Scopes on +init shortcut
* fix(apps): consume repository_url from +git-credential-init in +init
* feat(apps): add +init template flag and absolute-path dir resolution
* refactor(apps): use shared charcheck for +init --dir validation
* feat(apps): add meta.json, steering, and empty-repo helpers for +init
* feat(apps): add +init npx scaffold orchestration (init/upgrade branches)
* feat(apps): wire +init scaffold, already-initialized short-circuit, npx dep check
* docs(apps): document +init npx scaffold, --template, --dir, already-initialized
* docs(apps): correct stale +git-credential-init unreleased note in +init ref
* fix(apps): reject all control chars in +init --dir
* feat(apps): add +init progress logging and optional --template resolver
* refactor(apps): inline constant in +init scaffold progress log
* docs(apps): document +init optional --template and stderr progress contract
* feat(apps): treat README-only repo as empty and commit with --no-verify in +init
* docs(apps): explain README-seed match and --no-verify rationale in +init
* docs(apps): document README-seed empty detection and commit --no-verify
* feat(apps): add session conversation lifecycle shortcuts (#13)
* feat(apps): add +session-create shortcut
* fix(apps): remove unused sessionPath helper, assert empty +session-create body
* feat(apps): add +session-list shortcut
* feat(apps): add +session-read shortcut
* feat(apps): add +session-stop shortcut
* feat(apps): add +chat shortcut
* feat(apps): register session lifecycle shortcuts
* docs(apps): add session conversation skill reference
* docs(apps): clarify fullstack session_id source and fallback
* style(apps): gofmt apps_session_create.go
* docs(apps): add conversation/session triggers to skill routing description
* docs(apps): add conversation flow guidance (when to reuse vs new session, per-step user prompts)
* docs(apps): slim session reference per skill quality standard (4047->1726 tok)
* docs(apps): tighten session additions in SKILL.md (4394->4145 tok)
* fix(apps): align +chat with v7.8 contract (async, no turn_id in response)
* fix(apps): update +chat path to .../sessions/{id}/chat (backend endpoint change)
* docs(apps): align SKILL.md session command shape with v7.8 contract
* style(apps): gofmt apps_db_table_schema_dryrun_test.go
Go 1.19+ gofmt 文档注释列表缩进新规则(普通缩进 → tab 对齐),
修复 fast-gate CI 的 gofmt 卡点。
Change-Id: Ic246a659e016d9d6216182199ef300ae6f00ef9d
* feat(apps): split +init commit, plainer wording, align skill branches (#14)
* refactor(apps): plainer +init progress/help wording, keep scaffold key
* refactor(apps): add porcelain change classifier for +init commit split
* feat(apps): split +init empty-repo commit into code + config, reword subjects
* refactor(apps): scaffold-kind constants and pathspec assertions for +init split
* docs(apps): use +init in Path A; align app-repo branch to sprint/default
* docs(apps): align local-dev playbook to sprint/default + origin remote
* docs(apps): document +init two-commit split and plainer init wording
* docs(apps): require asking clone dir before +init, no assumed path
* fix(apps): stage +init commits by exact paths to avoid gitignore error
* refactor(apps): lowercase miaoda in +init commit subjects
* test(apps): cover +init upgrade path with real git
* fix: harden app git credential handling (#16)
* fix: harden git credential refresh fallback (#18)
* fix(apps): validate env-pull key names before writing to .env.local (#17)
* fix(apps): validate env-pull key names before writing to .env.local
S2 (medium-low) from security review: env-pull wrote server-returned
env KEYs to .env.local without validation. A compromised or MITM'd
backend could inject arbitrary lines via keys containing newlines.
- Add envKeyPattern regex to validate keys match [A-Za-z_][A-Za-z0-9_]*
- extractEnvPullVars now returns skippedKeys for invalid key names
- Invalid keys are skipped (not hard-fail) so remaining valid keys
are still pulled
- writeEnvPullPretty prints a warning listing skipped keys
* fix(skills): correct npm script syntax from 'npm dev run' to 'npm run dev'
* fix(skills): align env-pull guidance with implementation
🤖 Generated with [Aiden x Claude Code]
* test(apps): cover storage/git-credential error paths and fix tz-flaky env-pull tests (#19)
The coverage and unit-test CI jobs failed on two timezone-dependent
assertions in apps_env_pull_test.go: the code renders the database
expiry via time.Local() while the tests hard-coded a CST literal, so
they failed under CI's UTC. Compute the expected string from the same
timestamp with Local() instead, making the assertions timezone-agnostic.
Also add unit tests for the error branches codecov flagged as uncovered,
taking storage.go and git_credential.go to 100%:
- storage Read/Write/Delete/List filesystem-error paths
- +git-credential-remove ConfigWarning output (pretty and JSON)
- gitCredentialLocalError nil passthrough
* fix(apps): silence +init forbidigo, npx app sync -y --prefer-online (#20)
* fix(apps): add Subtype to env-pull error literals (#21)
typed_error_completeness lint requires all errs.XxxError literals to
set Problem.Subtype. Add the missing field to 11 error constructions:
- ValidationError (user input checks): SubtypeInvalidArgument
- ValidationError (API response parsing): SubtypeInvalidResponse
- InternalError (filesystem ops): SubtypeUnknown
* feat(apps): inject FORCE_DB_BRANCH=dev in env-pull output (#23)
* feat(apps): inject FORCE_DB_BRANCH=dev in env-pull output
Always write FORCE_DB_BRANCH="dev" into the resolved .env.local after
extracting upstream env_vars, so downstream tooling pinning the dev
database branch does not need a separate manual edit. Existing local
values are overwritten in place via the canonical merge path.
* docs(skills): document apps +env-pull in lark-apps skill
Add the env-pull entry to the lark-apps SKILL index and ship the
matching reference doc covering args, merge semantics, return shape,
error envelope subtypes, and dry-run behavior so AI agents can route
to it without reading the Go source.
* feat(apps): surface is_published and online_url in +list pretty view (#22)
* docs: refactor lark-apps skill per quality spec (#24)
Slim SKILL.md and references against the lark-cli skill quality spec
while preserving domain knowledge and safety guardrails.
- Compress SKILL.md (drop the MUST-read prelude, full command-index
tables, and content already owned by lark-shared: auth, scope,
exit-10, risk policy, _notice); add version field; zero CRITICAL
markers.
- Defer flag enumeration in references to `--help`; convert
narration-inducing prohibitions into positive defaults; de-duplicate
the per-file error.hint relay into a single resident SKILL.md rule.
- Fix stale facts found against shortcuts/apps source: drop the
non-existent +create --message and --enable-multi-env-db flags,
+list --filter (now --keyword), +db-multi-env-init (now
+db-dev-init), and the removed html-publish cwd hard-reject.
- Keep all safety guardrails: db-dev-init irreversibility/exit-10,
db-sql non-transactional multi-statement, git-credential token
handling, html-publish credential scan, access-scope confirmation.
- Restore intent lost during slimming: release_id is not an approval
instance (do not route to lark-approval); resolve access-scope
targets via contact/im; ask the user before publishing as a
side-effect; distinguish developing an existing app locally
(+init) from creating a new one (+create).
* test(apps): supplement shortcuts/apps unit-test coverage to 88% (#25)
* test(apps): cover db-table-list numeric/byte formatting helpers
* test(apps): cover db-sql cell/code/dml/error render helpers
* test(apps): cover env-pull newline/expiry/extract-vars helpers
* test(apps): cover db-sql render branches and env-pull expiry edge case
* test(apps): cover init empty-dir/meta/ls-files error branches
* test(apps): cover env-pull target/read/parent-dir error branches
* test(apps): cover stage-and-commit and commit-push error branches
* test(apps): cover access-scope target split and JSON validation
* test(apps): cover html-publish decode error and scaffold sync failure
* test(apps): cover apps-update body field combinations
* test(apps): cover access-scope body build branches
* feat(apps): pass --local to npx skills sync in +init (#26)
* feat(apps): pass --local to all npx miaoda-cli calls in +init
* feat(apps): pass --local only to npx skills sync in +init
* docs(apps): surface +publish and +init dir-choice in local-dev flow (#27)
* docs(apps): surface +publish as deploy action in skill routing
* docs(apps): add explicit deploy-after-local-edit section to local-dev
* docs(apps): promote +init dir-choice instruction to a domain rule
* docs(apps): make dev-method a signal-driven entry gate before routing (#28)
* docs(apps): restore three-path overview line in apps skill intro (#29)
* feat(apps): add executable Examples to shortcut --help and error hints (#30)
* test(apps): guard every shortcut has a help Example and no PII
* feat(apps): add help Examples to all 24 apps shortcuts
* feat(apps): add actionable hints to high-impact error paths
* test(apps): cover withAppsHint set-if-empty hint behavior
* feat(apps): use concrete enum value in access-scope-set Example
* docs(apps): clarify db-sql/db-table-list json default output behavior
两处仅补充注释,不改逻辑:
- +db-sql: data.results 在 json 默认路径原样透出全部行,CLI 不二次截断;
server 对单条 SELECT 有 1000 行硬上限、超出直接返报错,非无界 token 黑洞。
- +db-table-list: json 默认透出含每表完整 columns[] 系产品设计(list 接口本就
返回列定义,json 消费方一次拿全量、免逐表再调 +db-table-schema),pretty 仅摘计数。
Change-Id: I1a49de8defc4428bfe1e774e4fd7adb45e59e3af
* feat(apps): command-layer AI-friendliness governance (P0+P1) (#32)
* fix(apps): normalize --app-type case to align with server
* refactor(apps): migrate CallAPI to CallAPITyped for typed errors and retryable
* feat(apps): trim icon_url and created_at from +list default output
* feat(apps): add actionable hints to high-impact error paths
* feat(apps): add 2-3 help Examples to +chat and +access-scope-set
* docs(apps): add --jq filter tips to list/db commands
* docs(apps): sync +list reference with trimmed output fields
* test(apps): assert error hints and messages carry no secrets or PII
* fix(apps): prefix --jq tips with .data. so they run against the response envelope
* test(apps): expect --app-type uppercase normalization in create dry-run E2E (#33)
* fix(apps): scaffold via @latest miaoda-cli instead of @alpha (#34)
* feat(apps): rework lark-apps triggering, routing & confirm policy (#35)
* feat(apps): results-oriented triggering, pre-auth floors, terminal URL
Widen description WHEN to cover app-building openers (CRM/审批/HTML page)
with no Miaoda signal word, WHAT still anchored to 妙搭应用开发与托管.
Add a pre-authorization rule (auth words skip confirm) with two non-exempt
floors: destructive DDL (DROP/TRUNCATE/ALTER drop|modify column) dry-run,
and first public-URL publish (+publish/+html-publish) when no auth word.
Exempt html app_type from the local-vs-cloud dev-method gate, and scope
that gate to new-app creation only (existing-app ops route directly).
Require an accessible URL as the end-to-end terminal step.
* feat(apps): apply eval-fix behavior contracts across reference docs
init/local-dev: end-to-end default-directory escape hatch; end-to-end
new-build starts with +create. db-sql: additive DDL direct-exec when
authorized, destructive DDL stays dry-run. local-dev/publish-status:
return online_url via +list as the full_stack publish terminal step.
cloud-dev: generation != shareable URL, +publish handoff, background
until-poll snippet (sleep N && cmd intercepted; deprecate ScheduleWakeup),
multi-turn publish precondition. publish/publish-error-log: transient
failure (EAI_AGAIN/ETIMEDOUT/registry) discrimination, retry cap 2,
honest receipt. env-pull: first-launch fallback. local-dev/db-dev-init:
new full_stack ships dual DB, skip +db-dev-init.
* refactor(apps): apply review feedback — semantic criteria, drop overfit/unverified content
Per line-by-line review of the eval-fix changes:
- Entry routing reframed to objective/semantic criteria (new-vs-existing =
'can an existing app be identified'; dev-method = who-writes-code
preference), replacing keyword/example matching.
- db-sql DDL gate restated by effect (data-loss / reversibility), not a
keyword list.
- Pre-authorization judged by expressed intent (not a word list); single
non-exempt floor (destructive/irreversible DB dry-run); confirm policy in
its own section, error.hint in 'failure handling'.
- init.md slimmed to command facts (directory choice owned by local-dev,
no init<->local-dev cycle); local-dev defers new-vs-existing to the entry.
- Reverted unverified/redundant/runtime-coupled additions: cloud-dev
session-read preview-URL claim + background-poll snippet + queued_count
precondition; publish transient-retry/ScheduleWakeup; env-pull first-launch;
db-dev-init positive restatement; SKILL terminal-URL mandate.
- Fixed dangling section references after the rename.
* fix(apps): scope pre-authorization to hands-off intent, not 'wants a result' (#36)
Follow-up to #35. The merged pre-authorization rule treated 'wanting the
final result' as authorization, so '先在本地跑起来让我看看' was read as
pre-authorized and the agent silently picked a clone directory without
asking. Re-state the criterion as the user's hands-off intent (explicit
waiver, or an end-to-end directive), judged uniformly across the flow
(directory/clone, publish) — not a per-decision carve-out. Merely wanting
a result or asking to review is not authorization.
* docs: clarify apps cloud dev publish state
* fix(apps): require commit+push before publish, clarify deploy flow (#38)
* fix(apps): require committing changes before publish in local-dev flow
* fix(apps): make commit+push mandatory before publish in agent rules
* fix(apps): scope selective-add caveat to incremental deploy, not new-app flow
* fix(apps): make pre-publish commit conditional on local changes
* fix(apps): tighten pre-publish commit wording in agent rules
* fix(apps): cloud-dev does not auto-deploy, add explicit publish step
* docs(apps): document +chat init vs incremental turn cost (#39)
First +chat on a not-initialized app runs full design+gen server-side
(~20-50 min); chat on an already-initialized app is incremental and
finishes in minutes. Surface this in the +chat Go comment as a pointer
and put the init-state check + matching polling cadence (5-10s vs
60-120s) in the lark-apps cloud-dev skill reference as the canonical
source. Cloud-side init check uses +session-read committed-version
info or +list is_published:true.
* docs(apps): document +chat init vs incremental turn cost (#40)
First +chat on a not-initialized app runs full design+gen server-side
(~20-50 min); chat on an already-initialized app is incremental and
finishes in minutes. Surface this in the +chat Go comment as a pointer
and put the init-state check + matching polling cadence (5-10s vs
60-120s) in the lark-apps cloud-dev skill reference as the canonical
source. Cloud-side init check uses +session-read committed-version
info or +list is_published:true.
* feat(apps): surface online_url/error_logs in +publish-status output (#41)
* refactor(apps): extract shared release error-log table helper
* fix(apps): keep error-log table byte-identical for null error_logs
* feat(apps): surface online_url/error_logs in +publish-status output
* docs(apps): read online_url/error_logs from +publish-status in publish flow
* docs(apps): align local/cloud dev publish flow with +publish-status fields
* refactor(apps): rename +db-dev-init→+db-env-create, trim db-table-list columns
- +db-env-create(原 +db-dev-init):新增 --env 参数(调用方传入,目前只支持 dev),
--sync-data 改为 true/false 取值;服务端 URL 仍走 db_dev_init。
- +db-table-list:json 默认用白名单投影(dbTableListItem)只输出产品要求字段,
每表 columns[] 折算成 column_count、不再透出完整列定义(与 +db-table-schema 重复且放大
token);要完整列定义/索引/约束用 +db-table-schema。
- 同步对齐 db 相关 skill 文档(命令名、column_count、env-create 参数)。
- 单测 + cli_e2e dry-run 全绿。
Change-Id: I116ab11807679f8f06ed18221f705bab426d015c
* refactor(apps): rename +db-table-schema → +db-table-get
动词对齐 +db-table-list(list/get)。仅命令名 + 标识符 + 文档改名,行为/输出/URL 不变:
- AppsDBTableSchema→AppsDBTableGet,文件/测试/cli_e2e test 重命名
- buildDBTableSchemaParams→buildDBTableGetParams
- +db-sql / +db-table-list 里的交叉引用 hint、skill 文档同步
Change-Id: I36dfb8fd0d2613492a57dc7815bc58414c145480
* feat: auto-pull env vars after apps +init (#42)
* test: route apps +env-pull to its own fake-runner key
* feat(apps): add +env-pull envelope parsers for +init
* feat(apps): add pullEnv helper invoking sibling +env-pull
* feat(apps): +init auto-runs +env-pull after push (non-fatal)
* docs(apps): clarify db-sql --query @path is relative-only, use stdin for absolute paths
@path 受 lark-cli 全局文件安全策略约束,只接受 cwd 内相对路径;绝对路径 / cwd 不固定
场景改用 stdin(--query - < /abs/file.sql),无需先 cd。
Change-Id: Ib3453810cfc9303d72b4facf3493ad9688eeffd3
* docs(apps): refine db-sql --query path guidance wording
以 agent 视角重写:@ 仅接受工作目录内相对路径,绝对路径/越界路径被拒(CLI 文件访问统一约束);
工作目录外的文件经 stdin 传入。
Change-Id: Ic7db00934b3571368eb704451f4ce1776463806d
* feat(apps): make +db-sql high-risk-write (require --yes)
+db-sql 可含 DML/DDL,统一升级为 high-risk-write:框架对所有执行强制 --yes 确认关卡
(--dry-run 预览豁免),无 --yes 返 confirmation_required / exit 10。
- Risk: write → high-risk-write(去掉自定义门禁,直接用框架机制)
- skill 文档:命令骨架标注 --yes 要求;Agent 规则改为「执行需 --yes,只读可直接带、
破坏性先 dry-run 确认再带」
- 单测所有执行调用补 --yes
Change-Id: I57e78832b35fa170a485774e6fb7289109d678c3
* docs(apps): clarify app_ (Miaoda) vs cli_ (Feishu) app id (#46)
* 优化云端开发skill,明确执行模型,参数解释 (#44)
Co-authored-by: fushengdong.1 <fushengdong.1@bytedance.com>
* refactor: rename apps publish commands to release and session-get (#45)
* refactor(apps): drop +publish-error-log, rename release path constants
* refactor(apps): rename +publish to +release-create
* refactor(apps): rename +publish-history to +release-list, unify pagination to --page-size
* refactor(apps): rename +publish-status to +release-get
Renames apps +publish-status → +release-get (AppsPublishStatus → AppsReleaseGet),
updates --release-id desc to reference +release-create, and fixes the Execute
error hint to point at +release-list instead of +publish-history.
* refactor(apps): rename +session-read to +session-get
* docs(apps): rename publish references to release, +session-read to +session-get
* refactor(apps): clean up residual publish/session-read references
Fix six leftover references missed in Tasks 1-6: +publish-history in
jq-tip test wantCmds map and common_test hint fixture (×3), +session-read
in apps_chat.go comment+output string (×2), apps_session_stop.go flag
desc (×1), apps_chat_test.go comment (×1), and +publish-status in
lark-apps-list.md agent rule prose (×1).
* docs(apps): clarify release-get link contract and session-get vs session-list
* docs(apps): generalize release-list page-size rule to N records
* feat(apps): rename +list --scope flag to --ownership (#47)
* feat(apps): rename +list --scope flag to --ownership
* test(apps): update +list cli_e2e dry-run for --ownership rename
* docs(apps): document +list --ownership flag
* feat(apps): align +release commands with new release API format (#48)
* feat(apps): align +release-create scope to spark:app:write
* feat(apps): raise +release-list --page-size documented max to 500
* feat(apps): show commit_id in +release-get pretty output
* docs(apps): update release reference docs for page-size 500 and commit_id
* test(apps): cover empty commit_id in +release-get pretty output
* docs: align lark apps cloud dev release flow
* feat(apps): redesign +db-sql → +db-execute (--sql/--file, default env dev)
按 db 子域命令最终设计重做执行入口:
- 命令 +db-sql → +db-execute(动词收尾,对齐 +db-table-list/-get)
- --query 拆为 --sql(内联/stdin)与 --file(.sql 文件路径),二选一互斥;
--file 在 Validate 阶段读出归一化到 --sql
- 默认 --env online → dev(打生产库需显式 --env online)
- 文件/标识符/注册/测试/cli_e2e/skill 文档全部对齐重命名
- 新增测试:--sql/--file 互斥、--file 读取、默认 env=dev
不在本次范围:--transaction/--no-transaction(服务端 transactional 实为路径切换、
非真事务,需 dataloom 侧先支持真事务开关)、--max-rows/--timeout 等后续项。
Change-Id: I50c06faf83527471446e2a6651ccb51f6eedd6ff
* docs(apps): clearer --env online wording for +db-execute
把口语化的「打生产库需显式」改为「需要操作线上环境数据库时,显式指定 --env online」;
flag desc 同步去掉 hit production 措辞。
Change-Id: Iee82fccf17e08bddb4b760c3970a416746b10c4c
* docs(apps): drop 'ad-hoc' jargon from +db-execute description
中文文档/英文 description 去掉术语 ad-hoc;SELECT/DML/DDL 已表意,含义不丢。
Change-Id: Ie2cccc5fc3491fe5f57190a87b93ecd70405b156
* docs(apps): trim +db-execute when-to-use and --file path wording
- 何时用去掉「(查询 / 临时数据修复 / 应急 DDL)」枚举
- --file 路径说明去掉 .. /符号链接/统一约束 的技术化描述,改为「相对路径,
否则用 --sql - < 文件路径」的产品化口吻
Change-Id: Ie70e57895c78650230b6942b03d90a2d95c937f2
* docs(apps): note --file rejects absolute/cwd-escaping paths
简短补回 --file 的路径约束(绝对路径 / 经 ..、符号链接越界会被拒),去掉冗余评注。
Change-Id: I549893c82cafbe97529e08dcbc3ee5496927da18
* fix(apps): replace t.Chdir with os.Chdir in db-execute test (Go 1.23 compat)
t.Chdir 是 Go 1.24 API,但 go.mod 为 go 1.23.0,CI(Go 1.23)报
"t.Chdir undefined"。改用 os.Chdir + t.Cleanup 还原,1.23 兼容。
Change-Id: I550611773e5088275be1c4344d4f8269610ce74a
* feat(apps): refine +init description and refresh env on re-init
* fix(apps): treat accessible-link requests as publish intent (#53)
* refactor(apps): +db-env-create --sync-data string-enum → Type:bool
原实现用 string + Enum["true","false"] + == "true" 模拟 bool,啰嗦且非惯用。
改为 Type:bool(rctx.Bool):传 --sync-data 即开启、省略为 false。
同步更新测试、cli_e2e dry-run、skill 文档。
Change-Id: I3068e0577fa20a7cbaf414ca9af3d197f6ae8049
* fix(apps): declare --app-type as strict lowercase enum (#55)
* docs(apps): front-load routing, dedupe, and trim lark-apps skill (#56)
* docs(apps): front-load intent-routing table and dedupe skill body
* docs(apps): dedupe publish guardrail and polling rules in cloud-dev
* docs(apps): trim env-pull implementation detail to behavior contract
* docs(apps): add +env-pull routing entry in SKILL.md
* docs(apps): fix create.md cross-ref to actual SKILL.md section name
* feat(apps): add error.hint to command failures and a consistency gate (#57)
* feat(apps): add appIDListHint const and wrap 4 pure app-id command failure paths
Adds shared `appIDListHint` recovery hint to common.go and wraps the
CallAPITyped failure branch of session-create, session-list, update, and
release-list to surface an actionable next-step hint on 4xx errors.
Includes httpmock unit tests in apps_hints_more_test.go (TDD: red→green).
* feat(apps): add sessionStopHint and createHint for session-stop and create commands
Adds per-command recovery hints with specific guidance: sessionStopHint
points at +session-list and +session-get; createHint explains valid
--app-type values and permission failure. Wraps the CallAPITyped failure
branch in both commands.
* feat(apps): add recovery hints for db-env-create, db-table-get, db-table-list
Adds dbEnvCreateHint, dbTableGetHint, and dbTableListHint with actionable
cross-command guidance (e.g. pointing at +db-table-list for env conflicts,
+db-env-create for missing dev env). Wraps only the CallAPITyped failure
branch; requireAppID validation errors are left untouched.
* refactor(apps): make session-stop hint runnable and align hint test names
* test(apps): guard withAppsHint upstream-wins contract and new hint leak safety
* test(apps): add help-skill command consistency gate
---------
Co-authored-by: linchao5102 <linchao.5102@bytedance.com>
Co-authored-by: Wang <wangjiangwen@bytedance.com>
Co-authored-by: wangjiangwen-gif <286006750+wangjiangwen-gif@users.noreply.github.com>
Co-authored-by: 陈兴炀 <chenxingyang.1019@bytedance.com>
Co-authored-by: aihao-git <aihao.0331@bytedance.com>
Co-authored-by: bali <bali@bytedance.com>
Co-authored-by: hunnnnngry <chenxi.xichen@bytedance.com>
Co-authored-by: shengdongyc <1135978761fsd@gmail.com>
Co-authored-by: fushengdong.1 <fushengdong.1@bytedance.com>
Block 1 — field completion: audio renders <audio key="..." duration="Xs"/>
(falls back to [Voice: Xs]/[Voice]); post renders emotion -> :emoji_type:,
applies text.style (bold/italic/underline/lineThrough), passes through md;
sticker unchanged.
Block 2 — opt-in --download-resources (default off) on +chat-messages-list,
+messages-mget, +threads-messages-list: extract downloadable resource refs
during formatting (image/file/audio/video/media + post-embedded; sticker
excluded; merge_forward sub-items carry the top-level container message_id,
since the resources endpoint rejects sub-item ids with "234003 File not in
msg" and can only fetch a forwarded resource through the container; thread
replies get their own block), then download each distinct (message_id,
file_key) once into ./lark-im-resources/ with bounded concurrency (3), filling
back local_path/size_bytes; single-resource failures are isolated (error:true +
stderr warning). Path safety reuses normalizeDownloadOutputPath +
ResolveSavePath.
Batch download keys each file on disk by its unique file_key basename and only
appends an extension (from the Content-Disposition filename or MIME type) —
it does NOT substitute the server's Content-Disposition filename. Otherwise two
resources whose servers return the same filename (e.g. download.bin) would
resolve to the same ./lark-im-resources/ path and clobber each other
concurrently. The friendly "adopt the server filename" behavior is kept only
for an explicit +messages-resources-download with no --output.
Resource ref extraction guards against self-referential / cyclic merge_forward
prefetch maps (a real API sub-item list can include the container's own id or a
back-pointing merge_forward) via a visited set, so extraction terminates instead
of overflowing the stack. The container message_id is threaded through nested
merge_forwards as the download owner.
Also: document the feature (including the im:message:readonly scope requirement)
in skills/lark-im — SKILL.md is generated from skill-template/domains/im.md
(edit the source), plus the hand-written message-enrichment + 3 command
references.
Change-Id: I3a71d7d1b193130f551aaa2ec180ac1500d59ac4
Meego: https://meego.larkoffice.com/5e96d7bff4e7c525510f9156/story/detail/7331555925
Emit structured validation, API, network, file, and internal error envelopes for Markdown shortcuts so users and agents can recover from failed markdown workflows using stable type, subtype, param, and code fields.
Add Markdown domain errscontract and golangci guards to prevent legacy envelope and common helper regressions.
Emit structured validation, API, network, file, and internal error envelopes for Slides shortcuts so users and agents can recover from failed presentation workflows using stable type, subtype, param, and code fields.
Add Slides domain errscontract and golangci guards to prevent legacy envelope and common helper regressions.
Emit structured validation, API, network, file, and internal error envelopes for Sheets shortcuts so users and agents can recover from failed spreadsheet workflows using stable type, subtype, param, and code fields.
Add Sheets domain errscontract and golangci guards to prevent legacy envelope and common helper regressions.
Emit structured validation, API, network, file, and internal error envelopes for Doc shortcuts so users and agents can recover from failed document workflows using stable type, subtype, param, and code fields.
Add Doc domain errscontract and golangci guards to prevent legacy envelope and common helper regressions.
+csv-put --csv data.csv (a forgotten @) was silently written as one-cell content, because any string parses as valid CSV — unlike malformed JSON it never errored, so the filename landed in the sheet instead of the file's contents.
+csv-put's Validate now rejects a --csv value when it names a real file in the cwd subtree (guardCSVValueIsNotFilePath; fileIO.Stat, fail-open), hinting to use --csv @file or stdin (--csv -). Scoped to --csv only — no framework or other-flag change. Checking real existence (not name shape) lets inline content that merely ends in a filename pass through. Adds TestGuardCSVValueIsNotFilePath.
`append` always inserts at document end (equiv. `block_insert_after --block-id -1`),
but skill docs previously recommended it for the "skeleton + chapter-by-chapter fill"
pattern, causing all content to pile up after the last heading.
Changes:
- Remove `append` from skeleton workflow guidance in `lark-doc-create-workflow.md`
and `lark-doc-create.md`; recommend `block_insert_after` with explicit `--block-id`
- Fix `block_move_after` required params: remove `--content` (not supported),
only `--block-id` and `--src-block-ids` are valid
- Add bash language tag to code block for proper highlighting
Replace every command-facing error path in the event domain — the
consume/schema command layer, the +subscribe shortcut, EventKey
definitions, and the consume orchestration — with typed errs.*
envelopes, so consumers get stable type, subtype, param, hint, and
missing_scopes metadata for classification and recovery instead of
free-form message text.
- Input validation (--jq, --param, --output-dir, --filter, --route,
unknown EventKey, EventKey params) reports validation /
invalid_argument with the offending flag in param and an actionable
hint.
- Scope preflight reports authorization / missing_scope with the
machine-readable missing_scopes list; console-subscription and
single-bus preconditions report failed_precondition with recovery
hints.
- The consume API boundary passes already-typed errors through and
classifies transport, non-JSON HTTP, and unparsable responses; the
vc note-detail retry now matches the not-found code on typed errors
(it silently never fired against the legacy envelope shape).
- Previously-bare failures exited 1 with a plain-text "Error:" line
and now exit with their category code (validation 2, auth 3,
network 4, internal 5) alongside the typed stderr envelope.
- forbidigo and errscontract guards now cover the event paths so
regressions fail lint; AGENTS.md and the lark-event skill document
the typed contract for agent consumers.
Validation: make unit-test (race) green; event unit and e2e suites
assert category/subtype/param/hint and cause preservation against the
real binary; errscontract and golangci lint clean.
slides +create finished by calling /drive/v1/metas/batch_query just to
fetch the presentation URL. That call needs a drive scope the shortcut
never declares, so it 403'd for users who only authorized slides scopes
(both UserAccessToken re-auth and TenantAccessToken scope-not-opened),
producing a large share of the shortcut's failure telemetry — even though
the presentation itself was already created successfully.
slides creation never otherwise touches drive, so rather than gating a
drive-free operation behind a drive scope, build the URL locally from the
token via common.BuildResourceURL (the same brand-standard-host fallback
already used by drive +upload / wiki +node-create). The URL is now always
returned, no extra scope is required, and creation never blocks.
Tests are updated to match: drop the registerBatchQueryStub helper and its
call sites (the httpmock Verify cleanup was failing on the now-unconsumed
batch_query stubs), point url assertions at the brand-standard host, and
replace TestSlidesCreateURLFetchBestEffort with TestSlidesCreateURLBuiltLocally,
which asserts the url is produced with no drive call registered.
* docs(approval): restructure skill with intent table and scope boundaries
Rewrite the description for intent-based routing (situation framing
instead of method enumeration) and add the lark-task disambiguation.
Replace the bare method list with an intent-to-command table including
topic and add_sign_type enums, document the query-to-operate workflow
chain with a runnable example, and add an out-of-scope section routing
definition creation to the Feishu client/admin console.
Bump version to 1.1.0.
Change-Id: I33b7b13b7855d67f40954701a09b115e3c91176c
* docs(approval): strengthen description coverage of edge actions
Restore the "all processing operations" phrasing so edge actions like
remind route to this skill; weak-model routing evals regressed on the
narrower "query and process" wording (2 misses in 4 runs vs 0 after
the fix).
Change-Id: Ica1928dacf879b6c7a46dfda37e35b1be9391432
* docs(approval): drop misleading 已发起 from tasks query row
tasks query 查的是本人作为审批人的任务;已发起(本人发起的实例)应走
instances initiated,该路径已在下方表行列出。移除 tasks query 的「已发起」
标签与 topic=3 枚举,避免 agent 误用 tasks query topic=3 查已发起。
Land the high-value, low-risk items from the skill-quality audit; SKILL.md only.
- description: drop the '接口通过 XML 协议通信' impl detail; append a 不负责
out-of-scope clause so 'make a deck' / 'draw a diagram' stop mis-routing.
- replace the 权限速查 scope table with a ## 不在本 skill 范围 routing table
(doc / whiteboard / drive / sheets / base).
- reconcile the whiteboard boundary with the in-slide <whiteboard> element
(added on main, #1029): lark-whiteboard owns only standalone whiteboard
objects in cloud docs; flow/architecture diagrams drawn inside a slide stay
in this skill via <whiteboard>. Clarified in description and out-of-scope note.
- defer auth / permissions / global params to lark-shared as single source.
- move native-API resource hint into prose; reword schema reminder; move the
'schema is source of truth' note next to 核心规则.
Deliberately not adopted: moving Design Ideas out of the body, relocating the
wiki-token section, dropping the native-API schema guardrail, and the bulk
lark-slides- reference rename.
Failures from the minutes and video-conference commands now surface as
structured, typed errors carrying a stable category and subtype — spanning
input validation, missing permissions, network and file-I/O failures, and
remote API errors — so callers can branch on the error kind instead of
parsing free-form text. Batch commands report partial failures explicitly,
emitting per-item results with a non-zero exit instead of masking them.
- Add explicit NOT boundaries to the description and a dedicated
"不在本 skill 范围" section: file upload -> lark-drive, content
editing -> lark-doc / lark-sheets / lark-base.
- Move the Shortcuts table up, right after 快速决策, so command entry
points are discoverable first; keep the member-add flow and
target-semantics sections after it.
- Add an inline reminder under the delete-space guidance that a wiki
URL / name is not a space_id and must be resolved via
`wiki spaces get_node` first.
- Remove the duplicated permission (scope) table and the redundant
schema note so auth/permission guidance stays centralized in
lark-shared.
- Bump the skill version to 1.0.1.
- Keep skill-template/domains/wiki.md in sync with the SKILL.md
introduction narrative.
Change-Id: If2b4341f350191ee0a65bf3a2cab9afa2b76d931
lark-cli update currently discovers official skills by parsing unstable human-oriented `skills add --list` output. This prefers the stable official JSON index for skills discovery, while preserving the existing CLI-list fallback and full-install fallback for resilience.
Changes:
- Add official skills index JSON parsing in `internal/skillscheck/sync.go`
- Prefer JSON index discovery before existing CLI list parsing in `internal/skillscheck/sync.go`
- Add reason-chain details when both discovery layers fall back to `fallbackFullInstall`
- Add bounded HTTPS fetch for `https://open.feishu.cn/.well-known/skills/index.json` in `internal/selfupdate/updater.go`
- Add unit tests for parser behavior, discovery fallback order, and fallback detail reasons in `internal/skillscheck/sync_test.go`
Co-authored-by: zhaoyukun.yk <zhaoyukun.yk@bytedance.com>
Task commands now return structured, typed errors instead of the legacy
exit-code envelope: every failure carries a stable category, subtype, and
recovery hint, so callers can branch on the error class instead of parsing
messages. Exit codes derive from the error category — input validation exits 2,
a permission denial exits 3, other API errors exit 1.
Batch operations (adding tasks to a tasklist, creating a tasklist with tasks)
now report partial failure honestly: the per-item successes and failures stay
on stdout and the command exits non-zero instead of masking failures as a
success.
The okr and whiteboard commands now report every failure as a typed error
envelope. Invalid flags, malformed input, output-file conflicts, and API or
transport failures alike carry a stable category, subtype, the offending flag
or Lark error code, and a meaningful exit code — so scripts and agents can
branch on the error shape instead of scraping message strings.
Wire the board.whiteboard.updated_v1 EventKey into the consume pipeline so that lark-cli event consume automatically calls the per-whiteboard subscribe / unsubscribe OAPIs instead of requiring callers to manage server-side subscriptions out-of-band.
Change-Id: I94323807e8dc649d3296f6922311d2acaf92284e
Adds feed shortcut management to the im domain: pin chats to the user's feed sidebar, list pinned entries, and unpin them. Three new shortcuts wrap the im/v2/feed_shortcuts OpenAPI routes, which currently expose CHAT-type entries only and accept user identity only.
Calendar commands now return structured, typed error envelopes for every
failure mode — input validation, internal faults, and API responses —
instead of legacy generic errors. Callers and AI agents get consistent
exit codes and a machine-readable shape (type / subtype / code / hint),
and can tell bad input, an internal fault, and an API rejection apart.
Validation errors are attributed to the offending flag.
Server-supplied error details (e.g. why an event time was rejected) are
surfaced on the typed error's hint via a shared classifier improvement
that benefits every domain. Multi-step operations (create-with-attendees
rollback, multi-field update) preserve the real failure's classification
and report which steps completed.
The whole calendar domain is now lint-locked against reintroducing legacy
error constructors.
The recommend.allow list in scope_overrides.json special-cased a set of
calendar/contact/mail scopes into the auto-approve set on top of the
platform recommendations in scope_priorities.json. Remove all entries so
no scopes are special-cased anymore; auto-approve now reflects only the
platform recommend=true scopes (plus the recommend.deny removals).
Update registry tests to use a recommend=true scope (sheets:spreadsheet:read)
as the auto-approve sample and assert the override allow set is empty.
Change-Id: Ic555a2c664e2dbd742f79712253f2918dfabf7ce
Validate the example commands embedded in shortcut definitions (the
"Example: lark-cli ..." lines in each shortcut's Tips, shown in --help)
against the real command tree built by cmd.Build. Implemented entirely as
test-only code in cmd/ (package cmd_test), so it ships in no binary and is
not importable by product code; the truth source is cmd.Build, the same
tree the binary uses, so the check cannot drift. It runs in the standard
unit-test CI job (go test ./cmd/...); a renamed command or unaccepted flag
in an example fails that job.
* feat(mail): return typed error envelopes across the mail domain
Replace every produced error path in shortcuts/mail with typed errs.* envelopes, so consumers get stable category, subtype, param/params, hint, retryable, and log_id metadata for classification and recovery instead of free-form message text.
- Locally constructed mail errors move from output.Err* / output.Errorf / final fmt.Errorf / common legacy helpers to errs.* builders, with structured params on multi-flag validation and failed-precondition states kept non-retryable.
- API-call failures move from runtime.CallAPI / DoAPIJSON legacy boundaries to runtime.CallAPITyped or runtime.ClassifyAPIResponse, and mail-specific enrichers read errs.ProblemOf so typed code, subtype, hint, and log_id metadata are preserved.
- Batch draft-send partial failures now use runtime.OutPartialFailure so successful and failed draft sends stay in stdout while the command exits through a typed multi-status signal.
- Add mail-domain typed helpers, mail API code metadata, and guard wiring to keep shortcuts/mail from reintroducing legacy envelopes or legacy API calls.
- Keep genuine intermediate fmt.Errorf wraps in parser/builder layers annotated with nolint comments; command-facing paths wrap them into typed validation, API, network, or internal errors.
* fix(mail): report aborted draft-send batches as a single failure result
When an account-level failure interrupts a batch send after some drafts
already went out, the command previously produced two machine-readable
failure results: the partial-failure ledger on stdout and a second error
envelope on stderr. Consumers could not tell which one to recover from.
The batch ledger is now the only failure result for that case: it gains
aborted and abort_error fields carrying the typed cause, so callers can
see which drafts were sent, which failed, why the batch stopped, and how
to recover — all from stdout. A --stop-on-error stop keeps these fields
unset because stopping early there is the caller's own choice.
When triaging a public/shared mailbox, downstream AI consumers (e.g.
mail +message) need the mailbox_id to construct correct API paths.
Previously the triage output only included message_id, causing
/user_mailboxes/me/messages/{id} lookups that fail for public mailboxes.
- Add mailbox_id field to every normalized message in structured output
- Add mailbox_id to top-level JSON/data output envelope
- Add mailbox_id to table rows when mailbox is not "me"
- Update stderr next-step tip to include --mailbox for non-me mailboxes
- Update next-page hint to include --mailbox for non-me mailboxes
- Add unit tests covering list, search, and public mailbox paths
- Update triage skill docs to show mailbox_id in output examples
* feat(slides): add whiteboard element support and reference documentation
- Add lark-slides-whiteboard.md covering SVG and Mermaid modes, routing
rules, layout examples, known issues, and self-check checklist
- Register <whiteboard> in slides_xml_schema_definition.xml; remove it
from the undefined element type list
- Update SKILL.md quick-reference table and按需再读 section to point to
the new whiteboard reference
- Update xml-schema-quick-ref.md with <whiteboard> syntax examples
- Update slide create/get/replace references to include whiteboard as a
valid <data> child element
- Tighten fallback_if_missing descriptions in planning-layer.md and
asset-planning.md: replace "shapes" wording with neutral intent
language and add "whiteboard diagrams" to the fallback tool lists
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs(slides): refine whiteboard reference doc structure and content
- Restructure doc: common attributes and prerequisites moved to top
- Move design quality rules under SVG mode section
- Add z-order inline note to full-screen layout example
- Replace JS coordinate script with Python, broaden scope to decorative elements
- Delete redundant Mermaid examples (keep one complete whiteboard+flowchart)
- Add prerequisite link and references section
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs(slides): clarify chart vs whiteboard selection and fix doc gaps
- lark-slides-whiteboard: add chart vs whiteboard decision table at top;
fix intro and SVG use-case list to remove bar/line (those belong to <chart>)
- SKILL.md: split whiteboard quick-ref row into chart row + whiteboard row;
fix sidebar link label to match actual scope
- asset-planning: correct chart asset type — remove funnel/scatter (unsupported
by <chart> XSD) and note they fall back to <whiteboard> SVG
- visual-planning: add one-line whiteboard preference hint to
architecture-diagram and process-flow layout types
- validation-checklist: add Whiteboard Elements section noting slide.get
does not return SVG/Mermaid content; content correctness requires manual
visual sign-off
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs(slides): add SVG decorative visibility principles
Add two design rules to SVG quality requirements: check background
luminance before writing SVG (dark bg requires higher contrast), and
use non-linear brightness jumps (e.g. 0.10→0.40→0.70→1.0) instead of
linear opacity stacking (0.04→0.08→0.12) which produces near-identical
layers on dark backgrounds.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs(slides): add custom icon use case to whiteboard SVG
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs(slides): fix whiteboard SVG rendering rules
- content area is determined by child element bounding box union, not svg width/height/viewBox/xmlns
- viewBox only purpose: provide reference for percentage-based attribute values; omit when using absolute coords
- remove redundant attributes from all svg examples, use bare <svg> tags
- drop positive/negative coordinate guidance; rendering rule simplified to bounding-box auto-scale
* feat(lark-contact): route user_profiles batch_query in skill
- Add user_profiles batch_query row to the routing table.
- Add a worked example next to the search-user one, with `lark-cli
schema` first (best practice: don't guess `--data` / `--params`).
- Trim description: drop the duplicated trigger clause, add
personal_status / signature to the capability list so routing picks
this skill up for those queries.
* refactor(sheets): rebuild lark-sheets on sheet-skill-spec canonical + One-OpenAPI
Restart lark-sheets as a spec-driven downstream. Skill content (SKILL.md
and 16 references covering 13 operations skills + 3 workflow skills,
including the standalone filter-view skill) is mirrored from the
sheet-skill-spec canonical-spec; do not hand-edit, change upstream and
rerun npm run sync:consumers.
Drop the 11 legacy shortcut sources (spreadsheet / sheet management,
cell ops, dropdown, filter-view, float image, etc.) and 10 associated
tests. Wire up the new sheet_ai/v2 One-OpenAPI single entry that
dispatches by tool_name with JSON-string input/output, and land the
first canonical shortcut +workbook-info as a template that exercises
the public token XOR pair, Risk tiering, and zero-side-effect DryRun.
sheet_ai_api.go provides callTool / invokeToolDryRun and bypasses
runtime.CallAPI's silent swallowing of non-envelope responses so
gateway and business errors from the new endpoint surface precisely.
The remaining 55 shortcuts will be designed and landed separately,
canonical skill by canonical skill.
* feat(sheets): implement lark_sheet_workbook shortcuts (B1)
Land the 8 modify_workbook_structure shortcuts that round out the
lark_sheet_workbook canonical skill alongside the existing +workbook-info:
+sheet-create / +sheet-delete / +sheet-rename / +sheet-move / +sheet-copy
/ +sheet-hide / +sheet-unhide / +sheet-set-tab-color. All eight call
modify_workbook_structure via the One-OpenAPI invoke_write endpoint,
dispatched by the `operation` enum.
Helpers in helpers.go grow publicSheetFlags() / resolveSheetSelector() /
sheetSelectorForToolInput() / sheetSelectorPlaceholder() so future
sheet-level shortcuts share the public --sheet-id / --sheet-name XOR
treatment. +sheet-create intentionally drops the sheet selector pair since
create has no existing-sheet anchor (matches the spec fix in
tool-shortcut-map.json).
+sheet-delete is the first high-risk-write shortcut in the canonical
package; the framework requires --yes (exit code 10 otherwise).
+sheet-move's tool requires source_index in addition to target_index. The
CLI accepts an optional --source-index override and falls back to a
single get_workbook_structure read to derive it (and to resolve sheet_id
from --sheet-name). DryRun stays network-free by rendering <resolve>
placeholders for any field that would need that read.
* feat(sheets): implement lark_sheet_sheet_structure shortcuts (B2)
Add 8 shortcuts under the lark_sheet_sheet_structure canonical skill:
+sheet-info (get_sheet_structure) plus +dim-insert / +dim-delete /
+dim-hide / +dim-unhide / +dim-freeze / +dim-group / +dim-ungroup
(modify_sheet_structure, dispatched by operation enum).
Two reusable conversion helpers cover the impedance mismatch between
the CLI surface and the tool input:
- dimRange / dimPosition translate the CLI's 0-based exclusive-end
range into the tool's 1-based A1 notation. row 5..8 becomes
position "6" + count 3 (insert) or range "6:8" (range ops); column
26..29 becomes "AA:AC".
- infoTypeFromInclude maps the fine-grained --include vocabulary
(row_heights / col_widths / merges / hidden_rows / hidden_cols /
groups / frozen) to the coarse info_type enum the tool accepts;
mixed categories collapse to "all".
+dim-delete is high-risk-write (irreversible row/column removal).
+dim-freeze --count 0 auto-dispatches to operation=unfreeze. +dim-group
accepts --depth for forward-compat with a future server-side nested
group endpoint but does not pass it through today.
* feat(sheets): implement read_data / search_replace / write_cells shortcuts (B3)
Land 11 shortcuts across three canonical skills:
- lark_sheet_read_data (3): +cells-get / +csv-get / +dropdown-get
- lark_sheet_search_replace (2): +cells-search / +cells-replace
- lark_sheet_write_cells (6): +cells-set / +cells-set-style / +csv-put
/ +dropdown-set / +dropdown-update / +dropdown-delete
+dropdown-get reads the data_validation field via get_cell_ranges with
the range carrying its own sheet prefix (no --sheet-id needed). The
fine-grained --include vocabulary (value / formula / style / comment /
data_validation) maps to the tool's coarse include_styles bool plus
value_render_option enum. +csv-get's --include-row-prefix=false strips
the [row=N] prefix client-side because the tool only emits the
annotated form.
+cells-search / +cells-replace flatten the tool's options sub-object
into four independent flags (--match-case / --match-entire-cell /
--regex / --include-formulas) per the flat-flag rule, then repack them on the way
in.
+cells-set takes a raw --data JSON body whose `cells` array must match
the --range dimensions. +cells-set-style fans a single --style block
out to every cell in the range via a new fillCellsMatrix helper; the
range parser (rangeDimensions / splitCellRef / letterToColumnIndex)
only accepts rectangular A1:B2 forms — whole-column / whole-row need
sheet totals and are deferred.
+dropdown-set fans the validation block out to one range; +dropdown-
update / +dropdown-delete iterate sheet-prefixed --ranges and call
set_cell_range sequentially (partial failure leaves earlier ranges
already mutated; the Tip calls this out). +dropdown-delete is
high-risk-write and requires --yes.
+cells-set-image stays deferred to the cli-only batch (needs the
shared local-file upload helper alongside +workbook-create / +dim-move
/ +workbook-export).
* refactor(sheets): move +dropdown-update / +dropdown-delete to lark_sheet_batch_update
Follow-up to B3 after the spec re-mapped these two shortcuts to the
batch_update tool (atomic multi-range CRUD) instead of fan-out via
set_cell_range. Drop their Go implementations + helper validateDropdownRanges
+ splitSheetPrefixedRange from lark_sheet_write_cells.go and remove the
registrations from Shortcuts(); the shortcuts will reappear under
lark_sheet_batch_update during B7.
Also pull in the re-rendered reference docs:
- skills/lark-sheets/references/lark-sheets-write-cells.md
- skills/lark-sheets/references/lark-sheets-batch-update.md
* feat(sheets): implement lark_sheet_range_operations shortcuts (B4)
Land 8 shortcuts across four canonical tools:
- clear_cell_range → +cells-clear (high-risk-write)
- merge_cells → +cells-merge / +cells-unmerge
- resize_range → +dim-resize
- transform_range → +range-move / +range-copy / +range-fill / +range-sort
Three CLI↔tool vocabulary bridges live in this file:
- +cells-clear: --scope content normalizes to the tool's clear_type
"contents" (singular/plural spec mismatch is absorbed in the CLI).
- +dim-resize: --size <px> wraps as resize_{height,width}:{value:N};
--reset wraps as {reset:true}. The two flags are mutually exclusive
and at least one is required.
- +range-fill: CLI's five-valued --series-type collapses to the tool's
binary fill_type — `copy` → "copyCells", anything else → "fillSeries"
(the actual series progression is inferred server-side from the
seed cells in --source-range).
- +range-copy: --paste-type {values, formulas, formats} maps to the
tool's {value_only, formula_only, format_only}; "all" omits the
field entirely so the server applies its default.
+cells-clear is the second high-risk-write shortcut in the package;
the framework enforces --yes with exit code 10 as usual.
* feat(sheets): implement object-list shortcuts (B5)
Land 7 read shortcuts, one per object skill — chart / pivot table /
conditional format / filter / filter view / sparkline / float image. All
share the same shape (public sheet selector + optional <obj>-id filter)
so they're declared via newObjectListShortcut + an objectListSpec.
Notes:
- +cond-format-list exposes --rule-id, which is renamed to
conditional_format_id on the wire (the tool's full field name).
- +sparkline-list exposes --group-id (the higher-level handle); the
tool also accepts sparkline_id, intentionally not surfaced.
- +filter-list takes no id filter — at most one sheet-level filter
per sheet, so the listing is already unique.
- +filter-view-list is `cli_status: cli-only` but get_filter_view_objects
is in mcp-tools.json and dispatches through the same One-OpenAPI
endpoint; no special path required.
* feat(sheets): implement object CRUD shortcuts (B6)
Land 21 shortcuts — three (create / update / delete) per object skill —
backed by the manage_<obj>_object tools dispatched on the operation
enum. Five standard objects (chart / cond-format / sparkline /
float-image / filter-view) share an objectCRUDSpec factory; pivot and
filter are special-cased.
Shared wire contract:
excel_id + sheet_id|sheet_name + operation + [<obj>_id] + [properties]
CLI --data is passed through as the tool's `properties` field as-is, so
callers shape it per each object's spec doc.
Special cases:
- pivot adds optional --target-sheet-id / --target-position on create
(siblings of properties, not inside it).
- cond-format exposes --rule-id (short CLI name) wired to the tool's
conditional_format_id on the wire.
- sparkline uses --group-id (higher-level object handle) instead of
sparkline_id.
- filter has no separate id flag — at most one filter per sheet, so
filter_id is implicit. +filter-create promotes --range to a first-
class flag (instead of burying it inside --data).
- filter-view CRUD are `cli_status: cli-only` but
manage_filter_view_object is in mcp-tools.json, so they go through
callTool / One-OpenAPI alongside everything else.
All delete shortcuts are high-risk-write and require --yes.
* feat(sheets): implement lark_sheet_batch_update shortcuts (B7)
Land 4 shortcuts that all funnel through the batch_update tool's atomic
operations array:
- +batch-update raw passthrough; --data carries the full
{ operations: [{tool, params}, ...] } payload
plus optional continue_on_error. high-risk-write
since the caller may stuff anything inside.
- +cells-batch-set-style --data is [{ranges, style}, ...]; CLI flattens
each (entry × range) pair into a set_cell_range
op with a fan-out cells matrix carrying
cell_styles + border_styles.
- +dropdown-update --ranges + --options (+ --colors / --multiple /
--highlight) — installs/replaces one dropdown
across many ranges, each becoming a separate
set_cell_range op with data_validation in cells.
- +dropdown-delete --ranges — clears data_validation across many
ranges (high-risk-write).
Default is strict transaction: if any sub-tool fails the whole batch rolls
back. +batch-update exposes --continue-on-error to flip the policy; the
three fan-out shortcuts leave it strict (they're meant to be all-or-nothing).
Reinstates validateDropdownRanges + splitSheetPrefixedRange that were
removed during B3 → B7 relocation.
* feat(sheets): implement cli-only shortcuts (B8) — 70/70 complete
Land the four cli-only shortcuts that can't route through the One-OpenAPI
dispatcher (their backing capabilities aren't in mcp-tools.json):
- +workbook-create POST /open-apis/sheets/v3/spreadsheets
+ optional set_cell_range follow-up that zips
--headers and --data into the first sheet starting
at A1.
- +workbook-export POST /open-apis/drive/v1/export_tasks (type=sheet)
→ poll /export_tasks/:ticket up to ~30s
→ optional GET /export_tasks/file/:file_token/download.
CSV mode requires --sheet-id (single sheet export).
- +dim-move POST /open-apis/sheets/v2/spreadsheets/:token
/dimension_range
CLI is 0-indexed inclusive (--start / --end); the v2
endpoint expects half-open [startIndex, endIndex)
so the body uses endIndex = --end + 1. --sheet-name
is resolved client-side to sheet_id via
lookupSheetIndex when needed.
- +cells-set-image common.UploadDriveMediaAll
(parent_type=sheet_image, parent_node=token)
then callTool set_cell_range with cells carrying
rich_text: [{type:"embed-image", attachment_token, attachment_name}].
--range must be exactly one cell.
All four use runtime.CallAPI / DoAPI directly; only +cells-set-image
combines a legacy upload with the new One-OpenAPI for the second step
(set_cell_range is in mcp-tools.json so callTool is the right path).
This closes the migration: 70 shortcuts × 17 canonical skills × matching
the sheet-skill-spec v0.5.0 tool-shortcut-map.
* test(sheets): cover all 70 shortcuts with dry-run + execute-path tests
Twelve _test.go files alongside the implementation, mirroring the legacy
package's coverage style:
- testhelpers_test.go shared rig: TestFactory + Mount + dry-run
capture + JSON-input decode + envelope helpers.
- lark_sheet_*_test.go one test file per implementation file (9
files), table-driven dry-run cases per shortcut
plus targeted validation guards.
- execute_paths_test.go end-to-end execute paths via httpmock stubs.
Covers callTool unwrap, JSON-string output
decoding, two-step lookup (+sheet-move),
batch_update fan-out, dropdown atomic writes,
and the legacy OAPI shortcuts (+workbook-create,
+dim-move) including CLI inclusive → API
half-open index conversion.
Test coverage on the sheets package is 60.5 % of statements with -race
clean, meeting the dev manual's ≥ 60 % patch-coverage gate.
* refactor(sheets): inline cli-only shortcuts into their canonical skill files
Two naming cleanups:
- lark_sheet_cli_only.go is gone. The four shortcuts it grouped
(+workbook-create / +workbook-export / +dim-move / +cells-set-image)
were bundled by their implementation pattern (legacy OAPI direct
calls) rather than by canonical skill. The whole sheets package IS
the CLI implementation, so "cli only" wasn't a meaningful grouping
at the Go layer. Each shortcut now lives next to its skill peers:
+workbook-create / +workbook-export → lark_sheet_workbook.go
+dim-move → lark_sheet_sheet_structure.go
+cells-set-image → lark_sheet_write_cells.go
Per-skill shortcut counts now match tool-shortcut-map.json exactly
(workbook: 11, sheet_structure: 9, write_cells: 5). Helpers
(buildInitialFillInput, pollExportTask, downloadExportFile,
dimMoveBody) move with their shortcuts; nothing else in the package
referenced them.
- testhelpers_test.go → helpers_test.go. The _test.go suffix already
conveys "test"; the leading "test" was redundant. Matches the
helpers.go naming convention.
Behavior unchanged. go test -race -cover stays at 60.5 %.
* refactor(sheets): sync shortcut flags with sheet-skill-spec v0.5.0
Upstream hoisted a batch of high-frequency scalar fields out of --data
into independent flags and renamed several composite-JSON flags to
match their semantic content. CLI catches up.
Renames (drop-in, same payload semantics):
- +cells-replace --replace → --replacement
- +cells-set --data → --cells
- +workbook-create --data → --values
- +batch-update --data → --operations (now a bare array;
still accepts the envelope form for
back-compat with continue_on_error)
Flat-flag hoists out of --style / --data:
- +cells-set-style / +cells-batch-set-style
--style JSON drops; replaced by 11 flat style flags
(--background-color / --font-color / --font-size / --font-style /
--font-weight / --font-line / --horizontal-alignment /
--vertical-alignment / --word-wrap / --number-format) plus
--border-styles for the one field that's still nested. Both
shortcuts share styleFlatFlags() + buildCellStyleFromFlags().
- +cells-batch-set-style also drops the [{ranges, style}] array shape
in favor of one --ranges + the same flat style flags applied to
all of them.
Object CRUD --data → --properties everywhere (chart / pivot / cond-format
/ filter / filter-view / sparkline / float-image). Per-skill scalar
hoists merged into properties via an enhanceCreate/UpdateInput callback:
- +pivot-create adds --source (required), --range
(and continues to expose --target-sheet-id /
--target-position at top level)
- +cond-format-{create,update}
adds --rule-type (enum) + --ranges (JSON array);
merged into properties.rule.type and
properties.ranges respectively
- +filter-view-{create,update}
adds --view-name and --range; both override
their properties.* counterparts
- +filter-update adds first-class --range (was buried in --data)
Float-image is fully hoisted — no --properties flag at all. Ten flat
flags (--image-name / --image-token | --image-uri / --position-row /
--position-col / --size-width / --size-height / --offset-row /
--offset-col / --z-index) compose the properties block. Implemented as
its own factory (newFloatImageWriteShortcut) since it diverges from the
shared CRUD spec.
Tests track every flag renamed and add explicit cases for the new flag
combos. go test -race -cover stays at 60.3 %.
* refactor(sheets): align batch_update + cells-set with synced reference docs
Sync to upstream reference doc updates for 9 skills:
- batch_update sub-ops: rewrite wire fields tool/params -> tool_name/input
in CellsBatchSetStyle and DropdownUpdate/Delete fan-out (the actual
server contract per Schemas section); update --operations flag desc
and tests.
- +cells-set --cells: accept bare 2D matrix [[{cell},...],...] instead
of envelope {"cells":[[...]]}; spec example shows bare-array form.
- sparkline createDataDesc enum: win_loss -> winLoss (camelCase).
All other doc changes (float-image flat flags, cond-format
--rule-type/--ranges, pivot create-only --source/--range, filter /
filter-view extra flags, chart --properties) were already aligned in
commit ce33315.
* fix(sheets): repair cells-set-image rich_text embed payload
The server rejected set_cell_range calls from +cells-set-image with three
distinct errors: missing "text" property, missing image_width/image_height,
and unknown attachment_token field. Realign the rich_text element to the
embed-image schema (text/image_token/image_width/image_height) and decode
PNG/JPEG/GIF dimensions from the local file before the write.
* refactor(sheets)!: split +dim-resize into +rows-resize and +cols-resize
Sync to upstream spec change that splits the legacy +dim-resize shortcut
into +rows-resize and +cols-resize. Reasoning is that row vs column
resize has divergent semantics (only rows support auto-fit) and the
shared --dimension flag was hiding that.
Behavior changes (BREAKING):
- +dim-resize is removed; use +rows-resize or +cols-resize.
- --dimension and --reset flags are gone.
- --type enum replaces --size/--reset:
pixel (requires --size)
standard (reset to sheet default; no --size)
auto (auto-fit row height; +rows-resize only)
- --end is now inclusive (was exclusive). Old "--start 0 --end 5"
(5 rows) becomes "--start 0 --end 4".
- Wire payload for resize_height / resize_width changes from
{value: N} | {reset: true} to {type: "pixel", value: N} |
{type: "standard"} | {type: "auto"}.
Tests cover both shortcuts across pixel / standard / auto and the
new guard surface (--type pixel needs --size; standard/auto reject
--size; +cols-resize rejects --type auto; --end < --start).
Also pulls in synced reference docs for 5 skills (batch-update,
core-operations, range-operations, sheet-structure, visual-standards)
that update prose mentions of +dim-resize.
* feat(sheets): add --print-schema runtime introspection for composite JSON flags
Composite JSON flags (--cells / --properties / --operations /
--border-styles / --sort-keys / --options) carry non-trivial structured
payloads. Reference docs cover top-level fields but agents writing
those flags often need the full JSON Schema to build a valid payload.
This adds a system-level introspection contract so any shortcut whose
flags are tracked upstream can serve its schemas locally:
lark-cli sheets <shortcut> --print-schema --flag-name <name>
lark-cli sheets <shortcut> --print-schema # list flags
The schema data is embedded at build time from a synced artifact
(shortcuts/sheets/data/flag-schemas.json). Upstream is the source of
truth — never hand-edit the JSON; update the source Base table and
rerun the sheet-skill-spec sync.
Framework changes (shortcuts/common):
- types.go: Shortcut gains an opt-in PrintFlagSchema hook
(flagName -> bytes/error). When non-nil the framework auto-injects
--print-schema / --flag-name and short-circuits Validate/Execute.
- runner.go: register the two system flags when PrintFlagSchema is
set; intercept in runShortcut before identity/scope/config so
pure-local lookups don't trigger auth or network. Install a
PreRunE that relaxes cobra's required-flag gate when
--print-schema is set, since asking for a schema shouldn't need
unrelated required flags.
Sheets surface (shortcuts/sheets):
- flag_schema.go (new): go:embed data/flag-schemas.json; expose
printFlagSchemaFor(command) closure. When flagName is empty it
emits a JSON listing of introspectable flags for discovery;
otherwise it returns the schema subtree as pretty JSON.
- flag_schema_test.go (new): cover embed parsing, listing /
by-name lookup, unknown-flag error path, registration via
Shortcuts(), and the full system-flag short-circuit through
cobra (required flags relaxed, schema printed on stdout).
- shortcuts.go: Shortcuts() now wraps shortcutList() and attaches
PrintFlagSchema to every command present in flag-schemas.json,
so shortcuts opt in by being listed upstream — no per-shortcut
boilerplate.
- data/flag-schemas.json (new, synced from sheet-skill-spec):
19 entries, schema_version "2". Generated upstream from the Lark
Base source-of-truth (see sheet-skill-spec
scripts/fetch_cli_flag_schema_map.mjs); ships only per-flag
subtrees (not the full mcp-tools.json) to keep tool internals
out of the open-source repo.
Skill docs (skills/lark-sheets):
- SKILL.md: system-flag table gains --print-schema / --flag-name and
an "Agent 使用提示" note steering agents to prefer --print-schema
over guessing JSON shape from the cheatsheet.
- references/*.md: regenerated by upstream sync (Schemas-section
boilerplate updated, plus accumulated upstream prose refinements).
* docs(sheets): remove sandbox references and normalize tool names to CLI shortcuts
Replace export_sheet_to_sandbox / import_sandbox_to_sheet / doubao_code_interpreter
with local-script + batch csv-get/csv-put workflows; unify legacy MCP tool names
(set_cell_range, get_range_as_csv, etc.) to CLI shortcut format (+cells-set, +csv-get).
* feat(sheets): add flag-descriptions.en.json and wire applyFlagDescs into Shortcuts()
Embed data/flag-descriptions.en.json (synced from upstream spec) and
apply it at shortcut assembly time so every Flag.Desc is sourced from
the canonical JSON rather than hardcoded Go strings. Existing hardcoded
Desc values serve as fallback for flags not yet in the JSON.
Also sync reference doc updates from upstream.
* feat(shortcuts): support int64 and float64 flag types
Flag.Type previously could not express non-integer numbers. Add int64
and float64 cases to flag registration plus Int64/Float64 runtime
accessors.
* refactor(sheets): build shortcut flags generically from flag-defs.json
Replace flag-descriptions.en.json with the richer flag-defs.json (full
flag definitions: type / default / enum / input / hidden / required /
kind) synced from sheet-skill-spec. Add flagsFor(command) to materialize
each shortcut's []common.Flag straight from the JSON, skipping
system-kind flags the framework injects.
Migrate every sheets shortcut (including the CRUD/list/dim/merge/
visibility factories) to Flags: flagsFor("+command"), dropping all
hand-written flag literals plus the now-dead publicTokenFlags /
publicSheetFlags / styleFlatFlags helpers and enum vars. A coverage test
locks the Go-flags-match-JSON contract.
Align Go with the new spec where they diverged: +cells-get --ranges →
--range, font-size int → float64, +filter-view-create --range now
required, +sheet-create row/col-count defaults 200/20.
* docs(sheets): sync +batch-update CLI override schema (shortcut/input form)
Pulled from sheet-skill-spec:
- skills/lark-sheets/references/lark-sheets-batch-update.md: --operations
now documents the {shortcut, input} form; tool_name references gone
- shortcuts/sheets/data/flag-schemas.json: --operations resolves to the
CLI-side array<{shortcut(enum), input}> schema, sourced from spec's
canonical-spec/tool-schemas/cli-schemas.json (cli: prefix). +dropdown
--options also drilled one level deeper
NOTE: the binary still raw-passes --operations to MCP batch_update which
expects {tool_name, input}. A follow-up will add a shortcut→tool_name
translation layer (with per-shortcut operation field) before the docs
become actionable.
* feat(sheets): translate +batch-update sub-ops {shortcut,input} → MCP shape
Users now hand +batch-update --operations a CLI-shape array
([{shortcut, input}, ...]) and the binary translates each sub-op to the
underlying MCP batch_update shape ({tool_name, input(+operation)}) via
a new dispatch table in shortcuts/sheets/batch_op_dispatch.go.
Dispatch table covers 50 batchable write shortcuts. Excluded by design:
- all read ops
- fan-out wrappers (+batch-update self, +cells-batch-set-style,
+dropdown-update, +dropdown-delete) — nesting these = nested batch
- +dim-move — single shortcut uses legacy v2 /dimension_range endpoint,
not MCP, can't be batched
- +cells-set-image — multi-step image upload, not atomic-batch friendly
- +workbook-create — new workbook, not batch-on-existing semantics
Translator also rejects sub-ops that hand-fill input.operation (implied
by shortcut name) or input.excel_id / spreadsheet_token / url (set
once at +batch-update top level).
+dim-freeze always injects operation=freeze; the count==0 unfreeze
path of the single shortcut is intentionally not supported in batch —
callers should use the single shortcut for unfreeze.
Tests cover: end-to-end translation, --continue-on-error propagation,
13 rejection cases (banned shortcuts, malformed shapes, reserved keys).
Sync'd from sheet-skill-spec: skills/lark-sheets/references/
lark-sheets-batch-update.md + shortcuts/sheets/data/flag-schemas.json
pick up the corrected enum (+cells-set-style / +dropdown-set added,
+dim-move removed).
* fix(sheets): make +batch-update sub-ops reuse standalone flag→body translators
Sub-ops previously near-passed-through their input, so any shortcut whose
standalone translator renames fields broke inside a batch: +range-copy lost
range/destination_range (transform_range errored "range missing") and
+rows-resize lost range/resize_height ("No resize operation specified").
Introduce a flagView interface (satisfied by *common.RuntimeContext) and a
map-backed mapFlagView, then route every batchable sub-op through the SAME
*Input builder the standalone shortcut uses. mapFlagView seeds flag-defs.json
defaults for value reads while keeping Changed() user-driven, so a sub-op body
is byte-identical to the standalone body — locked by a batch-vs-standalone
contract test over all ~40 batchable shortcuts.
Also fix single-row/column resize: start==end now formats as "23:23" / "C:C"
(resize_range rejects a bare "23"); dimRangeFull keeps both sides while
dimRange's collapse stays for modify_sheet_structure consumers.
* fix(sheets): align +cells-get/+csv-get range flags with synced spec
sheet-skill-spec now declares +cells-get --range as a single string
(was string_array) and +csv-get --range as required. Match the
flag→body translators:
- +cells-get wraps the single --range into the tool's `ranges` array
and validates with Str() instead of StrArray(), which silently
returned nil against the now-String flag and broke the command.
- +csv-get gains a trim-based required-range guard.
Update read-data dry-run tests to single-range form and add a guard
test for the empty --range path.
* fix(sheets): push +batch-update sub-op validation down into xxxInput builders
Sub-ops that omit --sheet-id (or any other required flag) used to slip
past CLI validation — Validate ran only against the standalone shortcut
path, and batchOpDispatch's translators built bodies from whatever
flagView returned, so a structurally broken sub-op surfaced as an opaque
server "sheet undefined not found" after a network round-trip.
Push each batchable shortcut's check trio down into its xxxInput builder:
1. resolveSpreadsheetToken — stays in Validate (batch already does it
once at the top level; sub-ops don't repeat).
2. requireSheetSelector(sheetID, sheetName) — new helper; flagView-
agnostic XOR + control-char check, called at the top of every
xxxInput.
3. shortcut-specific required / range / enum checks (--dimension,
--range, --start <= --end, --type pixel needs --size,
--float-image-id, image-token XOR image-uri, ...) — moved out of
Validate into the builder body.
All ~30 batchable xxxInput builders now return (map, error). Standalone
Validate shrinks to validateViaInput(xxxInput); DryRun / Execute
propagate the error. batch_op_dispatch entries drop the noErrTranslate
wrapper and pass the builder directly — its error bubbles up wrapped
with "operations[N] (+shortcut):" context.
Tests:
- TestBatchOp_ErrorEquivalence (7 cases): XOR / logical-constraint
errors fire identically from standalone and batch sub-op paths.
- TestBatchOp_RejectsBadSubOpInput (8 cases): cobra-required flags that
standalone catches via MarkFlagRequired now also get rejected CLI-side
on the batch path (where cobra is not in the loop).
- TestBatchOp_BodyMatchesStandalone (~40 cases) and
TestBatchOp_DispatchCoversReportedBugs continue to pass — bodies stay
byte-identical.
- BOE smoke (spreadsheet ICFwstkUGheyfptGWS2bB7RgcDf, sheet 51991c):
+batch-update with a sub-op missing --sheet-id now returns
"operations[0] (+dim-insert): specify at least one of --sheet-id or
--sheet-name" before any network call.
sheetMoveBatchInput (xiongyuanwen's batch-only explicit-source-index
requirement) is preserved — it's an orthogonal batch-specific constraint
not affected by this push-down.
* fix(sheets): align +cond-format / +filter with server schema (#4 + #5)
Two latent bugs in the object_crud translator surfaced during BOE smoke
testing of +batch-update. Both are schema-alignment fixes against
manage_conditional_format_object / manage_filter_object as declared in
sheet-skill-spec/canonical-spec/tool-schemas/mcp-tools.json.
#4 +cond-format: rule_type path + enum vocabulary
---------------------------------------------------
condFormatEnhance used to write the user's --rule-type value into
`properties.rule.type` (nested under a `rule` object). The server
schema actually puts it at flat `properties.rule_type` and silently
drops the nested form — so every conditional-format create/update
secretly built the wrong document.
Worse, the CLI enum exposed via flag-defs.json was its own invented
vocabulary (cellValue / formula / duplicate / unique / topBottom /
aboveBelowAverage / dataBar / colorScale / iconSet / textContains /
dateOccurring / blankCell / errorCell) — none of those values were
the strings the server accepts.
Fix:
- condFormatEnhance now writes `properties.rule_type = <value>`
directly (no nested `rule` object).
- Synced flag-defs.json + lark-sheets-conditional-format.md enum
vocabulary from base to match the server: duplicateValues,
uniqueValues, cellIs, containsText, timePeriod, containsBlanks,
notContainsBlanks, dataBar, colorScale, rank, aboveAverage,
expression, iconSet.
- ⚠️ Breaking: scripts passing the old CLI-invented enum values
(e.g. --rule-type cellValue) now get a cobra "invalid value …
allowed: …" error listing the new vocabulary. No alias layer.
- TestObjectCRUDShortcuts_DryRun's +cond-format-update case updated
to assert the flat properties.rule_type shape + new enum.
#5 +filter-{update,delete}: auto-inject filter_id = sheet_id
-------------------------------------------------------------
manage_filter_object's contract is "filter_id === sheet_id" for the
sheet-scoped filter (per per-tool description in mcp-tools.json),
and update / delete operations MUST carry filter_id. Standalone
filterUpdateInput / filterDeleteInput never set it, so the server
rejected with "filter_id is required for update/delete operation"
on every call — both standalone AND inside +batch-update.
Fix:
- filterUpdateInput / filterDeleteInput now set
input["filter_id"] = sheetID.
- Because filter_id must equal sheet_id (not sheet_name), update /
delete reject when only --sheet-name is given — there's no
network lookup available inside the builder. The friendly error
points at +workbook-info for resolving sheet-name → sheet-id.
- create still omits filter_id (server requires that — id is
server-allocated on creation).
- New tests:
* TestObjectCRUDShortcuts_DryRun gains a +filter-update happy-path
case asserting filter_id is auto-injected + --range hoisting.
* +filter-delete case updated to assert filter_id presence.
* TestBatchOp_RejectsBadSubOpInput gains two cases asserting both
+filter-update and +filter-delete reject --sheet-name-only with
the friendly error.
Docs (#2 + #3 + #8) synced from sheet-skill-spec
-------------------------------------------------
Companion doc fixes that landed via npm run generate:cli + sync:cli
in sheet-skill-spec; included here because the regenerated flag-defs
and references markdown are byte-tracked in this repo:
- #2: lark-sheets-sheet-structure.md — +dim-{hide,unhide,group,
ungroup} --start/--end desc changed from "(0-based, inclusive)" to
"(0-based)" / "(exclusive)" to match the half-open range semantics
the code has always implemented (requireDimRange: end > start;
dimRange uses end - 1 for column end letters).
- #3: lark-sheets-workbook.md — +sheet-move section gains a note
about the batch-internal requirement to pass --sheet-id AND
--source-index explicitly (sheetMoveBatchInput's constraint).
- #8: lark-sheets-pivot-table.md — +pivot-create --properties
example drops the stale data_range field (the actual server
schema uses --source as a hoisted flag; properties only carries
rows / columns / values / filters / show_*_grand_total).
* feat(sheets): add +cells-batch-clear fan-out over batch_update
Clear content/formats across many sheet-prefixed ranges in a single atomic
batch_update (one clear_cell_range op per range), mirroring the existing
+cells-batch-set-style / +dropdown-{update,delete} fan-out wrappers. The
--scope to clear_type normalization is shared with standalone +cells-clear
(normalizeClearType) so the two stay in lockstep.
high-risk-write (requires --yes); rejected as a batch sub-op like the other
fan-out wrappers. flag-defs/flag-schemas and skill docs updated to match.
* docs(sheets): sync stdin guidance and sparkline reference
- skills/lark-shared/SKILL.md: drop the generic "prefer stdin" section
- skills/lark-sheets/SKILL.md: add expanded stdin guidance (use stdin over @file abs paths; don't cd or write into the project dir)
- skills/lark-sheets/references/lark-sheets-sparkline.md: document the group_id / sparkline_id two-tier model with worked examples
* fix(sheets): require sparkline_id on +sparkline-update items (#6)
manage_sparkline_object uses two layers of IDs: --group-id picks the
sparkline group, and properties.sparklines[i].sparkline_id picks each
item inside the group. The server contract requires sparkline_id on
every update item (server maps each entry back to an existing
sparkline by this id). Agents that called +sparkline-update without
the per-item ids hit an opaque server-side rejection that didn't
mention sparkline_id at all, then got stuck in a try-fail-list-retry
loop.
Pre-check CLI-side in objectUpdateInput via a new validateUpdateInput
hook on objectCRUDSpec. sparklineSpec wires validateSparklineUpdateItems,
which walks properties.sparklines[] and rejects with a message that
points at +sparkline-list:
+sparkline-update properties.sparklines[N] missing sparkline_id
(run `+sparkline-list --group-id <id>` first to read sparkline_id
for each item, then echo each id back on the corresponding update
entry)
Scope is update-only. config-only updates (properties.config without
sparklines) stay legal — the validator skips when sparklines is
absent. Delete is not pre-checked: objectDeleteInput doesn't pass
properties through, so the partial-delete branch can't be reached
today (separate follow-up).
Tests:
- TestObjectCRUDShortcuts_DryRun: positive case for update with
sparkline_id present.
- TestSparklineUpdate_MissingSparklineID: standalone path — error
contains both "missing sparkline_id" and "+sparkline-list".
- TestBatchOp_RejectsBadSubOpInput: batch sub-op missing sparkline_id
rejected with the same friendly error.
Docs synced from sheet-skill-spec (canonical change committed there):
skills/lark-sheets/references/lark-sheets-sparkline.md documents the
two-layer id model, the three "+sparkline-list first" cases, and both
delete modes.
* docs(sheets): sync lark-sheets skill from spec (audit 20260521)
Pull latest spec from sheet-skill-spec (PR ee/sheet-skill-spec!6 + earlier
develop commits) into skills/lark-sheets/ and shortcuts/sheets/data/.
Audit findings now reflected in CLI docs:
- A2 +cond-format-create example: --rule-type duplicate → duplicateValues
- A3 +cond-format-create Validate: cellValue/formula → cellIs/expression
- A5 +csv-put examples: --range → --start-cell; drop redundant --allow-overwrite
- A7 +sparkline-create: Validate / Examples aligned with real schema
(config/sparklines), executable JSON example added
- B13 cross-doc dead links: lark_sheet_*/cli-shortcuts.md → lark-sheets-*.md
- C2 +csv-put: `=` literal warning next to Examples
- CC5 +rows-resize/+cols-resize --type auto: single point of truth in
range-operations reference
flag-defs.json description / required sync (from base):
- A4 +float-image-update: image-name/position-*/size-* required → optional
(patch mode)
- A8 +dim-move --start/--end description cleanup
- B3 +pivot-create --properties: data_range → source (real field name)
Also picks up the +cells-batch-clear shortcut doc (introduced in spec
develop). Go-side implementation for that shortcut is intentionally not
in this PR — docs-only preview; runtime dispatch will land in a follow-up.
`go test ./shortcuts/sheets/...` passes.
* feat(sheets): add +cells-set --copy-to-range and sync skill spec
Sync lark-sheets skill references and flag schemas from upstream
sheet-skill-spec, and wire the newly-specced --copy-to-range flag into
+cells-set: it passes copy_to_range to the set_cell_range tool so a
template block written via --cells fans out across a larger range with
auto-shifted formula refs.
* docs(sheets): sync lark-sheets skill spec (chart/pivot wire mappings, --end semantics)
Sync skill references and flag-defs descriptions from upstream
sheet-skill-spec: clarify +chart-create properties structure
(snapshot.data), +pivot-create --target-position / --range wire-field
mappings, add a cross-command --end endpoint-semantics table
(insert/delete/hide/group exclusive vs move/resize inclusive), note
--group-state default, and rename reference identifiers to lark-sheets-*.
Description-only refinement; the existing CLI implementation already
matches the clarified wire mappings and --end semantics.
* fix(sheets): make --max-chars the single read cap for +cells-get / +csv-get
Drop --cell-limit (+cells-get) and --max-rows (+csv-get) from the CLI surface
and pin the underlying tool's cell_limit / max_rows to a very large sentinel so
the tool's own defaults never truncate before --max-chars. --max-chars stays the
only knob (default 200000, unchanged).
- lark_sheet_read_data.go: add unboundedReadLimit (1e9); cellsGetInput pins
cell_limit, csvGetInput pins max_rows; --max-chars still passed through
- data/flag-defs.json: synced from spec (drops the two flags)
- tests: spot-check moved to --max-chars; dry-run wantInput asserts cell_limit /
max_rows are pinned high
Mirrors sheet-skill-spec (Base flag records removed).
go build ./... + go test ./shortcuts/sheets/ green.
* docs(sheets): sync lark-sheets read docs — --max-chars as single read cap
Sync skills/lark-sheets references from spec: drop --cell-limit / --max-rows
guidance; 大表分批读 switches to --range row windows + --max-chars auto cap + has_more.
Mirrors sheet-skill-spec 58e7456 and handler change 2befc49.
* docs(sheets): sync lark-sheets skill spec from upstream
Refine reference docs and flag-defs descriptions from upstream
sheet-skill-spec (--depth wording for +dim-group / +dim-ungroup,
plus assorted reference clarifications). Description-only; no CLI
behavior or flag surface change.
* docs(sheets): sync chart properties schema (position/size required)
Regenerate flag-schemas.json from upstream sheet-skill-spec: the chart
properties schema now marks position and size as required, and the chart
reference doc reflects the same. flag-schemas.json is print-schema-only
(no client-side validation), so this is a generated-artifact + doc sync
with no CLI behavior change.
* docs(sheets): sync lark-sheets skill spec from upstream
Refine reference docs and flag-defs descriptions from upstream
sheet-skill-spec: clarify +workbook-export sheet flag scope, +filter-*
--properties optionality (omitted => empty filter on --range; rules must
be non-empty when provided), float-image reference_id wording, and
assorted reference cleanups. Description-only; existing CLI behavior
(filter passthrough, properties optional) already matches.
* docs(sheets): sync lark-sheets skill spec from upstream
Trim and refine reference docs from upstream sheet-skill-spec
(condense core-operations workflow, tidy write-cells / range-operations /
float-image / SKILL guidance). Description-only; no flag or CLI behavior
change.
* docs(sheets): sync lark-sheets skill spec from upstream
Refine reference docs from upstream sheet-skill-spec (core-operations,
formula-translation, visual-standards, SKILL guidance). Description-only;
no flag or CLI behavior change.
* fix(sheets): correct +workbook-create initial fill and +dim-move endpoint
+workbook-create: the v3 create response does not echo the default sheet's id, so the initial-fill set_cell_range was sent with an empty sheet_id and rejected ("sheet_id or sheet_name is required"). Resolve the workbook's first sheet via get_workbook_structure before filling.
+dim-move: the move request was POSTed to the v2 dimension_range endpoint (the add/update/delete surface, which requires a `dimension` object) and rejected with "[9499] Missing required parameter: Dimension". Switch to the native v3 move_dimension endpoint (sheet_id in path; snake_case source.{major_dimension,start_index,end_index} + destination_index). CLI --end and v3 end_index are both 0-based inclusive, so they pass through unchanged.
* fix(sheets): align +workbook-create, +dropdown-*, +dim-move, +range-sort with server schema
Five separate E2E failures in shortcuts/sheets/ that all trace back to a
CLI ↔ server contract mismatch. Each is independently scoped; bundling
them because they share the test-report citation and the same one-line
fix shape in most cases.
buildInitialFillInput sent {"sheet_id": ""} on the secondary
set_cell_range call after creating the workbook. The empty value was a
holdover from "...otherwise server picks first sheet" — but
set_cell_range rejects an empty selector with
"sheet_id or sheet_name is required" rather than falling back to the
default sheet.
Use sheet_name "Sheet1" instead. POST /sheets/v3/spreadsheets always
creates that sheet on workbook creation, and set_cell_range accepts
sheet_name as an equivalent selector — saves an extra
get_workbook_structure round-trip just to learn the auto-generated id.
buildDropdownValidation emitted four fields that don't exist in the
canonical set_cell_range.data_validation schema:
- "values" (options list) → renamed to "items"
- "multiple_values" → renamed to "support_multiple_values"
- "colors" (per-option color) → removed (not in schema; flag also
removed from data/flag-defs.json
for +dropdown-set / -update)
- "highlight_options" → removed (not in schema; flag also
removed)
The canonical schema lives at sheet-skill-spec/canonical-spec/tool-
schemas/mcp-tools.json (set_cell_range tool, data_validation property);
the colors / highlight knobs were CLI inventions the server never
accepted, so removing the flags is correct (renaming would leave the
flags broken). Skill reference docs (write-cells.md, batch-update.md)
synced.
validateDropdownOptionsColors lost its colors check; renamed to
validateDropdownOptions to reflect the narrower contract.
dropdownGetInput sent "Sheet1!C2:C6" verbatim as a ranges[] entry.
get_cell_ranges expects sheet_id / sheet_name as separate fields and
ranges entries without the sheet prefix; the server bounced with
"sheet not found, sheetId:" (empty).
Use the existing splitSheetPrefixedRange helper (declared in
lark_sheet_batch_update.go) to break "Sheet1!C2:C6" into ("Sheet1",
"C2:C6"), then thread the sheet name through sheetSelectorForToolInput
exactly like +cells-get does.
The shortcut was POSTing to /sheets/v2/spreadsheets/{token}/dimension_
range, which is the v2 insert-dimension endpoint and requires a top-
level {"dimension": {...}} body. Move uses a separate endpoint:
POST /sheets/v2/spreadsheets/{token}/move_dimension
body: { "source": {...}, "destination_index": N }
(camelCase "destinationIndex" → snake_case "destination_index" to
match the v2 contract.) Both DryRun and Execute updated, plus the
TestDimMove_DryRun and TestExecute_DimMove assertions.
transform_range.sort_conditions[i] requires both `column` (string) and
`ascending` (bool); rangeSortInput passed the --sort-keys array through
to the server unvalidated, so missing fields surfaced as opaque
"required property X missing" errors with no per-item context.
Walk the parsed array client-side, reject with item-pointing messages.
Test fixtures and a contract-test fixture switched from the historical
{col, order} vocabulary (which the server has never accepted) to the
correct {column, ascending}.
Server-schema citations and test-report case mapping in this branch's
plan file.
* revert(sheets): drop direct flag-defs.json edits — generated from spec
data/flag-defs.json is regenerated from the upstream sheet-skill-spec
canonical-spec; editing it here gets clobbered on the next sync. The
schema realignment for +dropdown-set / -update --colors / --highlight
removal needs to land on the base table first, then flow back through
sheet-skill-spec → larksuite-cli sync, not via a direct CLI-side edit.
Restore the previous flag entries verbatim. The Go-side change in
buildDropdownValidation still drops the wire fields, so:
- users passing --colors / --highlight today see the flag accepted
silently (no effect on the wire) until the upstream removal lands;
- after upstream removal + sync, both the flag declarations and the
Go-side handling will be in sync.
Functional fixes (#1 workbook-create, #3 dropdown-get, #4 dim-move,
#5 range-sort) and dropdown wire-shape rename (#2) are unaffected.
* revert(sheets): drop direct edits to skills/lark-sheets/references/
These md files are sync targets generated from sheet-skill-spec; editing
them here gets clobbered on the next sync, same as data/flag-defs.json.
The --colors / --highlight row removals belong on the upstream base
table → canonical-spec sync, not here.
Restore the previous --colors / --highlight rows in both
lark-sheets-write-cells.md (+dropdown-set) and lark-sheets-batch-update.md
(+dropdown-update). The Go-side change in buildDropdownValidation still
drops the wire fields, so:
- users passing --colors / --highlight today see the flag accepted
silently (no effect on the wire) until upstream removes the flag;
- after upstream removal + sync, both flag declarations, ref docs, and
Go-side handling will be in sync.
Functional fixes (#1 workbook-create, #3 dropdown-get, #4 dim-move,
#5 range-sort) and dropdown wire-shape rename (#2) are unaffected.
* docs(sheets): sync from sheet-skill-spec — remove dropdown --colors / --highlight
Upstream sheet-skill-spec base table deleted the --colors and --highlight
flags on +dropdown-set / +dropdown-update (the corresponding wire fields
data_validation.colors / .highlight_options were never accepted by the
server schema; see prior fix in this branch). Re-running the sync from
canonical-spec brings the CLI flag-defs and skill reference docs back in
line with the Go-side handling that already drops these fields.
Generated by `npm run sync:cli` in sheet-skill-spec @ ac7acef.
* fix(sheets): restore +dropdown --colors / --highlight, map to canonical fields
Reverses the --colors / --highlight removal from 7932ab2 (item #2 of the
batch-1 schema-alignment commit). That commit dropped both flags after the
test report flagged data_validation.colors / highlight_options as "unexpected
property" — at the time the canonical set_cell_range.data_validation schema
listed only help_text / items / operator / range / support_multiple_values /
type / values, so the flags had no server-side target and the removal was
correct.
Since then, set_cell_range.data_validation has gained two fields explicitly
modelling the dropdown highlight UI (mcp-tools.json in sheet-skill-spec
2026-05-22 base sync):
enable_highlight (bool) — show pill backgrounds
highlight_colors (string[]) — hex pill colors, length must match items
So the flags are back, but rewired:
--colors -> data_validation.highlight_colors (was: colors)
--highlight -> data_validation.enable_highlight (was: highlight_options)
--options -> items and --multiple -> support_multiple_values renames from
7932ab2 are kept.
Changes:
- buildDropdownValidation: re-add --colors / --highlight handling against
the new field names; --colors length check stays inline (so dropdownSetInput
Validate path catches it via validateViaInput, no separate guard needed).
- validateDropdownOptions -> validateDropdownOptionsColors: restore the
Validate-time --colors length check on +dropdown-update / +dropdown-delete
(called from lark_sheet_batch_update.go).
- TestDropdownSet_CellsShape: extend to assert highlight_colors /
enable_highlight emitted; assert legacy `colors` / `highlight_options`
absent.
- TestDropdownSet_ColorsLengthMismatch: new — covers the early Validate
error path.
- TestDropdownUpdate_BatchPayload: extend to cover dropdownBatchInput
propagation of --colors / --highlight through batch_update.
- skills/lark-sheets/references/lark-sheets-{write-cells,batch-update}.md,
shortcuts/sheets/data/flag-defs.json, flag-schemas.json: synced from
sheet-skill-spec generate output (MR !7).
* chore(sheets): re-sync from spec + loosen --colors length check
Catches up to sheet-skill-spec's 2026-05-25 base sync (MR !7) after
rebasing onto upstream feat/lark-sheets-refactor (12 new upstream commits
including the lark-sheets skill refactor + tools-schema migration).
Spec changes flowing in:
- highlight_colors description loosened: length may be **shorter than**
--options (server cycles remaining slots through a built-in 10-color
palette); previously the tool errored on any length mismatch.
- shortcuts/sheets/data/flag-schemas.json: mass re-mirror — generator now
emits `type` before `properties` and adds explicit `additionalProperties:
false` on object schemas (cosmetic, no behavior change).
- skills/lark-sheets/references/lark-sheets-{batch-update,chart,write-cells}.md:
--options gains the type='list' tag; data_validation inline field-count
goes 7 → 9 (catches up the highlight schema in the summary); chart
position / size marked optional per upstream.
Go-side adjustment:
- buildDropdownValidation / validateDropdownOptionsColors: change the
--colors length check from strict-equal to "must not exceed --options"
to match the relaxed schema.
- TestDropdownSet_ColorsLengthMismatch -> TestDropdownSet_ColorsLongerThanOptions
(now hits the overflow path with 3 colors vs 2 options).
- New TestDropdownSet_ColorsShorterAccepted: 2 colors vs 4 options is
legal and forwarded as-is.
* docs(sheets): sync dropdown --colors/--highlight clarification from spec
Mirrors sheet-skill-spec MR !7 changes:
- skills/lark-sheets/references/lark-sheets-write-cells.md: new "Dropdown
配色" section explaining how --colors (→ data_validation.highlight_colors)
and --highlight (→ data_validation.enable_highlight) compose — length
rule (shorter ok, longer rejected), --highlight gating, palette
fallback behavior, minimal +dropdown-set example.
- skills/lark-sheets/references/lark-sheets-batch-update.md: one-line
pointer to the write_cells section for +dropdown-update / -delete
(same rules).
- shortcuts/sheets/data/flag-defs.json: --colors / --highlight `desc`
fields gain the long-form server-field / length-rule descriptions
used by `--help`.
No Go-side change — earlier commit 538eb2e already loosened the
buildDropdownValidation length check to "must not exceed"; this PR step
just makes the docs / `--help` text catch up.
* feat(sheets): +dropdown-set/-update --source-range for listFromRange mode
Previously +dropdown-set / +dropdown-update only emitted
data_validation.type=list — agents wanting listFromRange (dropdown options
sourced from existing cells, kept in sync with that range) had to drop down
to +cells-set and hand-build a data_validation map. The flag now exposes it
natively as --source-range, paired with --options under XOR.
CLI changes:
- shortcuts/sheets/lark_sheet_write_cells.go:
* new dropdownTypeAndItems(runtime) — central XOR resolver: rejects 0 or
2 of {--options, --source-range}, returns (sourceSize, partial dv with
type+items|range filled in). Source size = options length for list
mode, rangeDimensions(--source-range) cell count for listFromRange.
* buildDropdownValidation rewritten to call the resolver, then layer
--colors / --multiple / --highlight on top — semantics unchanged
for callers, just two modes instead of one.
* validateDropdownOptions / -Colors renamed to validateDropdownSourceOrOptions
so the XOR + length check fires at +dropdown-update Validate time too.
* --colors length error message generalized: "must not exceed dropdown
source size (N)" (covers both modes).
- shortcuts/sheets/lark_sheet_batch_update.go: rename call site.
- shortcuts/sheets/lark_sheet_write_cells_test.go: 4 new tests —
ListFromRange (happy path: range + items absent + colors + highlight all
emit), ListFromRange_ColorsLongerThanCells (overflow against T1:T3 cell
count), XorBothSet, XorNeitherSet. Updated the existing
ColorsLongerThanOptions assertion to match the new "source size" wording.
Spec-driven changes (synced via npm run sync:cli from sheet-skill-spec
MR !7 2c298b6):
- shortcuts/sheets/data/flag-defs.json: --options Required flips to xor on
+dropdown-set/-update; new --source-range row gains long-form description
pointing at server data_validation.range + the XOR semantics.
- skills/lark-sheets/references/lark-sheets-write-cells.md: "Dropdown 配色"
section reorganized into "Dropdown 选项 + 配色" — XOR comparison table
(list vs listFromRange), shared config flag table (--highlight /
--colors), explicit length rule covering both modes, side-by-side
minimal examples, server-range-normalization gotcha callout.
- skills/lark-sheets/references/lark-sheets-batch-update.md pointer updated
to mention both modes + that +dropdown-delete is unaffected.
PPE smoke (ppe_lark_cli_sheet) on UFJxszjrZhZ1LVtc9FdcICSbn6b C column:
- +cells-set C1 → "性别" (bold + centered): updated_cells_count=1
- +dropdown-set --range C2:C21 --source-range "Sheet1!T1:T3" --colors
'["#cce8ff","#ffd6e7","#e6e6e6"]' --highlight: updated_cells_count=20
- read-back: data_validation.type=listFromRange + range=$T$1:$T$3 (server
normalizes the prefix away on storage; highlight_colors /
enable_highlight not echoed by get_cell_ranges, see byted-sheet read
projection TODO).
- error-path replay (both XOR violations + colors > source-size) all
rejected at Validate stage with the expected messages.
* docs(sheets): sync agent-voice rewrite of Dropdown 选项+配色 from spec
Mirrors sheet-skill-spec MR !7 60df610 — narrative now describes how the
flags interact (XOR, colors length rule, highlight gating, sheet-prefix
read-back gotcha) without exposing the underlying data_validation field
names or server-side normalization details that agents don't act on.
No Go-side change, no shortcut behavior change.
* chore(sheets): restore --colors in parseJSONFlag docstring example list
The earlier commit 49104ec swapped --colors out of parseJSONFlag's "Used
by" example list when it deleted the flag (item #2 there removed --colors
/ --highlight from +dropdown-set/-update). Subsequent commits 8672d8e /
538eb2e / fb90c8b reinstated --colors (and added --source-range) but did
not roll back this docstring tweak — leaving an orphan reference to
--properties where --colors used to be.
This restores the example list to its pre-49104ec form so the docstring
matches what the helper actually services on this branch's HEAD.
Pure docstring change — function behavior unaffected, no test movement.
* fix(sheets): post-rebase test fixups after dropping superseded fix#1
Two test fallouts from rebasing onto upstream 4be06c8 (which independently
re-fixed +workbook-create and +dim-move with a more thorough approach):
- shortcuts/sheets/lark_sheet_workbook_test.go: our PR's earlier
TestWorkbookCreate_DryRun "with headers and data → 2-step plan" subtest
asserted the expedient sheet_name="Sheet1" / no-sheet_id wire body that
matched our dropped fix#1 implementation. Upstream's fix#1 resolves
the workbook's first sheet via get_workbook_structure and fills with
the real sheet_id instead. Reset this file to upstream's version — our
superseded assertions disappear, upstream's tests cover the new wire
shape.
- shortcuts/sheets/execute_paths_test.go: TestExecute_RangeSort fixture
still used the legacy {col, order} sort-key shape because the rebase
resolution picked the upstream version of this file wholesale (it
contained other unrelated changes). Re-apply just the one fixture
update to {column, ascending} so fix#5's CLI-side rejection logic
exercises a valid input — server-side sort_conditions has required
fields `column` (string) and `ascending` (bool); the historical
{col, order} vocabulary was never accepted.
go build ./... + go test ./shortcuts/sheets/... -count=1 both green.
* feat(sheets): +dropdown --highlight tri-state via Changed() for opt-out
The server-side default for data_validation.enable_highlight flipped from
false to true (aligning with the UI behavior). With the previous code path
if runtime.Bool("highlight") { dv["enable_highlight"] = true }
omitting --highlight and passing --highlight=false both produced the same
"enable_highlight key absent" body, leaving CLI users with no way to opt
out of the (now-default) highlighting.
Switch to runtime.Changed() so the translator can distinguish all three
input shapes:
- omitted -> no enable_highlight key (server applies default=true)
- --highlight=true -> enable_highlight: true (explicit no-op vs default)
- --highlight=false -> enable_highlight: false (the only opt-out path)
flagView already exposes Changed() and mapFlagView (the +batch-update
sub-op adapter) implements it via raw-key presence — same pattern other
translators use for "Changed-only" branching (e.g. omit target_index
unless --index was set), so no interface surface change is needed.
Test coverage:
- TestDropdownSet_HighlightTriState pins all four shapes (omit / presence
form / explicit true / explicit false) and asserts the enable_highlight
key's presence/value
- TestBatchOp_BodyMatchesStandalone adds a --highlight=false sub-op case
so the batch sub-op path produces a body byte-identical to the
standalone +dropdown-set --highlight=false body
* chore(sheets): sync +dropdown flag desc + write-cells narrative from spec
Mirror sheet-skill-spec generated/ into shortcuts/sheets/data/ and
skills/lark-sheets/ for the +dropdown-set / +dropdown-update path. No
hand edits in this repo.
The +dropdown flag desc and the Dropdown 配色 narrative now match the
server-side enable_highlight default flip (true) and the tri-state
--highlight semantics introduced in the sibling commit:
* --highlight desc: 不传 = 开(按内置 10 色色板循环上色),
--highlight=false 关闭得到纯白下拉
* --colors desc: 单独传即生效(高亮默认开),--highlight=false 时忽略
* write-cells reference: 三种意图三条线(默认色板 / 指定颜色 /
纯白下拉)+ 新增 --highlight=false 示例
Source upstream: sheet-skill-spec MR !8.
* fix(sheets): validate +cells-set-image --image path in Validate
The unsafe-path check only ran at Execute (via FileIO.Stat), so --dry-run
printed a misleading success preview for an absolute / out-of-cwd --image
path that a real run would then reject. Move the path-safety check into
Validate (validate.SafeLocalFlagPath), so --dry-run and Execute fail
identically and both name the real --image flag. File existence stays
deferred to Execute, so legitimate relative paths still preview cleanly.
Add TestCellsSetImage_DryRunRejectsUnsafePath.
* feat(sheets): support local --image in +float-image-create
+float-image-create now accepts a local file via --image (XOR with
--image-token / --image-uri): the CLI uploads it as a sheet_image and
embeds the returned file_token, removing the previous "upload elsewhere
to get a token first" workaround. Path safety is checked in Validate,
--dry-run previews the extra upload step, and +batch-update rejects
--image (no upload phase). +float-image-update is unchanged (it does not
register --image).
Also syncs the lark-sheets skill docs/flag-defs from sheet-skill-spec:
the new --image flag, partial-merge / border-per-side / bare sheet-prefix
clarifications, and refreshed dropdown --colors/--highlight descriptions
(already pending in the source Base table).
* fix(sheets): +dropdown-get accepts --sheet-id/--sheet-name + bare --range
Align +dropdown-get with its get_cell_ranges siblings (+cells-get / +csv-get):
sheet selection is now via --sheet-id / --sheet-name (XOR) and --range is a
bare A1 reference. The previous shape required the sheet prefix inside --range
(e.g. "Sheet1!A2:A100") and was the odd one out among the read-data wrappers;
callers pasting the sheet-id form straight from the URL hit a misleading
"sheet not found, sheetId: , sheetName: <id>" error because the prefix was
unconditionally treated as sheet_name.
Flag schema + skill reference regenerated from the upstream Lark Base
Shortcut-flags table.
* fix(sheets): drop Sheet1! prefix from +cells-get / +csv-get / +csv-put flag examples
Server tools-schema.json for get_cell_ranges, get_range_as_csv and set_range_from_csv
does not accept a sheet prefix on --range / --start-cell; the sheet is selected via
--sheet-id / --sheet-name. +csv-put --start-cell also now states it must be a single
cell (no range notation).
Synced from spec repo.
* feat: 把环境变量提交上去
* fix(sheets): clarify batch --ranges prefix must be sheet display name
E2E test cases repeatedly trip on this:
$ lark-cli sheets +cells-batch-set-style \
--ranges '["7f8fba!A2:B3","7f8fba!C2:D3"]' --font-color '#3366FF' ...
→ tool "batch_update" failed: [900015206]
sheet "7f8fba" not found. Available sheets: [{id: "7f8fba", name: "Sheet1"}]
Callers paste the hex sheet-id (e.g. "7f8fba") from a spreadsheet URL /
+sheet-create response straight into the --ranges sheet prefix. The four
batch shortcuts (+cells-batch-set-style / +cells-batch-clear /
+dropdown-update / +dropdown-delete) fan each range out into a
batch_update sub-op (set_cell_range / clear_cell_range) and pass the
prefix through as sheet_name; the server only matches sheet_name
literally, so the lookup fails.
The set_cell_range tool schema is explicit: sheet_id is the
reference_id and "must be correct or it errors"; sheet_name is the
display name. CLI can't disambiguate purely from the literal because
users can rename sheets to anything (including six-char hex strings).
Cleanest fix is at the source: each batch shortcut's --ranges flag
description now states explicitly that the prefix must be the sheet
display name and that the sheet reference_id is rejected, so agents
reading the reference don't try the id form in the first place.
No Go changes; these files are regenerated from the upstream Lark Base
Shortcut-flags table via the sheet-skill-spec sync chain.
* docs(sheets): sync lark-sheets skill docs from upstream spec
- SKILL.md: clarify --url only resolves /sheets/ and /spreadsheets/ links; /wiki/ links must be resolved via wiki +node-get first (confirm obj_type=sheet, use obj_token)
- formula-translation: document IMPORTRANGE cross-workbook limits (max 5-level nesting, 100 refs per sheet)
- write-cells: document rich_text cells for hyperlinks, @mentions and @docs
* feat: 同步 tools-schema.json 改动
* fix(sheets): warn when +dropdown source-range exceeds 2000 cells with highlight on
byted-sheet's ListFromRangeValidation.checkOptionsValid() sets
isOptionError=true when shouldHighlightValidData is on and the source
range exceeds LIST_WITH_COLOR_MAX_COUNT (2000 cells) — the highlight +
large source combo is unsupported. CLI previously had no signal for
this, so users only learned by seeing the dropdown render as
option-error in the workbook.
Add a Validate-phase stderr warning in +dropdown-set and +dropdown-update
when --source-range covers >2000 cells unless --highlight=false. Soft
warning, never blocks the request. Inline --options is not subject to
this limit — server enforces no count or per-item length cap on inline
lists, so no warning fires there.
* docs(sheets): sync lark-sheets skill from spec — dropdown flag descs reflect server reality
Pulls sheet-skill-spec canonical-spec → generated → consumers chain for
dropdown flag desc corrections committed upstream (Shortcut-flags base
table rows for +dropdown-set / +dropdown-update --options and
--source-range).
Aligns flag descs with byted-sheet behavior:
- --options: dropped fabricated "≤500 items, each ≤100 chars, no commas"
promise. byted-sheet ListOfItemValidation enforces none of these.
- --source-range: appended note about the only real cap —
LIST_WITH_COLOR_MAX_COUNT=2000 when --highlight is on (server flags the
dropdown as option-error beyond that; CLI warns at Validate time per
bb7ccae).
Also picks up an unrelated upstream tools-schema.json drift (chart float
block schema + data_validation.items description tweak) that surfaced
via npm run check:tool-schemas; bundling keeps the spec sync gate green.
* revert(sheets): drop tools-schema drift mirror from previous spec sync
930c9c7 顺带 sync 了 spec 的 tools-schema bundling — 跟那条 commit 一起
误带进来 chart float block required 和 data_validation.items 描述微调,
这两处其实是上游 sheet-ai-skills 还在 pending 的 revert。
配套 sheet-skill-spec 的 revert commit (a3aa9f2 on
fix/dropdown-flag-desc-real-limits / !11),重跑 sync:consumers 拉回
正确的 generated mirror:
- shortcuts/sheets/data/flag-schemas.json(chart 部分)
- skills/lark-sheets/references/lark-sheets-{chart,batch-update,write-cells}.md(rendered schema 段)
dropdown 文案改动(flag-defs.json 4 处 desc + dropdown 段的 reference
渲染)不在本 commit 范围,保持 930c9c7 的状态。
* docs(sheets): sync lark-sheets skill from spec — +filter-view-update --properties desc
去掉 +filter-view-update --properties 描述里"pass at least one of
--properties.rules / --range / --view-name"的误导承诺。--properties
实际是硬必填(MarkFlagRequired),且 update 走 PUT 整组覆盖语义。
* fix(sheets): align +cells-search/+cells-replace option keys with server schema
The CLI emitted `options.regex` and `options.include_formulas`, but the
server-side `search_data` / `replace_data` tool schemas declare and
consume `use_regex` and `match_formulas`. Result: passing `--regex` or
`--include-formulas` always failed with `unexpected property ... is not
defined in schema`.
Keep the user-facing flag names (`--regex`, `--include-formulas`) — only
the JSON keys sent to the server change. Updates the dry-run test that
locked the wrong contract.
* docs(sheets): sync float-image reference from spec — fix non-runnable examples
Two examples in skills/lark-sheets/references/lark-sheets-float-image.md
didn't actually run against PPE; sync brings them in line with CLI behavior:
- +float-image-create local-path example missed --image-name (CLI rejects
with `required flag(s) "image-name" not set` even when path basename
already has the filename). Add `--image-name "logo.png"` + inline note.
- +float-image-update "only change position" example missed image source
(CLI rejects with `one of --image, --image-token, or --image-uri is
required`). Expand to two steps: list with --jq pulls the current
image_token, then update re-passes --image-token to satisfy the guard.
- Leading warning realigned: image source is mandatory on every update
call; "keep original image" still requires passing the token explicitly.
Upstream change: sheet-skill-spec MR fix/float-image-reference-examples.
* feat: 同步 tools-schema.json 改动
* fix(sheets): allow +float-image-update to omit the image source
The image source (--image-token / --image-uri) is the only optional part
of an update: omit all of them to keep the current image. image_name,
position and size stay required — the manage_float_image tool rejects an
update without them, and +float-image-list does not return image_name to
backfill. Previously the shortcut forced an image source even when only
position/size changed, so those updates were rejected CLI-side before any
API call (reported as a Fail case in the sheets e2e rerun).
- floatImageProperties: gate the image-source requirement on create only;
keep image_name/position/size required on both; emit image_uri only when set
- sync flag-defs.json + lark-sheets-float-image.md from sheet-skill-spec
(image-name/position/size now required on +float-image-update)
- tests: cover the image-source-optional dry-run; the single-required checks
move to the +batch-update sub-op path (cobra owns the standalone path)
* docs(sheets): sync lark-sheets skill from spec
Mirror the canonical-spec reference fixes into the consumer skill:
- search_replace output contract: `matches[]` with `address` (+ `has_more`/`next_offset`)
- workbook sheet fields: `sheet_name`/`is_hidden`/`*_count`, no `frozen_*`
- `+range-fill` example uses a non-overlapping target (A3:A100)
- drop the unimplemented `envelope.meta.verification` auto-readback claim; advise
manual list/get verification instead
* fix(sheets): allow +pivot-create to omit both sheet selectors
manage_pivot_table_object treats sheet_id / sheet_name as the placement
target — when both are absent, handleCreate() auto-creates a new sub-sheet
to host the pivot table. The CLI's flag schema didn't reflect this:
- Exposed a third flag --target-sheet-id that mapped to the same wire
field as --sheet-id, leaving the caller unsure which one to use
- --sheet-id / --sheet-name had "XOR with the other" descriptions that
read like "operation context", so callers (especially LLM tool callers)
felt obligated to set one — frequently the source sheet — which
silently disabled the backend's auto-create guardrail and dropped the
pivot at A1, overlapping the source data
Wire change (synced from sheet-skill-spec): drop the duplicate
--target-sheet-id flag; rewrite --sheet-id / --sheet-name descriptions
to make the placement-target semantics explicit and call out that
omitting both is the recommended path.
Implementation change (this PR): add an at-most-one sheet-selector
helper and let object create-shortcuts opt into it.
- helpers.go: new optionalSheetSelector (both empty allowed; both set
still rejected; control-char validation unchanged). requireSheetSelector
is untouched — every existing caller keeps the exactly-one contract.
- lark_sheet_object_crud.go: objectCRUDSpec gains
allowEmptySheetSelectorOnCreate; objectCreateInput dispatches to
optionalSheetSelector when it's set. Only pivotSpec opts in;
chart / cond-format / sparkline / filter-view / float-image keep
the existing require semantics. DryRun and Execute switch to direct
flag extraction (same pattern Validate already used) so the XOR
check happens in exactly one place (the builder).
- pivotSpec: drop the enhanceCreateInput branch that read the now-removed
--target-sheet-id flag.
- Tests: TestPivotCreate_SheetSelectorSemantics covers both-empty /
both-set / single-set; TestObjectCreate_RequiresSheetSelector
regresses chart / cond-format / sparkline / filter-view to lock the
scope of the relaxation.
* docs(sheets): clarify filter/filter-view rules update is whole-set PUT
Synced from upstream tools-schema. The rules field on manage_filter_object and manage_filter_view_object now documents update as whole-set PUT semantics: submitted rules become the complete rule set, all existing columns' rules are cleared first, columns not listed lose their old rules (no merge), and [] clears everything. Description-only change, no structural/field change.
* refactor(sheets): switch dim-* / rows-cols-resize to A1-string range schema
The 9 row/column-region shortcuts used to share two int flags --start /
--end with inconsistent end semantics across commands — +dim-insert /
-delete / -hide / -unhide / -group / -ungroup treated --end as exclusive,
while +dim-move / +rows-resize / +cols-resize treated it as inclusive.
The skill reference even called this out as "the highest-frequency
off-by-one source", patched in docs rather than at the surface. Three
underlying tool schemas (position+count, A1 range string, 0-based int
pair) were all flattened onto the same --start/--end pair, which forced
a different normaliser per command and pushed mental math (count =
end - start) onto every caller.
Schema (sourced from base, regenerated via sheet-skill-spec, mirrored
into shortcuts/sheets/data/ and skills/lark-sheets/):
+dim-insert --position + --count
rows: "3"; columns: "C". --count rows/columns
inserted *before* --position.
+dim-delete / -hide / -unhide / -group / -ungroup
--range
+rows-resize / +cols-resize --range
A1 closed range. Rows: "3:7" or "5". Columns: "C:F" or "C".
Mixing letters and digits in one range is rejected.
+dim-move --source-range + --target
--target must match --source-range's dimension (both row or both
column). The move places the source block *before* --target.
Wire-shape preserved: modify_sheet_structure still receives `position`
+ `count` (insert) or a `range` A1 string (other dim-* ops); v3
move_dimension still receives 0-based inclusive ints (CLI parses the
A1 strings into them); resize_range still receives a two-sided A1
range (single-element form is expanded to "N:N" before send).
This is a flag-surface break (--start / --end / --dimension flags
removed from these 9 shortcuts); --dimension stays only on +dim-freeze
since it has no range to derive from.
Code: A1 parser added (parseA1Range / parseA1Position /
letterToColumnIndex reused from write_cells); dimRange / dimRangeFull /
dimPosition deleted; dim-move switches to source-range + target parsing;
resize gains a same-dimension guard so +rows-resize rejects "A:C" with
a clear "+rows-resize expects row numbers" message.
Tests: TestSheetStructureShortcuts_DryRun / TestDimMove_DryRun /
TestDimMove_Column / TestDimMove_MismatchedDimension /
TestDimRange_Validation / TestParseA1Range / TestResize_TypeAndSizeGuards
/ TestRangeOperationsShortcuts_DryRun all rewritten against the new
schema. Batch contract trio (BodyMatchesStandalone /
ErrorEquivalence / RejectsBadSubOpInput) and
TestBatchOp_DispatchCoversReportedBugs likewise. Full
`go test ./shortcuts/sheets/` passes.
* docs(sheets): sync +pivot-create placement reference from spec
Companion sync from sheet-skill-spec — the canonical reference rewrites
+pivot-create's "5 placement-related flags" rundown into a clearer
"4 placement-related flags" form (--target-sheet-id was already removed
in #1130, this updates the prose accordingly), and clarifies that
--sheet-id / --sheet-name on +pivot-create are the *placement* sheet
(not the source-data sheet), with omit-both as the strongly-recommended
default.
Also picks up a base-side --target-position description tweak that
dropped the now-stale "与 --target-sheet-id 配套" reference.
No CLI surface change.
* docs(sheets): sync +pivot-create summarize_by lowercase enum values from spec
* docs(sheets): wrap sheet names in single quotes in A1 examples
Synced from spec. Affects 3 reference md (pivot-table / batch-update /
write-cells) and 2 generated flag-data JSONs.
A1 examples like `Sheet1!A1:D100` now read `'Sheet1'!A1:D100` so models
default to single-quoted sheet names. Excel A1 notation requires single
quotes for sheet names containing hyphens / spaces / non-ASCII chars;
always-quoting is also valid for plain names, so this is the safer default
to teach.
Affected flags:
- +pivot-create --source
- +dropdown-update --ranges / --source-range
- +dropdown-delete --ranges
- +dropdown-set --source-range
- +cells-batch-set-style --ranges
- +cells-batch-clear --ranges
* docs(sheets): wrap A1 sheet names in handwritten examples + bash histexpand guide
Synced from spec. Affects 4 reference md (chart / pivot-table / sparkline /
write-cells) and SKILL.md.
In addition to wrapping sheet names in single quotes in all remaining
handwritten examples (covers chart refs.value / nameRef, sparkline source,
write-cells --source-range, pivot-create narrative), SKILL.md gains a new
"Shell quoting for A1 references with !" section.
The new section addresses bash history expansion: in interactive bash
(e.g., ShellExec sandbox), unescaped `!Word` after `"..."` triggers
`bash: !A1: event not found`, dropping the command before lark-cli sees
it. The section gives 4 quoting strategies (shell single-quote outer,
`set +H` prefix, mixed quoting, sheet-rename fallback) and an anti-pattern
list.
Affected files:
- skills/lark-sheets/SKILL.md (new section)
- skills/lark-sheets/references/lark-sheets-chart.md
- skills/lark-sheets/references/lark-sheets-pivot-table.md
- skills/lark-sheets/references/lark-sheets-sparkline.md
- skills/lark-sheets/references/lark-sheets-write-cells.md
* docs(sheets): drop bash histexpand section, fix write-cells table escape
Sync from spec, refining the bash-quoting deep-dive added in 0f695b6:
- Drop the `## Shell 调用注意事项` section in SKILL.md and the inline
`⚠️ bash 引号` callouts in lark-sheets-pivot-table.md and
lark-sheets-write-cells.md. The 4-scenario quoting table + anti-pattern
list turned out too verbose for the SKILL intro; single-quoted examples
in the references are themselves enough nudge.
- lark-sheets-write-cells.md L146: fix the table cell escape from the
malformed `'''Sheet1''!T1:T3'` (consecutive `''` are no-op empty
strings) to `''\''Sheet1'\''!T1:T3'`, matching the bash example at
L191 verbatim.
Net: 1 insertion, 40 deletions across 3 files.
* feat(sheets): rename +pivot-create sheet selector → --target-sheet-{id,name}
+pivot-create's placement selector (where the pivot table lands) is no
longer the generic --sheet-id / --sheet-name; it is now
--target-sheet-id / --target-sheet-name. The new names mark this as the
*output* sheet, distinct from the *data-source* sheet (which lives
inside --source as `'Sheet'!Range`). The other +pivot-{list,update,delete}
shortcuts keep --sheet-id / --sheet-name (their semantics are
"sheet that hosts the existing pivot", same as every other shortcut).
Motivation: an LLM agent reading the previous CLI surface saw +pivot-create
expose --sheet-id and assumed (as it had to) that it pointed at the data
source, like every other shortcut. The new flag name makes the intent
unambiguous at the call site, without relying on the agent having read
the narrative caveat in the reference doc.
Background: evaluation case U046 spent multiple rounds tripping on this
exact confusion before working around it with +sheet-rename.
Implementation:
- objectCRUDSpec gains createSheetIDFlag / createSheetNameFlag (with
default-fallback accessors sheetIDFlagOnCreate / sheetNameFlagOnCreate);
newObjectCreateShortcut + objectCreateInput consult the spec instead of
hard-coded "sheet-id" / "sheet-name". pivotSpec sets target-sheet-*;
every other create spec inherits the defaults.
- optionalSheetSelector (only used by pivot create) takes the two flag
names as parameters so its mutex / control-char errors quote the names
the user actually typed (--target-sheet-id, not --sheet-id).
- batch_op_dispatch: introduce sheetSelectorFlagsForSubOp(shortcut) →
(idFlag, nameFlag) returning target-sheet-* for "+pivot-create" and
the defaults otherwise; translateBatchOp uses it so +pivot-create
sub-ops in +batch-update accept the same renamed input keys.
- Tests:
- lark_sheet_object_crud_test.go: pivot-create cases switch args and
expected error wording to target-sheet-*; extra assertion that the
mutex error quotes the renamed flag (regression guard against
flag-name drift between code and error message).
- batch_op_contract_test.go: +pivot-create sub-op test uses
target-sheet-id / target-sheet-name input keys; the body-vs-standalone
contract loop reads the selector via sheetSelectorFlagsForSubOp so
every other shortcut keeps using sheet-id / sheet-name.
Synced reference docs (skills/lark-sheets/{SKILL.md,
references/lark-sheets-pivot-table.md}) mirror the spec's new flag names,
narrative, 3-placement-strategy block, and SKILL.md exception bullet that
explains why +pivot-create's badge says 无 sheet 定位 yet still has
placement selectors (just under different names).
flag-defs.json synced from spec picks up the renamed flags + kind=own.
All sheets-package tests pass.
* docs(sheets): strip migration-history language from pivot reference / SKILL
Synced from spec. Removes "renamed from / no longer called / not
--sheet-id" style migration-history language that snuck into the
previous sync. Reference and SKILL now describe the current flag names
directly without referencing the old names.
* docs(sheets): require +workbook-info before guessing sheet name
Synced from spec. SKILL.md adds a new rule under the sheet-locator
section: unless the user has explicitly named a sheet, the agent must
call +workbook-info first to fetch sheets[].sheet_id / sheets[].title
rather than guessing the default `Sheet1`. The Chinese-language tables
this CLI is typically used against rarely use that literal name —
"数据" / "Sheet" (no digit) / "工作表 1" / business-named sheets are
far more common — so guessing wastes a round-trip before the agent
ends up calling +workbook-info anyway.
The 统一调用范式 example also switches its `--sheet-name "Sheet1"`
placeholder to `<真实表名>` to remove the inadvertent suggestion that
`Sheet1` is a sensible default.
* docs(sheets): tell agent to `set +H` for A1 references containing `!`
Synced from spec. The sheet-locator section now warns: when a flag value
contains `!` (--source / --range / --ranges with a cross-sheet prefix),
run `set +H` at the start of the bash session to disable history
expansion — otherwise interactive bash (e.g. inside an agent's shell
sandbox) lexes "Sheet1!A1" as a history reference and fails with
`event not found` before lark-cli ever sees the argument.
When the sheet name itself contains hyphens / spaces / non-ASCII
characters, the A1 reference also needs single quotes around the sheet
name per A1 notation, e.g. --source "'Sales-2025'!A1:D100".
Also flips the previous `--range` example to `--range 'Sheet1!A1:B2'`
(shell single-quote) for consistency.
* feat(sheets): add schema-driven JSON flag validation
Validate composite JSON flags (--properties, --cells, --options,
--border-styles, --sort-keys) against the embedded flag-schemas.json
on every standalone and +batch-update sub-op invocation, replacing
ad-hoc per-shortcut guards.
Supports the JSON Schema subset actually used upstream: type / enum
/ oneOf / required / properties / items / nullable / minimum /
maximum / minItems / maxItems / additionalProperties (true | false
| <schema>). Enum errors quote the failing value, truncate beyond 8
entries, and surface case-only "did you mean" hints (SUM -> sum).
Coverage: 18 / 19 (shortcut, flag) pairs. +batch-update --operations
stays validator-skipped; its translator already does richer per
sub-op checks. mapFlagView.Command() routes batch sub-ops through
the same (command, flag) -> schema pipeline as standalone.
loadFlagSchemas() is now sync.Once-guarded so parallel first access
from t.Parallel test sets and concurrent shortcut invocations is
race-free.
Removes superseded hand-written guards:
- +pivot-create validateCreateInput / validatePivotCreateProps
- +range-sort sort-keys per-item shape check
Test fixtures updated to be schema-conformant (chart position/size,
pivot summarize_by lowercase, cells 2D-array shape).
* feat(sheets): add --rows-json output flag to +csv-get
+csv-get --rows-json returns structured rows ({row_number, values:{col→cell}})
instead of the CSV string, so callers can address cells by row_number / column
letter without parsing [row=N] or RFC-4180 CSV. Same read, alternate output
shape — a flag on +csv-get (default stays CSV), not a separate shortcut, since
the two differ only in representation.
- CsvGet.Execute: --rows-json reshapes the response via assembleRowsJSON
(parses annotated_csv into per-row records keyed by column letter; every
logical row emitted; embedded newlines parsed into cell values)
- surfaces the under-read hint structurally as data_not_fully_read
- flag-defs.json + read-data reference synced from spec
* feat(cli): agent-friendly errors, proxy silencing, +csv-put --range
Agent-experience fixes distilled from analyzing 50 real sheets
trajectories, where the top failures were hallucinated command/flag
names, proxy warnings corrupting JSON on stdout, and --range carried
over from +csv-get to +csv-put.
- did-you-mean: unify the duplicated Levenshtein into a shared
internal/suggest package and wire its prefix-weighted ranker into
unknown-subcommand and unknown-flag errors; flag-parse errors now
return a structured envelope with suggestions plus the full valid list,
so agents recover from semantic typos (e.g. --query vs --find).
- proxy: suppress the one-time proxy warning in non-interactive
(agent/CI/piped) runs so a 2>&1-merged stderr line cannot corrupt
stdout JSON; interactive sessions still warn.
- sheets +csv-put: accept --range as an alias for --start-cell (parity
with +csv-get / +cells-set) and echo the computed writes_range in
dry-run and the success envelope, so agents see the paste footprint
before it overwrites neighbours.
- docs(sheets): add an intent->command cheat-sheet to SKILL.md, a
runtime-prerequisites section, and document the --range alias and
writes_range behaviour.
* feat(sheets): close P0-4 pivot gaps — enum case, clear→pivot-delete hint, placement warning
Last open P0 from the 50-trajectory analysis — the two pivot black holes:
upper-cased summarize_by, and pivots built over the source sheet that hit
#REF! and then couldn't be removed.
- enum case tolerance: validateAgainstSchema rewrites a case-only enum
mismatch to the canonical (lower-case) spelling in place ("SUM" -> "sum")
before the request is sent, killing the whole class instead of only
hinting at it. Covers every nested enum (values[], calculated_fields[]);
genuinely unknown values still fail with the existing did-you-mean message.
- +cells-clear / +cells-batch-clear: when the backend reports "can not find
embedded block" (the range overlaps a pivot/chart), annotate the error
with the real fix — clearing cells can't delete an embedded object; remove
it with +pivot-delete / +chart-delete (id via +pivot-list / +chart-list).
Applied to both shortcuts, a Tips line, and the cells-clear reference.
- +pivot-create: a --help Tips block making "omit --target-* -> backend
auto-creates a sub-sheet, zero overwrite" the can't-miss default, plus a
placement_warning (dry-run + execute output) when an explicit target sheet
is set with no offset — definite when the target name matches the source
sheet, conditional otherwise. Local-only, advisory, never blocks the call.
The placement_warning is structured output, not a stderr line, so it
survives non-interactive proxy-warning silencing and isn't swallowed by 2>&1.
* feat(sheets): strip UTF-8 BOM from stdin/@file flag input
resolveInputFlags now strips a leading UTF-8 BOM from content read via stdin
or @file, so it cannot corrupt the first CSV cell or break JSON parsing of
payloads like --operations / --cells downstream.
Also pulls the synced lark-sheets skill docs from sheet-skill-spec and drops
scheme-number tags from two test comments.
* fix(sheets): drop dead --value-render-option flag from +csv-get
+csv-get wraps get_range_as_csv, which has no value_render_option support
(absent from its input type, executor, and published tool schema — it always
returns formatted display text via getText()). The CLI passed the flag through
as a silent no-op: callers asking for raw_value/formula got formatted values.
Remove the flag from flag-defs, drop the value_render_option passthrough in
csvGetInput, and clean the stale SKILL references. The real value_render_option
capability is unchanged on +cells-get (get_cell_ranges) via --include formula.
* chore: rename ppe x-tt-env lane to ppe_moa_canvas
* docs(sheets): sync skill description from spec (cloud-drive alias, lark-drive search, doubao routing)
* feat(sheets): restore pre-refactor shortcuts under backward/ for compatibility
The lark-sheets refactor renamed every shortcut (verb-noun → noun-verb,
e.g. +create-sheet → +sheet-create) and dropped the old commands. External
callers and the tests/cli_e2e/sheets suite still drive the legacy command
names (+create, +read, +write, +create-sheet, ...), which broke.
Re-add the pre-refactor implementations verbatim from main as an isolated
shortcuts/sheets/backward package (package rename only) and register
backward.Shortcuts() alongside sheets.Shortcuts(). Both sets mount under the
`sheets` service; their command names are fully disjoint (38 new vs 42 old,
zero overlap), so old and new commands coexist without collision.
* fix(sheets): resolve 30 golangci-lint v2.1.6 issues — copyloopvar, nilerr, unused
Removed 25 Go 1.22+ loop variable copies (copyloopvar) from test files where
tc := tc / tt := tt / c := c are no longer needed. Fixed 4 nilerr false
positives in flag_schema_validate.go by making intentional error discards
explicit (schema validation failures skip silently — best-effort guard).
Dropped unused batchOpDispatchKeys helper in batch_op_dispatch.go.
* feat(sheets): flag pre-refactor backward aliases via _notice and --help grouping
Nudge users whose lark-sheets skill predates the refactor to migrate off
the pre-refactor aliases (+read, +write, ...), without requiring anyone
to read --help.
- internal/deprecation: process-level pending Notice slot (mirrors
internal/skillscheck), surfaced in the JSON "_notice" envelope under a
"deprecated_command" key.
- internal/cmdutil: shared DeprecatedGroupID cobra group + helper so both
--help rendering and the unknown-subcommand path classify aliases the
same way.
- shortcuts/register.go: applySheetsCompatGroups splits the aliases into a
dedicated "update your skill" help group with "(-> +new)" pointers;
wrapSheetsBackwardDeprecation records the notice from Validate/Execute so
direct callers that never read --help still get flagged.
- cmd/root.go: extract composePendingNotice (now unit-testable) and split
availableSubcommandNames into current vs deprecated buckets while still
ranking unknown-subcommand suggestions across both.
* chore: drop hardcoded ppe lane routing from base security headers
The x-tt-env/x-use-ppe headers forced every request onto the
ppe_moa_canvas pre-release lane; they were only meant for exercising the
sheets refactor against the staging backend. Remove them so the CLI
routes to production by default.
* chore(sheets): promote lark-sheets skill to 2.0.0
Drop the -draft suffix now that the refactored sheets skill is ready to
ship.
* fix(sheets): correct +dropdown-get sheet-locator doc, finalize skill to 2.0.0
+dropdown-get requires a mandatory sheet selector — its Validate calls
resolveSheetSelector — so drop it from the "no sheet locator" exception
list in SKILL.md. It was wrongly grouped with +dropdown-update/+dropdown-delete,
which take only --ranges. +dropdown-get's own per-shortcut badge (公共四件套)
was already correct. Also finalize the skill version 2.0.0-draft -> 2.0.0.
* fix(sheets): enforce required-flag contract in batch sub-ops
Batch sub-ops reuse each shortcut's shared *Input builder through mapFlagView,
which seeds flag-defs defaults — so any required check that lives OUTSIDE the
builder (cobra MarkFlagsOneRequired, or a shortcut's own Validate) is silently
bypassed and the default value wins. Two gaps surfaced in PR review:
- +csv-put: with neither --start-cell nor --range set, start-cell's "A1"
default won and the paste silently anchored at A1. Require an explicit anchor
(guard on Changed, mirroring the standalone MarkFlagsOneRequired).
- +sheet-move: --index (plus >=0 bounds for index / source-index) was not
enforced in the batch path; a missing --index silently moved the sheet to the
front. Mirror SheetMove.Validate.
Also from the same review:
- +batch-update: an explicit --continue-on-error=false now wins over an
--operations envelope's continue_on_error:true (guard on Changed, not value).
- validateDropdownRanges rejects malformed sheet!range ("!A1", "Sheet1!",
"Sheet1!bad") at Validate instead of deferring to the server.
Tests added/updated for each path; full sheets suite green.
* fix(cli): surface skill in deprecated_command notice
deprecation.Notice carries Skill, but the _notice.deprecated_command payload
dropped it, forcing callers to parse `message` to learn which skill to update.
Emit `skill` when set, alongside the existing `replacement`.
* fix(sheets): harden batch type-checking and +workbook-create edge cases
From the branch code-review doc (3 findings):
- +batch-update sub-ops: `operations` is skipped by parse-time schema
validation and mapFlagView coerces a type-mismatched scalar to its zero
value, so "index":"abc" or "multiple":"true" silently became 0 / false and
wrote to the wrong place. translateBatchOp now runs validateRawTypes, which
checks each sub-op scalar against its flag-defs type and rejects mismatches.
- +workbook-create with empty arrays: buildInitialFillInput returned (nil,nil)
for empty rows while the caller wrote fill["excel_id"] unconditionally, so
--values '[]' panicked on a nil map and --headers '[]' produced an illegal
"A1:1" range. It now also returns nil when no cells survive (maxCols==0
guard) and Execute/DryRun skip the fill when fill==nil.
- +workbook-create partial failure: after the spreadsheet was created, a
first-sheet lookup or fill failure returned a bare fmt.Errorf, losing the new
token. It now returns a structured partial_success error carrying
spreadsheet_token in the detail so callers can retry or clean up.
Tests added for each path; sheets suite green.
* fix(cli): structured errors for unknown flags, print-schema, deprecated aliases
From the branch code-review doc (3 findings):
- pure-group UnknownFlags: installUnknownSubcommandGuard whitelists unknown
flags so a mistyped subcommand still reaches the suggestion path, but a lone
unknown flag before any subcommand (`sheets --badflag`) was swallowed and the
group fell through to help + exit 0. unknownSubcommandRunE now recovers the
swallowed tokens (from os.Args captured at Execute entry) and fails with a
structured unknown_flag error; a misplaced but known flag (e.g. --format)
still prints help.
- deprecated-alias notice: a backward-compat alias that fails a cobra-level
required flag short-circuits before RunE, so the Validate/Execute-wrapped
deprecation notice was dropped. Added Shortcut.OnInvoke, fired from PreRunE
(ahead of ValidateRequiredFlags); and the root legacy error fallback now
routes through the structured envelope when a deprecation is pending so the
migration hint survives. Non-deprecated errors keep the plain output.
- --print-schema: runShortcut returned the bare error from PrintFlagSchema. It
is now wrapped as a structured output.ExitError (type print_schema_error) so
agent introspection can parse the failure.
Tests added for each path; cmd + sheets suites green.
* fix(sheets): resolve --sheet-name via title + keep bare sheet selectors verbatim
Two review findings on the backward-compat layer:
- lookupSheetIndex matched only sm["sheet_name"], but get_workbook_structure
surfaces the sub-sheet display name as "title". Every --sheet-name path that
relies on the lookup (e.g. +sheet-move) failed to resolve. Fall back to
"title" when "sheet_name" is absent so either field resolves.
- +read / +write / +append fell back to --sheet-id when --range was omitted,
then routed that bare sheet id through the range normalizer. A sheet id that
looks A1-ish (letters+digits, e.g. "shtABC123") got mangled into
"shtABC123!shtABC123:shtABC123". Split the sheet-only path from the
range-normalization path: read/append pass the selector through verbatim,
write builds the rect from the selector's A1.
Regression tests added for both paths; sheets suite green.
* fix(sheets): silence nilerr/copyloopvar lint in batch type-check additions
- flag_view.go: annotate the fail-open return in validateRawTypes with
//nolint:nilerr (matches the repo convention for intentional fail-open).
- execute_paths_test.go: drop the redundant tc := tc copy (Go 1.22+ scopes
the loop var per iteration).
* test(sheets): data-driven required-flag parity contract for batch sub-ops
Adds TestBatchOp_RequiredFlagParity, the systematic standalone-vs-batch parity
check the branch review asked for. Data-driven over batchOpDispatch + flag-defs,
it asserts that for every batchable shortcut a +batch-update sub-op which
satisfies the sheet locator but omits the shortcut's business-required flags
fails in translateBatchOp, never silently defaulting.
This generalizes the hand-picked TestBatchOp_ErrorEquivalence / GuardsBeyondCobra
cases to the full 50-command surface and auto-covers shortcuts added later, so a
future refactor that moves a required check out of the shared *Input builder
(the failure mode behind the csv-put / sheet-move gaps) is caught here. 45
sub-tests run; locator-only commands (+sheet-delete / +sheet-hide / ...) have no
business-required flag to omit and are skipped. A missing-locator error is also
rejected so a bad fixture can't mask a real gap.
* refactor(sheets): drop unused int64 flag-type plumbing
No sheets flag-def declares an int64 type and RuntimeContext.Int64 had
zero callers, so remove the premature support: the RuntimeContext.Int64
helper, the registerShortcutFlagsWithContext int64 branch, the flagView
Int64 method + mapFlagView impl, and the typedDefault/validateRawTypes
int64 cases. float64 (consumed by --font-size) is kept.
* test(sheets): drop redundant copyloopvar copy in required-flag parity test
Go 1.22+ scopes the loop var per iteration, so `cmd, business := cmd, business`
in TestBatchOp_RequiredFlagParity is a no-op that trips the repo's copyloopvar
linter (same cleanup as 2132472). Behavior unchanged; 45 sub-tests still pass.
* revert(cli): drop non-interactive proxy-warning silencing
WarnIfProxied's interactivity gate is a generic CLI/agent-UX change
unrelated to the sheets refactor / backward-compat scope of this branch.
Split out to a dedicated PR; restore WarnIfProxied to its single-arg form
here (warn.go, warn_test.go, factory_default.go callers).
* docs(sheets): correct +workbook-info output field and batch +sheet-move index requirement
Sync from spec: +workbook-info returns sheet display name as 'title'
(sheet_name only as legacy fallback), and +sheet-move inside +batch-update
also requires --index, not just --sheet-id/--source-index.
* fix(sheets): reject non-integer numbers for batch int flags
validateRawTypes treated int and float64 identically (both only required a
JSON number), but mapFlagView.Int() truncates float64 via int(t), so a batch
sub-op accepted 1.9 for an int flag (e.g. --index) and silently floored it to
1. Standalone cobra rejects non-integer input for int flags at parse time;
enforce the same in the batch path with a math.Trunc check so batch/standalone
parity holds and positional fields can't land on a floored value.
* fix(cli): align flag-before-subcommand unknown_flag detail schema
The flag-before-subcommand recovery path emitted a Type: unknown_flag whose
detail only carried unknown_flags + command_path, diverging from
flagDidYouMean's unknown_flag detail (unknown, command_path, suggestions,
valid_flags). A consumer keyed on Type then saw two shapes for one Type.
Emit the same keys from both paths: add unknown (the offending flag; joined
when multiple), plus empty suggestions/valid_flags — the subcommand isn't
resolved at this point, so there is no meaningful flag universe to suggest
from, and the group's own flags would mislead. unknown_flags is retained as
the authoritative multi-flag field. Test locks the shared schema.
* perf(sheets): compile flag specs to Go to drop startup JSON parse
Every lark-cli invocation (sheets or not) unmarshaled data/flag-defs.json
(122KB) and data/flag-schemas.json (256KB) during package init, before
main(): flag-defs via the shortcut package vars (flagsFor runs at init),
flag-schemas via shortcuts.init() -> Shortcuts() -> commandsWithFlagSchema().
On a 0.5-core sandbox this cold-start cost lands on every command.
Compile both specs to Go at build time instead of parsing at runtime:
- flag-defs.json -> flag_defs_gen.go: flagDefs is a compiled map literal;
loadFlagDefs() returns it directly (no embed, no Unmarshal).
~3.3ms/4110 allocs -> ~0.57ms/539 allocs at sheets package init.
- flag-schemas.json -> flag_schemas_gen.go: only the command-name set
(commandsWithSchema) is compiled in; registration and the validate
fast-path gate on it without touching the 256KB blob. The blob stays
embedded and is unmarshaled lazily only on --print-schema or when
validating a command that has a schema. Removes the 256KB parse from
init entirely.
data/*.json remain the canonical source; *_gen.go are committed, derived
artifacts regenerated with `go generate ./shortcuts/sheets/...`
(shortcuts/sheets/internal/gen). *_gen_test.go guard source/generated drift.
No behavior change: flag rendering, required/enum/default, --print-schema,
and composite-flag schema validation verified unchanged; ./shortcuts/...
tests pass.
* ci(sheets): exempt internal/gen generators from forbidigo
The shortcuts/sheets/internal/gen code generator is a standalone
`package main` run via go:generate, not shortcut runtime code, so the
forbidigo bans on log.Fatal / os.ReadFile / fmt.Printf do not apply.
Making it "compliant" is impossible anyway: a structured error return
needs os.Exit (also banned), and the vfs alternative is blocked by
depguard shortcuts-no-vfs. Exempt shortcut internal/gen paths, matching
the existing _test.go and internal/vfs forbidigo exemptions.
* fix(cli): fail structured on flags before a missing subcommand
A pure group invoked with flags but no subcommand (e.g. `im --format=json`,
`sheets --format json`) silently fell through to help + exit 0, so an agent
could mistake a malformed call for success. The unknown-subcommand guard's
FParseErrWhitelist swallows the flags and leaves RunE with empty args; it now
recovers the raw flag tokens and fails structured:
- unknown flag(s) -> unknown_flag (unchanged)
- valid flag, no subcmd -> missing_subcommand (new, exit 2)
- bare group -> help, exit 0 (unchanged)
Because the group RunE is hook-wrapped, returning a real error also makes
plugin observers record the call as failed instead of ok (the lifecycle Err
is no longer flipped to nil).
Hardening from the same review:
- document the cobra error-text contract unknownFlagName relies on, in
both cmd/root.go and go.mod, so an i18n/reword is caught on upgrade.
- guard the reserved --print-schema/--flag-name registration with a Lookup
so a shortcut declaring same-named flags can't panic pflag.
Tests cover the new missing_subcommand path and the reserved-flag collision.
* fix(cli): don't flag group-valid globals as a missing subcommand
9f8dfa72 made a pure group invoked with flags but no subcommand fail with
missing_subcommand, keying on "any flag defined in the tree". That also matches
inherited global flags (--profile, ...), so `lark-cli --profile p im` and
`lark-cli im --profile p` errored with a misleading "flag --profile belongs to
a subcommand" instead of printing the group's help — a regression, since a bare
group carrying a global flag should print help.
Only treat a flag as missing_subcommand when it is valid on a subcommand but
not on the group itself or inherited (subcommandOnlyFlagTokens). A bare group
carrying only group-valid/global flags falls through to help; flags that
genuinely belong to an omitted subcommand (`im --format json`) still fail
structured, and unknown flags (`im --badflag`) still report unknown_flag.
Test covers a global flag on a bare group resolving to help.
---------
Co-authored-by: zhengzhijie <zhengzhijie.j@bytedance.com>
Input pre-check failures shared by every shortcut — @file/stdin input
resolution, enum validation, and unsupported --dry-run — now leave the
CLI as typed validation envelopes naming the offending flag, so scripts
and AI agents can branch on `param` instead of parsing prose. Wire type,
exit code, and message text are unchanged; the new fields are additive.
The shared layer also gains typed replacements for its legacy
error-producing helpers, so each business domain can migrate to typed
errors without rebuilding common plumbing, and a path-scoped lint guard
keeps migrated domains from sliding back.
Changes:
- Shared pre-check failures (input flags, enum values, dry-run support)
return typed validation errors carrying the offending flag as `param`.
- Every legacy error-producing helper in shortcuts/common has a typed
replacement that preserves the existing message text: validation and
flag-group checks, chat/user ID validation (callers name the flag so
`param` is ground truth), "me" open-id resolution, safe-path checks,
input-stat and save-error wrapping. Legacy helpers stay for
not-yet-migrated domains, marked deprecated — including the legacy
API-result classifier, whose typed route is runtime.CallAPITyped.
- A new errscontract rule rejects legacy common-helper calls on migrated
paths, so a migrated domain cannot silently reintroduce legacy
envelopes; drive is the first locked path and its last legacy
ID-helper calls are replaced.
The card message converter (shortcuts/im/convert_lib/card.go) previously
rendered a subset of card fields and had several mode-gated behaviors that
caused information to be silently dropped in concise mode. This PR audits
every element handler and brings the output up to full fidelity:
missing header fields are rendered, collapsible panels always expand, rich
element metadata (images, audio, video, overflow URLs, person names) is no
longer hidden behind cardModeDetailed, and several format bugs are fixed.
Change-Id: I422474ab6b7505e48ab5697793900df035be6e29
* feat(base): add base block shortcuts
* fix(base): use block scopes for base block shortcuts
* fix(base): split base block shortcut scopes
* docs(base): consolidate base block help
* docs(base): simplify block help wording
* test(base): cover base block shortcut execution
* feat(base): filter base block list by type
* docs(base): clarify base block ids
* docs(base): simplify docx block help
* docs(base): refine base block agent help
Add CLI-side validation for --message-ids in the mail +messages shortcut
to catch obviously invalid inputs before making any API call. The batch_get
endpoint would otherwise only reject malformed IDs server-side, returning
unclear errors.
Validation rules:
- Reject empty message-ids list
- Reject entries exceeding the server-mirrored batch limit of 20 IDs
- Reject entries with leading/trailing whitespace
- Reject entries containing control characters, whitespace, or path separators
- Reject duplicate message IDs
sprint: S2
Improve the --markdown vs --text guidance in the lark-im send/reply reference docs. Reposition --markdown as the recommended default for agents, add explicit selection rules, and reframe the docs around usage scenarios rather than caveats.
* feat(api): add --json flag as no-op alias for --format json
* feat(service): add --json flag as no-op alias for --format json
* feat(shortcut): add --json flag as no-op alias for --format json
Skip registration when a custom --json flag already exists on the
command (e.g. base shortcuts use --json for body input).
Change-Id: If66236cadeea7fa81811061cce775deff51b92ce
* refactor: extract FetchTAT sharing the TAT-rejection classifier
doResolveTAT minted the tenant access token inline. Extract the HTTP call
into FetchTAT(ctx, httpClient, brand, appID, appSecret) so callers that
already hold plaintext credentials — notably the post-config-init probe —
can validate them without a second keychain round-trip.
FetchTAT routes a non-zero TAT body code through the same
classifyTATResponseCode the credential layer already uses, so a rejection is
the canonical CategoryConfig / SubtypeInvalidClient (10003 / 10014) typed
error — identical to what every token-resolving command returns. Transport,
HTTP-status and JSON-parse failures stay raw (untyped) so callers can use
errs.IsTyped to separate a deterministic credential rejection from upstream
noise. doResolveTAT now delegates to FetchTAT; observable behavior unchanged.
* feat: validate credentials after config init
After config init saves the App ID / App Secret, fire a best-effort probe:
mint a tenant access token with the just-saved credentials, then POST the
application probe endpoint. When the credentials are deterministically
rejected, FetchTAT returns a typed errs.* error and runProbe propagates it,
so config init exits non-zero with the canonical ConfigError / invalid_client
envelope (the same one every other command shows for the same bad creds)
instead of letting the user discover the mistake on a later request.
Ambiguous failures (transport, HTTP non-200, JSON parse, timeout,
http-client init) come back untyped and are swallowed (errs.IsTyped is the
discriminator), so a valid configuration is never blocked by upstream noise.
The probe is wired into all four init paths and skipped when the user reused
an existing secret. The saved config is not rolled back on rejection: stdout
still records what was saved, stderr carries the typed error envelope.
Drive-domain errors now leave the CLI as typed, machine-branchable
envelopes — a stable `type` plus `subtype` and named fields (param,
params, retryable, log_id, hint) — so scripts and AI agents can branch on
structure and act on a recovery hint instead of parsing prose.
Changes:
- Every error produced in the drive domain — validation, file I/O, and the
failures returned from its Lark API calls — is emitted as a typed errs.*
error; the exit code is derived from the error category. Drive's API calls
now go through a shared typed classifier, so failures carry subtype,
troubleshooter, a recovery hint, and the request's log_id whether the
server returns it in the response body or the x-tt-logid header; an
already-typed network/auth error is never downgraded into a generic API
error.
- Known API conditions (resource conflict, cross-tenant, cross-brand, ...)
carry a recovery hint keyed by their error class; a command can refine
that hint with command-specific guidance.
- Batch partial failures (+push / +pull / +sync, where some items succeed
and some fail) now report an honest ok:false multi-status result on
stdout — the summary and every per-item outcome stay machine-readable —
and exit non-zero, instead of a misleading ok:true success envelope.
- Duplicate rel_path conflicts report each colliding path as a structured
params entry (RFC 7807 invalid-params style).
- Static guards lock the drive path so legacy error construction — direct
envelopes or the auto-classifying API helpers — cannot be reintroduced,
making drive the template for the remaining domains.
Output changes worth noting for consumers:
- Error envelopes now carry typed type/subtype and named fields; exit
codes follow the error category (malformed or incomplete API responses
are reported as internal errors rather than generic API errors).
- Batch partial failures (+push / +pull / +sync) emit an ok:false result
envelope on stdout (summary + per-item items[]) and exit non-zero; the
per-item results stay on stdout rather than in a stderr error envelope.
Errors surfaced through shared cross-domain helpers (scope precheck, media
import upload, metadata lookup, save-path resolution) are not yet typed;
they migrate with the shared layer in a follow-up change.
Interactive card messages (msg_type: interactive) can contain @user elements in their card
body. The json_attachment.at_users field stores resolved user info, but the user_id there is
the sender-side platform user_id — not the reading app's canonical open_id. When the backend
populates a mention_key on each at_users entry, it signals that the API-level mentions[]
array carries a more authoritative open_id and display name for the reading context. This PR adds
support for this two-level lookup: it threads the raw mentions[] array into the card converter,
indexes it by mention_key for O(1) access, and renders the canonical open_id + display name
whenever the link is resolvable. All existing fallback paths (no mention_key, nil mentions) are
preserved without behavioral change.
Change-Id: I00f846d76482adba315d07361c35909b71ca74c7
Follow-up to #1223. The hand-written FLAGS block in `lark-cli --help`
restated leaf-command flags at the root level — flags that are not
registered on the root command (they error "unknown flag" there). Even
trimmed to an illustrative example list, it duplicated information Cobra's
per-command `--help` already renders authoritatively, and any static list
in root help drifts from the real per-command flag sets over time.
Drop the section entirely: Cobra's per-command `Flags:` output is the
single source of truth. `USAGE:`/`EXAMPLES:` still show flags in context,
and the `Flags:` block at the bottom of root help lists the actual root
flags. Also removes the now-obsolete TestRootLong_FlagsSectionPointsToCommandHelp.
The hand-written FLAGS block in `lark-cli --help` listed --params, --data,
--as, --format, --page-all, --page-size, --page-limit, --page-delay, -o,
--jq, -q and --dry-run as if they were global flags. None are registered
on the root command — they all error "unknown flag" at the top level and
exist only on leaf commands (api, service). The block also contradicted
the Cobra-generated "Flags:" section rendered directly below it, which
shows only -h/--help, --profile, -v/--version.
Replace it with a short illustrative example list (common flags first) and
a pointer to `lark-cli <command> --help` for the full per-command set.
Root help stays a discovery signpost without claiming the flags are global
or restating defaults/descriptions that drift from the real flag sets.
Change-Id: Ia1cab889dd70b6b49a61dac468dedfd7fe39043f
Simplifies the markdown-to-post rendering pipeline in the IM shortcut. The previous
implementation split markdown at blank-line boundaries into multiple post paragraphs,
using zero-width space (\u200B) sentinel characters to preserve visual spacing.
While well-intentioned, this approach introduced fragility around edge cases such as
blank lines inside fenced code blocks, messages with only blank lines, and interactions
with the heading-normalization pass. This change consolidates rendering back into a
single {"tag":"md"} segment, making the output more predictable, the code significantly
easier to follow, and the test surface easier to maintain.
Change-Id: Ic2870ecbcb31ae7d36121f120102f2ff964f5169
* feat: unconditionally inject --format flag for all shortcuts
Removes three HasFormat guards in runner.go so every shortcut
gets --format regardless of the Shortcut.HasFormat field value.
Shortcuts that already define a custom 'format' flag in Flags[]
are skipped to avoid redefinition panics (e.g. mail +triage, +watch).
HasFormat is retained in the struct but marked deprecated.
Change-Id: I5e8fe07e839d5aed4cefaf7d753dabbaee68fb6e
* test: isolate config dir in format-universal test
Change-Id: I3a59942aa8a6753cd949ca42f2a19a72f032ff55
* test: revert unnecessary config-dir isolation (mount-only test)
Change-Id: I0146e5a2f57f5419863bdeeaa1a662fd8f70bddf
internal/util imported internal/proxyplugin (SharedTransport, FallbackTransport,
NewHTTPClient, and WarnIfProxied via proxyPluginStatus), so a foundational util
package depended up into a feature package, pulling binding/core/vfs into the
transitive cone of every util importer.
Move internal/proxyplugin -> internal/transport and make it the single owner of
outbound transport: fold the two SharedTransport functions into one Shared()
(proxy-plugin override -> LARK_CLI_NO_PROXY -> http.DefaultTransport), and move
Fallback/NewHTTPClient/WarnIfProxied/DetectProxyEnv/noProxyTransport out of the
now-deleted internal/util/proxy.go into the new package. The proxy-plugin probe
is demoted to a private pluginTransport(); the duplicate redactProxyURL collapses
to one. internal/util keeps no proxy code and is a leaf again.
Re-point all consumers (registry, doctor, config, auth, cmdutil, update) to
internal/transport. Behavior-preserving: package move + symbol rename + dedup.
Two new tests lock the fail-closed contract (plugin overrides NO_PROXY; malformed
config never falls through to direct egress).
Every failure on the authentication, authorization, and configuration
path now surfaces as a typed structured error instead of an ad-hoc
envelope. Users and scripts that consume CLI output get:
- a fixed nine-category taxonomy on the wire, each mapped to a
stable shell exit code (authentication/authorization/config = 3,
network = 4, internal = 5, policy = 6, confirmation = 10)
- identity-aware detail fields (missing_scopes, requested_scopes,
granted_scopes, console_url, log_id, retryable, hint) carried
uniformly on the envelope
- a single canonical policy envelope at exit 6; the legacy
auth_error carve-out is retired
- per-subtype canonical message + hint that preserves Lark's
diagnostic phrasing and routes recovery to the right actor:
app developer (app_scope_not_applied), user (missing_scope,
token_scope_insufficient, user_unauthorized), or tenant admin
(app_unavailable, app_disabled)
- wrong app credentials classify as config/invalid_client whether
surfaced by the Open API endpoint (99991543) or the tenant
access-token mint endpoint (10003 / 10014), instead of
collapsing to a transport error or api/unknown
- local shortcut scope preflight emits the same
authorization/missing_scope envelope (identity + deterministic
missing-scope set) used by the post-call permission path, so AI
consumers read the same structured shape from precheck and from
server-returned permission denial
- streaming download/upload failures keep the same network subtype
split (timeout / TLS / DNS / transport) as the non-stream path
instead of collapsing every cause to a generic transport failure
- console_url is carried only on the bot-perspective
app_scope_not_applied envelope (where the recovery action is
"developer applies the scope at the developer console"); the
user-perspective missing_scope envelope drops the field, since
the only actionable user recovery is `lark-cli auth login --scope`
and pointing an end user at a console they cannot modify is
misleading
- bind workflows (Hermes / OpenClaw / lark-channel) flatten dynamic
Type tags to wire 'config' with the original module name kept
as a metric label
All 10 typed errors are cause-bearing, nil-safe on .Error() and
.Unwrap(), and defensively clone slice setter inputs. Four lint
rules (CheckNilSafeError / CheckBuilderImmutable / CheckUnwrapSymmetry
/ CheckBuildAPIErrorArms) lock these invariants on migrated paths.
* feat(platform): support multiple policy rules per plugin
Extend the command policy framework from single-Rule to multi-Rule
semantics. A plugin (or policy.yml) may now contribute several scoped
Rules; the engine combines them with OR -- a command is allowed when it
satisfies every axis of at least one rule. This lets one integration
apply different risk ceilings and identity restrictions to different
command groups.
The cross-plugin fail-closed boundary is preserved: two distinct plugins
both calling Restrict still aborts startup (multiple_restrict_plugins).
Single-Rule behaviour is fully backward compatible -- the rejection
reason_code / rule_name / envelope shape are byte-for-byte unchanged;
multi-rule rejection surfaces the aggregate reason_code no_matching_rule.
- engine: New keeps single-rule compat, add NewSet for OR over rules
- resolver: dedupe by owner (one plugin may contribute many rules),
return []*Rule; yaml gains a top-level rules: list
- registrar/builder/staging: Restrict may be called more than once;
retire the double_restrict error
- config policy show / config plugins show: emit a rules array
- inventory: PluginEntry.Rules is now a slice (fixes last-rule-wins
overwrite when a plugin contributes multiple rules)
* fix(platform): clone rules in Builder.Restrict and inventory snapshot
Address review feedback. Builder.Restrict stored the caller's *Rule
directly, so reusing and mutating one Rule object across multiple
Restrict calls collapsed entries to the last mutation; clone the rule and
its slices on append, mirroring the staging registrar.
BuildInventory likewise reused the source Allow/Deny/Identities slices;
copy them when building the RuleView snapshot instead of relying on
cloneInventory downstream.
Add a regression test: reusing and mutating one Rule across two Restrict
calls now yields two independent rules.
* fix(platform): skip yaml when a plugin owns policy; reject empty rules list
Two policy-config robustness fixes from review:
- A malformed ~/.lark-cli/policy.yml could abort a plugin-governed
binary. applyUserPolicyPruning read yaml before resolving, and
build.go fail-closes on any policy error when a plugin is present.
Plugin rules shadow yaml anyway, so skip reading yaml entirely when a
plugin contributed rules -- an unrelated broken file on the user's
machine can no longer lock the CLI.
- A present-but-empty "rules: []" collapsed to a single all-zero Rule
that allows every annotated command ("looks like policy, enforces
almost nothing"). yaml.Parse now distinguishes absent from
present-but-empty (Rules is a pointer) and rejects the empty list.
Add regression tests for both.
Fix 3 occurrences of --minute-token (singular) to --minute-tokens
(plural) in lark-vc-recording.md to match the actual CLI flag
definition in minutes_download.go.
The size==1 (64-bit "largesize") branch of all three MP4 box walkers
(findMP4Box, readMp4DurationBytes, readMp4Duration) set boxEnd to the raw
largesize instead of offset+largesize — even though the 32-bit branch right
below correctly uses offset+size. Two consequences:
- Correctness: for any MP4 that carries a 64-bit box size at a non-zero
offset, the box walk is computed from the wrong end, so the moov/mvhd
lookup is truncated and the media duration is silently lost.
- Robustness/security (CWE-190): the unguarded uint64->int(64) conversion of
a largesize with the high bit set yields a negative boxEnd. The in-memory
walkers then assign it to offset and feed it back as a slice index
(data[offset:]), panicking with "slice bounds out of range" and crashing
the CLI on a crafted or corrupt MP4. This is reachable via URL-sourced IM
media, whose bytes the caller does not control.
Fix: compute boxEnd as offset+largesize (matching the 32-bit branch) and
reject largesize values smaller than the 16-byte header or larger than the
remaining input. Malformed media now honours the parsers' best-effort
contract by returning 0/-1 instead of panicking, and the bounds guarantee
the conversion can no longer overflow.
Add regression tests covering both the overflow (must not panic) and a
64-bit box at a non-zero offset (must walk correctly).
Add a new --types flag (string_slice; values from {group, p2p}) to
+chat-list, backed by the new GET /open-apis/im/v1/chats `types` query
parameter. Accepts CSV (--types group,p2p) and repeated-flag forms
(--types group --types p2p).
Defaults to groups-only (backward compatible). Under user identity,
p2p single chats appear with chat_mode="p2p" plus p2p_target_type /
p2p_target_id fields. Under bot identity:
- --types=p2p alone is rejected at validation
- --types=p2p,group is silently downgraded to types=group (no runtime
notice; skill docs document this contract)
Updates Shortcut.Description, lark-im SKILL.md (frontmatter trigger
+ shortcut table row), and the chat-list reference doc with command
examples, the new parameter, output field documentation, and a
dedicated "Bot identity and p2p" section.
Change-Id: I637ce23b3c6ce4ec350f0ac26dbac8120761bb71
* fix(install): detect curl version before using --ssl-revoke-best-effort
(cherry picked from commit da14737702)
* test(install): cover curl version gate and refactor for testability
Extract the version comparison out of curlSupportsSslRevokeBestEffort()
into a pure isCurlVersionSupported(output), so the >= 7.70.0 logic is unit
testable without spawning curl. Add cases for 7.55.1 / 7.69.0 / 7.70.0 /
8.x plus the unparseable and libcurl-token edge cases (the regex must read
the leading "curl X.Y.Z", not the trailing "libcurl/X.Y.Z").
Memoize the `curl --version` probe: curl's version is invariant for the
install's lifetime while download() runs once per mirror URL, so probe at
most once instead of re-spawning curl on every attempt.
---------
Co-authored-by: EllienTang <146210093+Ellien-Tang@users.noreply.github.com>
Co-authored-by: liangshuo-1 <266696938+liangshuo-1@users.noreply.github.com>
Two issues caught in review of #1132 that the existing tests missed because
they constructed RuntimeContext/CliConfig directly, bypassing the credential
edge where the bug lives.
P1 — Lang dropped at credential boundary
credential.Account had no Lang field, so AccountFromCliConfig and
ToCliConfig silently dropped cfg.Lang. The production Factory builds
CliConfig via acct.ToCliConfig() (factory_default.go Phase 3), which
meant RuntimeContext.Lang() always returned "" in production and
shortcuts/mail/mail_signature.go always fell back to zh_cn — defeating
the whole point of persisting --lang.
Fix: add Lang i18n.Lang to Account and copy it in both directions.
Regression test: TestFullChain_LangSurvivesProductionPath walks the
real path (SaveMultiAppConfig -> DefaultAccountProvider.ResolveAccount
-> ToCliConfig) and asserts Lang survives, so any future field added
to CliConfig forces the same audit.
P2 — priorLang ignored CurrentApp in multi-profile workspaces
priorLang scanned all Apps and returned the first non-empty Lang. If a
user had multiple profiles and the active one disagreed with Apps[0],
a re-bind without --lang would silently inherit the wrong profile's
preference.
Fix: read multi.CurrentAppConfig("").Lang instead.
Regression tests cover CurrentApp wins over Apps[0], single-app
fallback, and malformed bytes.
Change-Id: If7a276605f84f398cec329c2c942b471b4c32749
Follow-up to #1095. The reactions auto-enrichment shipped, but on busy chats the strictly-serial per-resource fetches in EnrichReactions, ExpandThreadReplies, and merge_forward expansion stretched the command's wall time above 14s — enough that wrapper agents (30–60s wall-clock budgets) saw timeouts even though the CLI itself never errored. This PR parallelizes all three with the same bounded-concurrency pattern, batches the follow-up contact-API sender resolution so it doesn't fan back out into a serial stall, and fixes two correctness bugs that surfaced during review. Scoped to convert_lib/{reactions,thread,merge,content_convert}.go + tests + the 4 shortcut Execute hooks + the reference doc.
Change-Id: I0206d10ad204382170bd42aec67f82578923736e
The six TestDriveInspectExecute_* tests set
t.Setenv("LARKSUITE_CLI_CONFIG_DIR", t.TempDir()) but build the CLI via
cmdutil.TestFactory(t, cfg), which provides an in-memory config closure
(func() (*core.CliConfig, error) { return config, nil }) and never reads the
filesystem. Per the repo learning from PR #343, this env var should only be
set for tests exercising the real NewDefault() factory path. None of these
tests use NewDefault(), so the calls are dead and removed.
No behavior change; all TestDriveInspect* tests still pass.
Co-authored-by: kyalpha313 <kyalpha313@users.noreply.github.com>
妙搭/spark consolidated the apps domain onto spark:app:read / spark:app:write.
The standalone spark:app:publish and spark:app.access_scope:* scopes are retired.
- +html-publish: spark:app:publish -> spark:app:write
- +access-scope-get: spark:app.access_scope:read -> spark:app:read
- +access-scope-set: spark:app.access_scope:write -> spark:app:write
Verified against the official docs for upload_html_code_and_release,
get_app_visibility and update_app_visibility. +create/+update/+list were
already correct (spark:app:write / spark:app:read).
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add `lark-cli mail +draft-send` shortcut that takes one or more existing
draft IDs and sends each via POST /drafts/:draft_id/send sequentially.
Per-draft failures are isolated and aggregated into a structured output;
fatal failures (auth, permission, network, mailbox quota) abort the
entire batch immediately while recoverable failures honor --stop-on-error.
Also extend internal/output with six mail-send-specific errno constants
(LarkErrMailboxNotFound=4013, LarkErrMailSendQuota{User,UserExt,TenantExt},
LarkErrMailQuota, LarkErrTenantStorageLimit) consumed by isFatalSendErr.
Risk is "high-risk-write" so the framework's --yes gate applies; the
shortcut declares only the minimal mail:user_mailbox.message:send scope
to avoid asking users for permissions it does not need.
- Pull messages now auto-call im.reactions.batch_query and attach a
reactions block (counts + details) to each message. Stops AI from
misjudging "user already reacted" as "no response yet" and
re-sending duplicate reactions. Server caps queries[] at 20 per
call, so messages are split into batches of size <= 20.
- Edited messages additionally surface update_time. The server echoes
update_time == create_time for unedited messages too, so the field
is only emitted when updated == true; otherwise every message
output would look "edited". The value is read via an explicit
string assertion + TrimSpace so empty strings are filtered properly
(the previous `v != ""` was a no-op for non-string types).
- All four message-pulling shortcuts (+messages-mget,
+chat-messages-list, +messages-search, +threads-messages-list) get
a --no-reactions opt-out flag for callers that want to skip the
extra round-trip.
- Each shortcut declares im:message.reactions:read on its
UserScopes/BotScopes (or Scopes for the user-only search command) so
the auth flow covers the new dependency.
- Each shortcut's --dry-run output now lists the
reactions/batch_query call (or omits it when --no-reactions is set),
so callers can audit the full set of API calls before execution.
- Warnings go through runtime.IO().ErrOut (forbidigo lint requires
IOStreams over os.Stderr in shortcut code).
- Duplicate message_id inputs (e.g. mget --message-ids om_a,om_a)
attach the reactions block to every entry while still querying the
API only once per distinct id.
- EnrichReactions walks msg["thread_replies"] recursively, and mget/
chat-messages-list call it after ExpandThreadReplies, so replies
receive reactions in the same batched call as their parent message.
- When the batch_query call fails or returns per-message failures,
the affected messages get reactions_error=true (mirroring the
thread_replies_error flag from thread.go) so consumers can
distinguish "fetch failed" from "no reactions exist" by reading
stdout alone, without depending on the stderr warning channel.
- lark-im skill docs: the default-enrichment contract lives in a
standalone references/lark-im-message-enrichment.md so the generated
SKILL.md can't strand it on regeneration. The four read references
and the raw reactions API reference link to it, and the template
source skill-template/domains/im.md carries a durable pointer.
Change-Id: Ia9ea74b11945644262bb25c6503fb9b2003c6c98
Affordance examples previously carried a title plus a structured input
object mirroring the inputSchema. Replace that with a description plus a
command string holding a ready-to-run lark-cli invocation, which is what
an AI agent driving the CLI actually consumes.
No affordance data exists in the registry yet, so this only reshapes the
consuming AffordanceCase type and its tests; the data pipeline
(registry-config.yaml -> gen-registry.py -> meta_data.json) forwards the
new keys verbatim.
* feat(schema): add envelope types and ordered properties container
* feat(schema): build meta_data.json key-order index for property ordering
* feat(schema): implement convertProperty with file/enum/range/nested handling
* feat(schema): build inputSchema with x-in / file binary / yes injection
* feat(schema): build outputSchema wrapping responseBody
* feat(schema): build _meta with scopes/risk/access_tokens normalization
* feat(schema): scaffold affordance overlay loader (PR-1 stub)
* feat(schema): wire up AssembleEnvelope main entry point
* feat(schema): parse dotted and space-separated path arguments
* feat(schema): batch envelope assembly with optional method filter
* feat(schema): implement L1-L3 envelope lint (structure/type/cross-field)
* feat(schema): measure L4 coverage and gate all envelopes through L1-L3
* feat(schema): add golden test harness with UPDATE_GOLDEN refresh
* test(schema): seed 20 golden envelopes covering edge cases
* feat(schema): output MCP envelope as default JSON, preserve pretty mode
Rewrites cmd/schema/schema.go so the default --format json branch emits
MCP-spec envelopes via schema.AssembleAll/AssembleService/AssembleEnvelope.
The legacy --format pretty branch is preserved verbatim and still uses
printServices / printResourceList / printMethodDetail.
Args max raised from 1 to 8 so the path can be supplied either as a single
dotted argument (im.reactions.list) or as space-separated segments
(im reactions list); both forms route through schema.ParsePath and produce
byte-identical output.
The completeSchemaPath function is extended to drive tab-completion for
both forms: legacy dotted prefix when len(args) == 0, and per-segment
resource/method completion when args already contains earlier segments.
BREAKING CHANGE: default JSON output shape changes from the raw meta_data
structure to an MCP envelope array/object. Existing scripts parsing the
old shape must either pin --format pretty or migrate to the new envelope
fields (name, description, inputSchema, outputSchema, _meta).
* test(schema): cover envelope JSON output, space-form path, yes injection
Replaces TestSchemaCmd_NoArgs with two variants reflecting the new default
shape: TestSchemaCmd_NoArgs_Pretty asserts the legacy "Available services"
text appears only under --format pretty, and TestSchemaCmd_NoArgs_JSON_IsArray
asserts the default JSON output parses as an envelope array with at least 180
entries.
Adds six new tests:
- TestSchemaCmd_JSONIsEnvelope: single-method output has name / description
/ inputSchema / outputSchema / _meta keys and envelope_version "1.0".
- TestSchemaCmd_SpaceSeparatedPath_EqualsDotted: dotted and space forms
produce identical output bytes for the same command path.
- TestSchemaCmd_ServiceListIsArray: schema <service> returns a JSON array
whose every entry's name starts with "<service> ".
- TestSchemaCmd_HighRiskYesInjection: high-risk-write commands inject
inputSchema.properties.yes.
- TestSchemaCmd_NoYesForReadRisk: read-risk commands do not inject yes.
- TestSchemaCmd_PrettyUnchanged_KeyTextPresent: --format pretty still
surfaces the legacy section markers (Parameters:, Response:, Identity:,
Scopes:, CLI:).
* feat(schema): assemble envelope from embedded data only for stability
* chore(schema): lint cleanup
* fix(schema): preserve dotted resource segments in envelope name
Nested resources whose meta_data key contains a dot (e.g. chat.members,
user_mailbox.templates) were previously split on '.' and rejoined with
spaces, producing envelope names like 'im chat members bots'. AI
consumers doing name.split(' ') and feeding the result back as argv
got 'lark-cli im chat members bots' which the CLI rejects — the actual
invocation form is 'lark-cli im chat.members bots'.
Pass the dotted resource key as a single argv segment so the envelope
name 'im chat.members bots' round-trips through name.split(' ') back
to the CLI. Mirror the same convention in the golden harness so its
single-method assembly matches the live AssembleService walk.
* fix(schema): align MCP envelope output with JSON Schema 2020-12 contract
- coerce enum literals to typed JSON values (integer to int64,
number to float64, boolean to bool) so type:"integer" fields no
longer emit string enums; sort numeric/boolean enums while
preserving meta_data order for string enums that carry semantic
priority
- translate non-standard meta_data type:"list" to JSON Schema
type:"array" with items:{} fallback when element shape is absent
(covers the two mail attachment_ids fields)
- render inputSchema.required even when empty so consumers see a
stable envelope shape ("[]" means no required fields, not "field
is missing")
- reject trailing path segments in both JSON and pretty modes so
schema im.messages.delete.foo errors instead of silently
returning the delete method
- drop dead "list type" entry from lint_test isKnownDataInconsistency
whitelist now that list values are translated upstream
* fix(schema): address CodeRabbit findings and stabilize CI tests
CI fix
- Replace hard-coded absolute key-order assertions in TestKeyOrderIndex_*
and TestBuildInputSchema_* with set-membership and propagation invariants;
the upstream meta_data API does not guarantee stable JSON key order across
fetches, so the old tests were flaky on CI by design.
- Skip byte-level TestGoldenEnvelopes when CI=true; golden snapshots are a
manual refresh artefact tied to a specific meta_data fetch, not a CI gate.
- Add TestMain to isolate registry-backed tests from any host ~/.lark-cli
cache (LARKSUITE_CLI_CONFIG_DIR + LARKSUITE_CLI_REMOTE_META=off) so the
suite gives the same answer on every machine.
CodeRabbit review actionables
- EmbeddedServiceNames returns a defensive copy so callers cannot mutate
the package-level slice and affect subsequent assembly determinism.
- coerceEnumValue is now also applied to default literals: integer fields
no longer ship default: "500" — they ship default: 500 (same idea as the
earlier enum coercion fix).
- options-branch string enums preserve meta_data source order, matching the
enum-branch policy; only numeric/boolean enums get sorted.
- validatePropertyTypes now validates the array element schema itself
(type, nested items), not only items.properties — previously a primitive
element with an invalid type (e.g. items.type="list") slipped past lint.
- OrderedProps.MarshalJSON falls back to alphabetical key order when Map
has entries but Order is empty, instead of silently emitting {}.
Tests pass locally and with CI=true env (simulating GitHub Actions).
* chore(schema): refresh golden envelopes after meta_data drift
Re-generated with UPDATE_GOLDEN=1 against the current meta_data.json
snapshot. The bulk of the diff is upstream noise (description wording,
enum entries, field order) which the CI snapshot diff can no longer
reasonably gate (see previous commit). Side-effects of the code fixes
in the parent commit are also captured:
- integer-typed defaults now emit numeric literals (e.g. page_size
default 500, not "500") thanks to coerceEnumValue
- mail.user_mailbox.templates.create _meta.risk corrects to "write"
(assembler already emitted "write"; the old golden was stale)
* fix(schema): address CodeRabbit round-3 review findings
- TestMain: cleanup now runs reliably. os.Exit skips deferred functions,
so the previous defer os.RemoveAll(dir) never executed. Replace defer
with explicit cleanup, and fail fast if MkdirTemp errors instead of
silently running against the host cache (which defeats isolation).
- convertProperty default coercion: when the literal cannot be coerced to
the declared type (e.g. default:"" on integer field, used by meta_data
to mean "no default"), omit the field entirely rather than emit a
type-mismatched default. Removes a contract violation flagged on
im.reactions.list.json#page_size.
* feat(schema): wire affordance overlay into envelope _meta
Replace the loadAffordance stub (which always returned nil and read
from an empty embedded annotations/ directory) with parseAffordance,
which lifts the affordance block from method["affordance"]. The block
is authored under larksuite-cli-registry's registry-config.yaml in the
overrides: section and flows through gen-registry.py's deep_merge into
the embedded meta_data.json.
Simplify buildMeta signature: the service/resourcePath/method args
existed only to feed the old dotted-path lookup.
Refresh 9 golden envelopes for unrelated upstream meta_data.json drift.
* refactor(schema): drop x-in extension from inputSchema
x-in (path/query/body) was an HTTP-shape leak in a CLI-facing tool spec.
AI consumers call the CLI by name with named args — they never construct
HTTP requests directly, so the path-vs-body-vs-query distinction is the
CLI's internal concern, not part of the contract.
Execution path (cmd/service/service.go) already reads location from
meta_data.json directly, so removing x-in does not affect routing.
Drop:
- Property.XIn field
- validXIn map and the two lint rules that depend on x-in
(L1 "top-level missing x-in" and L2 "path field must be in required")
- contains() helper, no longer referenced after the path-required rule
went away
Refresh 20 goldens for the now-absent x-in lines.
* refactor(schema): wrap inputSchema into params/data/flags sub-objects
Replace the flat inputSchema with a 3-bucket nested structure that mirrors
the CLI's actual flag layout, so AI consumers can directly map envelope
fields to lark-cli invocation:
inputSchema:
properties:
params: { ...path + query fields } → CLI --params JSON
data: { ...body fields } → CLI --data JSON
flags: { yes: ... } → CLI --yes (only for high-risk-write)
Each sub-object only appears when the method has the corresponding source,
so read-only GETs have a single `params` block, body-only POSTs have a
single `data` block, etc.
The `flags` wrapper carries an explicit description marking it as a CLI
control bucket (not API fields), so AI does not confuse `yes` with a
backend parameter.
Lint:
- L2 walkForL2 helper recurses into params/data sub-objects so leaf
invariants (format:binary on non-string, min<max, required-in-properties)
still apply.
- L3 yes-presence check now navigates flags.properties.yes.
Refresh all 20 goldens for the new shape.
* refactor(schema): drop flags wrapper, put yes at top level alongside params/data
The flags wrapper added one extra layer for a single field. Flatten so
inputSchema.properties has three siblings:
inputSchema:
properties:
params: { ...path + query } → CLI --params
data: { ...body } → CLI --data
yes: { boolean, default:false } → CLI --yes (only when risk == high-risk-write)
`yes` description strengthened to mark it as a CLI confirmation gate
(consumed by lark-cli, not sent to the backend), so AI can still
distinguish it from API fields without needing a wrapper.
Lint L3 yes-presence check goes back to top-level Properties.Map["yes"].
Refresh 20 goldens.
* feat(schema): add `file` top-level sub-object for binary upload fields
Splits file fields out of `data` into their own sibling, so the four
top-level slots in inputSchema map 1:1 to CLI flag dispatch:
inputSchema.properties:
params { path + query fields } → --params JSON
data { non-file body fields } → --data JSON
file { type:file body fields, format:binary } → --file <key>=<path>
yes boolean → --yes (only when risk == high-risk-write)
Each slot is conditional: only registered when the method actually has
fields for that source. This matches the CLI's own conditional flag
registration (cmd/service/service.go:170-195), so what AI sees in the
schema is exactly what flags exist for that method.
The file sub-object carries a description explaining its semantics so AI
knows to use --file for those fields rather than embedding the binary
in --data JSON.
Refresh im.images.create golden (the only file-upload method in the
golden set).
* test(schema): cover L2 lint recursion into params/data sub-objects
Add two negative test cases that stuff bad values inside the wrapped
inputSchema sub-objects (rather than at top-level), to lock in
walkForL2's recursive coverage:
- format:binary on a non-string field nested under params
- sub-object Required referencing a key not in its Properties
Regression guard so future walkForL2 refactors do not silently lose
recursion and let leaf-field violations slip past lint.
* fix(schema): coerce example, aggregate nested required, fix path hint
- coerce `example` literal to the declared JSON Schema type (rename
coerceEnumValue -> coerceLiteral, drop on coerce failure to match the
`default` policy). Without this, integer/boolean/number fields emitted
string examples and failed strict validators.
- aggregate child field `required:true` into the enclosing nested
object's `required[]` (both object and array-items shapes). Previously
only the top-level params/data sub-objects scanned `required`, so
envelopes silently under-reported the real call contract.
- check method existence before reporting trailing-segment failure in
both JSON and pretty `schema` paths. A typo like `schema im messages
typo extra` now reports "Unknown method: im.messages.typo" instead of
the misleading "Method 'typo' exists but trailing segments ..." hint.
- extract risk level constants (RiskRead / RiskWrite / RiskHighRiskWrite)
in internal/cmdutil/risk.go; replace literal usages in schema, lint,
and confirm helpers so the typo radius is one file.
- reconcile AssembleEnvelope docstring with implementation reality (the
package-level currentMethodOrder + assembleMu serialize concurrent
callers; output is deterministic per inputs).
- drop testdata/golden/ and golden_test harness. End-to-end envelope
shape regression now relies on real CLI invocations and the existing
property-level unit + lint coverage.
* fix(schema): emit items:{} for all typeless arrays, restore lint gate
The list→array fallback only added items:{} when the source type was
"list", leaving ~64 natively-typed array fields (e.g.
approval.instances.cc.cc_user_ids) as {type:"array"} with no items.
These violated the L1 lint rule, but TestAllEnvelopesPass skipped the
"array missing items" error as a known data inconsistency, so the MCP
tool contract was not actually lint-clean.
Relax the fallback to cover every array lacking element shape regardless
of source type, and drop the lint-test skip so the gate is hard again.
Parse keywords from minutes artifacts API in vc +notes and document
the field in lark-vc skill references.
Co-authored-by: Cursor <cursoragent@cursor.com>
The standup workflow and the +get-my-tasks reference both implied a
"pending todo summary" use case but did not pass --complete=false in
the example commands. As a result, completed tasks were surfaced into
standup/daily summaries as if they were still pending.
This change updates the workflow and reference docs only — the
underlying command behavior is unchanged.
Closes#993
Introduce a typed error contract framework for lark-cli so in-process
Go callers can branch via errors.As(&errs.XxxError{}) and shell scripts,
AI agents, and protocol adapters can branch on stable JSON type/subtype
fields instead of regex-parsing free-form messages.
Adds:
- Canonical taxonomy under errs/ (9 categories + typed Error structs
embedding a shared Problem, RFC 7807-aligned)
- Centralized Lark code metadata + identity-aware BuildAPIError dispatch
- Typed JSON envelope writer alongside the legacy envelope writer
- MCP / OAuth (RFC 6750 Bearer) projection adapters
- Five CI lint guards preventing ad-hoc taxonomy drift
Backward compatibility: legacy *output.ExitError producers (ErrAPI,
ErrWithHint, Errorf, ErrBare) and business shortcuts that use them
continue to render the legacy envelope unchanged. SecurityPolicyError
wire format and exit code are preserved via a carve-out; taxonomy
migration is deferred to PR 2. Domain-specific business migration is
staged across PR 3+.
Framework-direct paths now return typed *errs.*Error: ErrAuth /
ErrValidation / ErrNetwork emit category literals on the wire
(authentication / validation / network), *core.ConfigError is promoted
at the cmd/root boundary with exit code aligned from 2 to 3, and Lark
API permission denials classified by BuildAPIError exit 3.
At the SDK boundary, WrapDoAPIError preserves any already-classified
error (legacy *output.ExitError or typed *errs.*) so output.ErrAuth
from missing credentials surfaces with the auth category and exit 3
intact instead of being downgraded to a network error. Policy responses
classified by BuildAPIError (codes 21000 / 21001) extract challenge_url
and the canonical hint from the response body, matching what the
auth transport already surfaces at the HTTP layer; non-https
challenge URLs are dropped.
First PR in the feat/error-contract-* series.
* feat(apps): replace +html-publish cwd hard-reject with credential-file scan
The previous --path == "." block was a coarse heuristic: it caught the
common foot-gun of publishing a repo root, but also rejected legitimate
clean cwds, and let a ./dist with a forgotten .env ship the secret
through anyway (the sensitive-paths scanner was advisory and never ran
on the Execute path).
Move the gate from path shape to path content:
- Validate now walks --path candidates and rejects publishes that
include well-known credential files (.env / .env.* / .npmrc / .netrc
/ .git-credentials / .aws/credentials / .gcloud/credentials* /
.docker/config.json / .kube/config). Living in Validate (not DryRun)
means dry-run returns non-zero on hit too, so the dry-run preview
matches Execute.
- Narrow the credential pattern set. .git/, SSH private keys, *.pem
and *.key are out of scope -- they're not env-token files and the
false-positive rate (public certs, docs about key formats) is high.
- Add --allow-sensitive as the escape hatch for legitimate cases
(e.g. a docs site shipping .env.example on purpose). DryRun surfaces
the waived list in sensitive_waived so the caller can relay it.
- Drop the cwd defense-in-depth in runHTMLPublish. A clean cwd is now
a valid publish target.
The lark-apps skill and the html-publish reference are updated to
describe the new gate, the override flag, and the patterns now
explicitly out of scope.
* feat(apps): drop .gcloud/* from credential-file scan
The .gcloud/credentials pattern matched a non-existent path: gcloud's
actual config dir is ~/.config/gcloud/ (XDG-based), and the real
credential files there are credentials.db / access_tokens.db /
application_default_credentials.json -- none of which would land under
a .gcloud/ segment in a publish payload.
Drop the rule rather than fix it: the realistic gcloud foot-gun would
require recognizing the .config/gcloud/* tree by file basename, which
is a broader change than the targeted env/cred scan in this PR. The
remaining 7 patterns (.env / .env.* / .npmrc / .netrc /
.git-credentials / .aws/credentials / .docker/config.json /
.kube/config) cover the common Node/Python/CLI-tooling foot-guns.
* fix(apps): close credential-scan bypass when --path is the parent dir itself
isSensitiveRelPath anchors cloud-SDK matchers on adjacent parent/file
segments (.aws/credentials, .docker/config.json, .kube/config), but
walker strips that parent via filepath.Rel when --path is the conventional
parent dir (e.g. ./.aws), yielding a bare RelPath="credentials" that
slipped through silently. Same bypass for the single-file form
--path ./.aws/credentials (walker sets RelPath = Base(rootPath)).
Wrap the scan in isSensitiveCandidate: keep the fast RelPath scan, and
on miss fall back to filepath.Abs(AbsPath) so the parent segment is
visible again. isSensitiveRelPath itself is unchanged; existing tests
still pin its pure-function contract.
* fix(apps): drop filepath.Abs from sensitive scan to satisfy forbidigo lint
The previous fix called filepath.Abs(c.AbsPath) — banned by the repo's
forbidigo rule because shortcuts must not reach into the filesystem for
path resolution.
Reframe the same fix without fs access: re-prepend the root's basename
(or, for the single-file form, the parent dir's basename of rootPath)
to RelPath and re-scan only the parent-anchored credential pairs
(.aws/credentials, .docker/config.json, .kube/config). Leaf matchers
(.env / .npmrc / ...) stay scoped to RelPath — incidentally closing a
latent false-positive where --path /home/alice/.env/dist would have
flagged every file under it just because .env appeared in the
absolute path.
* fix(apps): read app object from data.app for +create and +update
The Miaoda OpenAPI returns the application object nested under
data.app for both POST /apps and PATCH /apps/{appId}. The CLI text
helper was reading common.GetString(data, "app_id"), which yields an
empty string against the wire format -- so `lark-cli apps +create
--format pretty` printed `created: ` with no ID.
Navigate the new nested path via GetString(data, "app", "app_id") for
both create and update. Update unit-test mocks to wrap the response
under `app`. Refresh the lark-apps skill references (example response
shape + jq paths) so agents reading them follow the right path.
Wire format is passed through to the user's JSON envelope untouched
-- no unwrapping in CLI. Consumers reading the response should use
.data.app.app_id.
The GET /apps list endpoint is unchanged: per the design doc its
items[] are flat objects, no wrapper.
* docs(apps): add required --app-type HTML to scenario 2 snippet
The "用户没有 app_id" snippet in lark-apps-html-publish.md was missing
the required --app-type flag, so copy-pasting it triggered Validate
("--app-type is required") and left $APP empty -- the following
+html-publish then failed with --app-id "". Bring the snippet in line
with every other apps +create example in the skill.
* docs(apps): simplify auth-recovery rule to error.type == missing_scope
Every apps shortcut declares Scopes, so the precheck path in
shortcuts/common/runner.go:825 is always the one that fires on scope
violations and the envelope's error.type is the stable discriminator.
Drop the keyword-sniffing of error.hint, the chain explanation, and the
bot caveat — they all reduce to one boolean: error.type == "missing_scope"
→ run `lark-cli auth login --domain apps`.
Also collapse the corresponding bullet in 快速决策 to point at this rule.
* fix(common): escape special chars in multipart form filenames
MultipartWriter.CreateFormFile concatenated the fieldname and filename
into the Content-Disposition header without escaping, so a filename
containing a double-quote, backslash, CR, or LF produced a malformed
header. For example, uploading `report "draft" v2.pdf` via
`task +upload-attachment` made the server see `filename="report "`
(truncated at the first internal quote) and drop the rest.
Drop the custom override and let CreateFormFile be promoted from the
embedded *multipart.Writer, which applies the stdlib's quoteEscaper
(backslash and double-quote get a backslash prefix; CR and LF get
percent-encoded). The Content-Type ("application/octet-stream") and
the wrapper API are unchanged, so the existing `task +upload-attachment`
call site is unaffected -- filenames with special characters just now
round-trip correctly.
Add helpers_test.go covering plain, quoted, backslashed, mixed, and
unicode filenames. The test asserts both the on-wire encoding and a
round-trip through mime.ParseMediaType (bypassing Part.FileName, whose
filepath.Base is platform-dependent for backslash on Windows).
* test(common): cover CR/LF/CRLF in multipart filename escaping
Per code-review feedback, extend the helpers_test.go cases table with
CR, LF, and CRLF filenames so the test exercises both legs of the
stdlib's quoteEscaper:
- backslash and double-quote use backslash escaping (quoted-pair);
these round-trip exactly through mime.ParseMediaType.
- CR and LF use percent encoding to prevent header injection; the
MIME parser does not decode percent escapes, so the read-side
filename param contains literal "%0D"/"%0A".
The cases table grows a wantParsed column so each case can declare its
expected post-parse value (same as filename for backslash-escaped chars,
percent-encoded for CR/LF).
* refactor(common): polish doc comments and regroup test cases
Two follow-up tweaks suggested by a re-read of the PR:
- helpers.go: stop naming the stdlib's internal `quoteEscaper` in the
doc comment. Describe the observable behaviour ("escapes special
characters") instead, so the comment stays valid if the stdlib ever
renames or reimplements its escaping.
- helpers_test.go: rename the vague `with both` case to
`backslash and quote`; split the table-driven cases into three
visually-separated groups (happy path / backslash escaping /
percent encoding) so it is obvious why two cases have a different
wantParsed than filename.
No behaviour change; tests still pass 8/8.
* test(common): drop CR/LF filename cases that depend on Go 1.24+ stdlib
CI runs against the toolchain pinned in go.mod (1.23.0), whose
multipart/Writer.quoteEscaper escapes only backslash and double-quote.
Percent-encoding of CR and LF was added to the stdlib later, so the
three CR / LF / CRLF cases I added on review feedback fail on CI: the
literal CR/LF lands in the Content-Disposition header and the parser
reports `malformed MIME header: missing colon`.
Drop those three cases. The fix in the prior commits still covers the
real-world bug — backslash and double-quote in filenames — which is
what the original `report "draft".pdf` example demonstrates. CR or LF
in a filename is essentially never legal on any supported OS, so
leaving that edge case to a future stdlib upgrade keeps the test
stable across toolchains.
Also dropped the now-unused wantParsed column from the cases table:
with only round-trippable characters left, mime.ParseMediaType returns
the original filename byte-for-byte, so a single tc.filename comparison
suffices.
---------
Co-authored-by: Wang-Yeah623 <Wang-Yeah623@users.noreply.github.com>
When creating wiki nodes under the same parent concurrently, the API
returns error code 131009 (lock contention) ~5-15% of the time. This
adds automatic retry with exponential backoff (250ms, 500ms; max 2
retries) so callers no longer need to implement retry logic themselves.
- Retry loop in runWikiNodeCreate: only retries on code 131009, respects
context cancellation, prints progress to stderr
- wrapWikiNodeCreateRetryError preserves Err/Raw/Detail.Code in ExitError
- 6 unit tests covering retry success, exhaustion, non-contention error,
single-retry success, context cancellation, no-retry on success
- 8 dry-run E2E tests for wiki +node-create request shape and validation
Per issue #1049 (third point), wiki +node-get used --token while sibling
commands (+node-delete / +node-copy / +move) use --node-token. The
inconsistency forced humans and AI agents to remember which adjacent
command takes which flag.
Make --node-token the canonical flag and keep --token as a hidden,
deprecated alias so existing scripts continue to work. pflag's
MarkDeprecated prints "Flag --token has been deprecated, use --node-token
instead" to stderr on use, guiding callers to migrate. Conflict between
the two with different values is rejected upfront.
Skills docs (lark-wiki, lark-base) updated to prefer --node-token.
Change-Id: I3415a98f079613c0b1a0b989cf54a09cbb8986fb
Wiki write-path operations (most commonly `wiki +node-create` against the
same parent) surface code 131009 "lock contention" under concurrent calls.
Currently this falls through to the generic "api_error" classification,
giving users no hint that it is transient and safe to retry.
Mirror the existing `LarkErrDriveResourceContention` (1061045) treatment:
add a named constant, classify as "conflict", and emit a hint that points
the caller toward exponential backoff or serializing sibling-node writes.
Refs: #1012
In buildFanoutResponse, when every fanout query fails AND the first failure
has no Lark API code (i.e. transport, parse, panic, or context-cancel),
the returned ExitError was carrying an empty Hint. This is the only
output.ErrWithHint call in shortcuts/ that ships an empty hint.
AGENTS.md states: "every error message you write will be parsed by an AI
to decide its next action. Make errors structured, actionable, and
specific." An empty hint gives the agent nothing to do.
Populate the hint with the actionable next step for this branch — retry,
and if it persists, narrow --queries to a single term to isolate the
failing input. The companion test exercises the no-code path and asserts
the hint is non-empty and mentions "retry".
Co-authored-by: Wang-Yeah623 <Wang-Yeah623@users.noreply.github.com>
Replace 8 bare fmt.Errorf calls with output.ErrValidation across 3 files
so validation errors consistently return structured JSON (type: validation,
exit 2) matching the rest of the codebase.
Affected functions: validateExpectedFlag (sheets), validateSendTime,
validateComposeInlineAndAttachments, validateEventFlags (mail),
validateSignatureWithPlainText (mail)
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
When AutoGrantCurrentUserDrivePermission encounters lark code 99991672/99991679,
extract permission_violations from the underlying ExitError and surface
lark_code, required_scope, and console_url on the result map. Override the
generic fallback hint with one pointing at the developer console — the
concrete next step a user can take.
Refactor extractRequiredScopes / SelectRecommendedScope wrapping / console URL
construction out of cmd/root.go into internal/registry/scope_hint.go so both
the top-level enrichPermissionError path and the best-effort sub-call path in
shortcuts/common share one implementation.
Change-Id: Ida63ed160d1167b7961b6faac5c2cf9b7f971c65
- description: switch from trigger-word enumeration to a general
principle (any HTML artifact intended to be independently accessible
falls under this skill; defer the deploy-vs-demo decision to the
skill body)
- surface apps +access-scope-get in prerequisites list and Shortcuts
table so agents can find the read side of access-scope
- add "writing HTML hard constraints" section: index.html is the
required entry filename, --path cannot equal cwd (both are CLI-side
hard rejects that previously only lived in the html-publish ref)
* feat(sidecar): support multi-client identity isolation in server-demo
When multiple CLI sandbox environments share a single sidecar instance,
user tokens (UAT) were not isolated -- the last user to log in would
overwrite previous users' tokens, causing identity cross-contamination.
This change introduces per-client HMAC key isolation:
- Each client gets a unique client-*.key file for data-plane HMAC signing,
allowing the sidecar to identify request origin.
- A new auth_bridge.go handles management endpoints (login/poll/status)
with explicit client-to-feishuOpenId binding.
- User token resolution is strictly bound to the matched client -- no
fallback to other users' tokens when a client has no mapping.
- The shared proxy.key is reused across restarts instead of regenerated,
fixing a race condition when multiple sidecar instances start together.
Wire protocol (sidecar package) is unchanged; existing single-client
deployments are fully backward compatible.
Signed-off-by: Gao Yang <grany@yeah.net> (topwin.tech)
* fix(sidecar): address review feedback on filesystem and safety
- Replace os.ReadFile/WriteFile/ReadDir with vfs.* equivalents for test
mockability, consistent with project coding guidelines.
- Limit auth bridge request body to 64KB to prevent memory exhaustion.
- Log errors in saveUserMap instead of silently discarding them.
- Reject client keys that collide with the shared proxy key.
- Reject duplicate client keys instead of silently overwriting.
Signed-off-by: Gao Yang <grany@yeah.net> (topwin.tech)
* refactor(sidecar): remove workspace-specific naming and backward compat
- parseClientID: only accept "client_id" field, remove legacy fallback
- loadClientKeys: scan all *.key (excluding proxy.key), no prefix required
- Remove legacy file migration logic in newAuthBridge
- Update flag description to reflect generic key scanning
Signed-off-by: Gao Yang <grany@yeah.net> (topwin.tech)
* refactor(sidecar): extract multi-tenant demo and add unit tests
Address review feedback from sang-neo03:
1. Extract multi-client code into sidecar/server-multi-tenant-demo/,
keeping server-demo as the minimal single-tenant reference.
2. Add unit tests for the isolation guarantee:
- loadClientKeys: shared-key collision and duplicate keyHex are skipped
- verifyWithClientKeys: correct client matched, unknown key rejected
- loadUserMap/saveUserMap: round-trip persistence across restart
3. Cross-link READMEs between server-demo and server-multi-tenant-demo.
Signed-off-by: Gao Yang <grany@yeah.net> (topwin.tech)
* docs(sidecar): rewrite multi-tenant demo README with problem statement and client guide
- Explain the multi-app credential isolation problem (app_secret must
not be exposed to client environments)
- Document typical deployment topology with multiple sidecar instances
- Add complete client setup guide: env vars, multi-app switching, login
flow, and end-to-end workflow example
- Document design decisions and management endpoint details
Signed-off-by: Gao Yang <grany@yeah.net> (topwin.tech)
* fix(sidecar): address CodeRabbit review feedback on tests and docs
- Make TestProxyHandler_AcceptsAllowedAuthHeaders fully offline by using
httptest.NewTLSServer instead of depending on open.feishu.cn
- Isolate TestRun_RejectsSelfProxy config state with t.Setenv and temp dirs
- Check os.MkdirAll error in test fixture setup
- Add language identifiers to fenced code blocks (MD040)
- Validate user-supplied CLI paths with validate.SafeInputPath
Signed-off-by: Gao Yang <grany@yeah.net> (topwin.tech)
---------
Signed-off-by: Gao Yang <grany@yeah.net> (topwin.tech)
- Bump version to 1.0.38
- Update CHANGELOG.md with the apps brand gating change since v1.0.37
- Backfill the [v1.0.38] link reference at the bottom of CHANGELOG.md
Change-Id: I6fd0d1243e2219a1eaa1fae5fae4ff6d8de361da
* feat(apps): gate apps domain off on Lark brand
The Miaoda apps OpenAPI is Feishu-only. On Lark brand:
- shortcut subtree is registered + hidden, RunE returns a structured
brand-restriction error so users see a clear message instead of
cobra's generic "unknown command"
- auth login `--domain apps` is treated as unknown; `--domain all`
skips apps; help text omits it
- scope collection skips apps shortcuts so spark:* scopes are never
requested
The leaf-stub pattern mirrors internal/cmdpolicy/apply.go::installDenyStub
(DisableFlagParsing + ArbitraryArgs + leaf-level PersistentPreRunE
override) so cobra can't short-circuit the stub with a missing-flag or
parent-PreRunE detour.
Change-Id: I5817e87ae6fedabdb5faf05d0d32ea988f7effc9
- +member-add: wrap POST /spaces/{id}/members; --member-type / --member-role
enums, optional --need-notification query (omitted entirely when the flag
is unset, instead of forcing need_notification=false), my_library
resolution under --as user, flattened single-member output
- +member-remove: wrap DELETE /spaces/{id}/members/{member_id}; surfaces the
required member_type + member_role body the API expects, my_library
resolution, fallback to echoing the caller's inputs when the API omits
the member echo
- +member-list: wrap GET /spaces/{id}/members; reuses the +space-list /
+node-list pagination contract (single page by default, --page-all walks
every page capped by --page-limit, --page-token resumes a cursor)
- All three reject bot identity + my_library upfront with a clear hint and
declare the narrowest scope the API accepts (wiki:member:create /
wiki:member:update / wiki:member:retrieve) so tokens carrying only the
narrow scope are not false-rejected by the exact-string preflight
- skill docs: reference pages for the three new shortcuts + SKILL.md
shortcuts table; switch the membership flow guidance from raw
`wiki members create` to the new +member-add path
Change-Id: I158a86aa7f00bb7cecc7a4e99346f3fb151b3c09
When a resource is created with bot identity, the CLI attempts to
auto-grant full_access to the current user. If the user open_id is
missing or the grant API call fails, the result was only written to
the JSON permission_grant field and easily overlooked.
Changes:
- Add stderr warnings when auto-grant is skipped or fails
- Add 'hint' field to permission_grant JSON output with failure reason
and actionable next step (e.g. auth login, check scope, retry)
- Add end-to-end skipped/failed tests across all affected shortcuts
(doc, drive, sheets, slides, wiki, markdown, base)
Closes#963
strings.Fields("") returns an empty slice, causing --scope "" to bypass
validation and return ok: true. Replace the false-positive success path
with an ErrValidation error so callers correctly detect the invalid input.
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
docs +search is in maintenance and will be removed; cloud-space resource
discovery is consolidated onto drive +search. Two related doc/help fixes:
1. Redirect guidance: docs +search -> drive +search
- skill-template/domains/{doc,sheets}.md
- lark-base/SKILL.md: --filter '{"doc_types":["BITABLE"]}' -> --doc-types bitable
- lark-sheets/SKILL.md: body + frontmatter description, add drive-search ref link
Same server API, equivalent capability; only flattens the entry from
nested --filter JSON to flags. reference links repointed to lark-drive.
2. Fix creator_ids/--mine semantic: creator -> owner
The server matches creator_ids (incl. --mine / --creator-ids) by owner
(document owner), not original creator, despite the OpenAPI field name.
- shortcuts/drive/drive_search.go: --help Desc and Tip
- lark-drive/references/lark-drive-search.md: identity section, params, rules, examples
- lark-drive/SKILL.md: top-level guidance
- lark-doc/references/lark-doc-search.md: creator_ids usage note (now self-consistent)
Wire field name creator_ids kept (aligned with the server).
Docs/help strings only, no logic change; gofmt / go vet / package build pass.
Change-Id: If3ebf5a247b7e38b58050c677dc888a310f1c6b6
* feat(doc): warn before overwrite when document contains whiteboard or file blocks
Before executing an overwrite in v1 mode, pre-fetch the current document
and scan the Markdown for <whiteboard> and <file> resource blocks. If any
are found, print a warning to stderr listing the counts and suggesting the
user take a backup with `docs +fetch` first.
Overwrite replaces the entire document and cannot reconstruct these blocks
from Markdown; previously the data was lost with no indication to the caller.
The check is best-effort: a failed pre-fetch silently skips the guard rather
than blocking the overwrite.
* test(doc): add validateSelectionByTitleV1 tests and drop redundant empty-md guard in warnOverwriteResourceBlocks
* fix(doc): use regex for resource block detection, add latency/coverage comments, document skip_task_detail purpose
Switch `drive +export --file-extension markdown` from the legacy V1
GET /open-apis/docs/v1/content API to the V2
POST /open-apis/docs_ai/v1/documents/{token}/fetch API for
higher-quality Lark-flavored Markdown output.
- Update DryRun and Execute paths to use V2 endpoint with JSON body
- Add docx:document:readonly scope for the new API
- Validate V2 response structure (fail fast on missing document/content)
- Encode token in URL path via validate.EncodePathSegment
- Update unit tests and add V2 response validation error path tests
- Add E2E dry-run test for markdown export path
- Update skill documentation
* fix(identitydiag): harden verify path and tighten status semantics
Follow-ups to #957:
- bound bot/user verify calls with a 10s timeout (mirrors the doctor
endpoint probe) so a hanging server cannot wedge `auth status --verify`
or `doctor`
- return StatusNotConfigured (not StatusMissing) when the user-identity
path is blocked by missing app config, matching the bot side
- surface the `{code, msg}` envelope on bot-info HTTP 4xx responses so
callers see why bot auth was rejected, not just the bare HTTP code
- introduce identity{User,Bot,None} constants in cmd/auth/status.go and
use the exported StatusMessage() in the human-readable note instead of
raw status codes like "not_configured"
- collapse the duplicated verify-failed identity construction in the
user path into a local helper
- cover the new failure paths with unit tests (HTTP 4xx with envelope,
business error code, user server-rejected, expired user token,
strict-mode user-only, missing app config for user)
Change-Id: I581348a65f15b1452a6f48a3e3245d09257314ac
* fix(identitydiag): decode bot/v3/info from "bot" field, not "data"
`/open-apis/bot/v3/info` returns `{code, msg, bot: {...}}` — the bot
payload is under `bot`, not `data` as the newer Lark API convention
would suggest. The decoder was reading from a non-existent `data`
field, so `envelope.Data.OpenID` was always empty and every successful
verify was reported as `Bot identity: verify failed: open_id is empty`.
The pre-existing test mocks used `{"data": {...}}` matching the buggy
decoder, so unit tests passed while production reads of every Lark
account failed verification.
Fix:
- change the JSON tag on the envelope from `json:"data"` to `json:"bot"`
- update mocks in identitydiag and cmd/auth/status tests to emit `bot`
Verified locally: `lark-cli doctor` now reports `bot_identity: pass`
for both a normal account and a bot-only profile, restoring the
behavior that #957 set out to deliver.
Change-Id: Ib26dfdd5a0cc37d2d62537ae2bf5e854e67cb83c
* fix(shortcuts/common): decode bot/v3/info from "bot" field, not "data"
Same schema bug as the one fixed in identitydiag — `RuntimeContext.
fetchBotInfo` reads from a non-existent "data" key, so every successful
call would report "open_id is empty" once a caller starts depending on
it.
There are no production callers of `RuntimeContext.BotInfo()` yet
(only tests + the `TestNewRuntimeContextWithBotInfo` helper), so this
bug is dormant — but the pre-existing tests pass with the same wrong
schema in their mocks, so the first real consumer would silently break.
Fix: tag `json:"data"` → `json:"bot"` plus aligning the four mock
fixtures in runner_botinfo_test.go. The Go field name `Data` is kept
to minimize the diff; only the JSON contract is corrected.
Change-Id: I11e1e871603e5349f8df29b1d58e35d07b628dfd
* feat(drive): add +inspect shortcut for document URL inspection with wiki unwrapping
Implements #662: `lark-cli drive +inspect --url <url>` inspects any
Lark/Feishu document URL to get its type, title, and canonical token,
with automatic wiki URL unwrapping via get_node API.
- Add ParseResourceURL (inverse of BuildResourceURL) in common
- Extract FetchDriveMetaTitle as public shared helper
- Add drive +inspect shortcut with wiki unwrapping support
- Add skill reference docs and update SKILL.md
- Dry-run E2E tests for docx URL, wiki URL, and bare token
* refactor: move host validation from ParseResourceURL to +inspect
ParseResourceURL is a general-purpose URL parser that should not
hardcode domain lists — future Lark domains would silently break.
Move isLarkHost/larkHostSuffixes to drive_inspect.go where host
validation is a business decision of the +inspect command.
Add E2E test for non-Lark host with Lark-like path.
* refactor: remove host validation from +inspect
Lark supports custom enterprise domains, so a hardcoded suffix list
can never be exhaustive and would falsely reject valid URLs.
Path-based matching in ParseResourceURL is sufficient; invalid URLs
will fail naturally at the API call stage.
* fix(wiki): surface real node url for +node-create / +node-copy
The create-node and copy-node OpenAPI responses carry a real `url`
field (present in practice though absent from the documented schema).
Both shortcuts ignored it: +node-create synthesized a link via
BuildResourceURL, and +node-copy emitted no URL at all.
Parse `url` into the shared wikiNodeRecord and add a wikiNodeURL helper
that prefers the response url, falling back to BuildResourceURL only
when it is blank. Wire +node-create and +node-copy to the helper so
both surface the canonical link when available.
Change-Id: I0ca5f91b02c24e81d083793e6a8e4f8c966aeec3
* refactor(wiki): move wikiNodeURL to shared wiki_helpers.go
The helper is consumed by both +node-create and +node-copy, so its
placement should reflect the broader usage rather than living in the
create command's file. Pure move; no behavior change.
Change-Id: I9990c12da042f631fe2519911c6a9d663fd5c22b
* feat(mail): bot+mailbox=me validation and dynamic --as help tests
Add validateBotMailboxNotMe helper to shortcuts/mail/helpers.go and
wire it as a Validate callback into +message, +messages, +thread and
+triage, so bot identity combined with the default --mailbox me is
rejected early with a clear fixup hint instead of a late opaque API
error.
The --as help text was already dynamic via AddShortcutIdentityFlag;
add TC-10/TC-11 tests in internal/cmdutil/identity_flag_test.go to
pin that behaviour, and TC-1 through TC-9 in
shortcuts/mail/mail_shortcut_validation_test.go to cover the new
Validate callbacks.
+watch is excluded: its AuthTypes is ["user"], so bot is never valid.
sprint: S2
* test(cmdutil): add Hidden and DefValue assertions to identity flag tests
* fix(mail): add bot+mailbox=me validation to +template-create and +template-update
* fix(mail): add bot+mailbox=me validation to +template-update
* fix(mail): gofmt mail_template_create.go
* fix(mail): gofmt mail_template_update.go
* fix(mail): skip bot+mailbox=me check for print-patch-template local path
Add a Priority field to DraftProjection populated from the EML header pair
X-Cli-Priority (CLI/OAPI primary) → X-Priority (RFC fallback for IMAP-回灌
historical drafts), with case-insensitive lookup via the existing
headerValue helper and a local mapping table aligned with the backend
gopkg/mail_priority.PriorityValueToType vocabulary. When neither header is
present (the symmetric read of --set-priority normal=remove_header) the
projection emits "unknown" so agents have a stable read-side surface.
Append one notes entry to buildDraftEditPatchTemplate documenting the
--set-priority flag and the X-Cli-Priority translation contract.
The write-side (--set-priority flag, parsePriority helper, translation
branch in mail_draft_edit.go, EML header target) is unchanged — already
shipped on master.
sprint: S4
Test files legitimately need to construct dangerous Unicode inputs
(RLO, ZWSP, BOM, etc.) to verify validation logic rejects them.
bidichk treats decoded \u escape literals as Trojan Source risks,
which is a false positive for intentional test data.
Change-Id: I555028a992ab008da16129eb41075c333d0099b8
- +node-get: wrap wiki.spaces.get_node; accepts node_token, obj_token,
or a Lark URL (URL path auto-infers obj_type); formatted output with
creator / updated_at. No synthesized url — get_node returns none and a
BuildResourceURL fallback is a non-canonical link that misleads in a
read/confirm command (sibling read shortcuts omit it too)
- +node-delete: wrap space.node delete; high-risk-write (--yes gated),
async delete-node task polling, auto-resolves space_id via get_node
when --space-id omitted, actionable hints for codes 131011 / 131003.
The delete-node task result lives under the gateway's generic
`simple_task_result` key (NOT `delete_node_result`)
- +space-create: wrap spaces.create; user-only identity, --name
required (no empty-name spaces), flattened space output, no url
- factor the shared wiki async-task poll loop into wiki_async_task.go;
preserve upstream Lark Detail.Code on poll exhaustion (no longer
rebuilt via lossy ErrWithHint)
- drive +task_result: add wiki_delete_node scenario so +node-delete's
async-timeout next_command actually resolves
- skill docs: reference pages for the 3 new shortcuts + SKILL.md
shortcuts table (no raw nodes.delete API exists — it's shortcut-only,
so it is intentionally absent from API Resources / permission table);
drop the circular TestWikiShortcutsIncludeAllCommands change-detector
Change-Id: I316f78290cec5bc50f80d629173e3bf2a35dd005
* feat(auth): add QR code support for device auth flow
* docs: update login QR code display hints for AI agent
* feat(auth): add ASCII QR code support for auth flow
* docs: add comments for login and auth helper functions
* chore: remove unused qrCodeToBase64 helper function
* fix(auth/login): clarify verification_url handling in login hint
Bidirectional sync between a local directory and a Drive folder with
diff detection (new_local, new_remote, modified, unchanged) and
conflict resolution strategies (--on-conflict: remote-wins, local-wins,
keep-both, ask).
Key behaviors:
- Type conflict detection: hard-fail when local file vs remote non-file
or local directory vs remote file
- Keep-both: rename local with __lark_<hash> suffix, then pull remote;
occupied map includes localDirs to prevent suffix collision
- Local-wins partial-success: prefer returned file_token on upload failure
- Empty directory mirroring: pre-create local dirs on Drive via
drivePushWalkLocal before scope preflight
- Structured errors throughout (output.Errorf / output.ErrWithHint)
Includes unit tests and E2E tests (dry-run + live workflow).
Two DryRun functions in the sheets shortcuts called json.Unmarshal without
checking the return value. This looks like a bug, but Validate already
parses and validates the same --style / --data JSON before DryRun runs,
so the error is structurally impossible at this point.
Use _ = assignment + comment to silence the unchecked-error lint warning
and make the safety invariant explicit to future readers.
Co-authored-by: KhanCold <KhanCold@users.noreply.github.com>
* feat(extension): introduce Plugin / Hook framework with command pruning
Add a single public extension contract under extension/platform: integrators
implement the Plugin interface and register Observers, Wrappers, Lifecycle
handlers, and pruning Rules through the Registrar in one Install call.
Command pruning:
- Rule (Allow / Deny / MaxRisk / Identities) with doublestar globs
- 4-axis AND evaluation, parent-group aggregation, unknown-risk allow
- Sources: Plugin.Restrict (single-rule) and ~/.lark-cli/policy.yml
- Plugin path is fail-closed (envelope on rule error / multiple Restrict);
yaml path is fail-open (warning, CLI continues)
- strict-mode stubs now also write the denial annotation so the hook
layer's denial guard physically isolates Wrap chains on them
- HOME path never leaked through policy_source label
Hook framework:
- Observer (panic-safe, Before/After), Wrapper (middleware, may short-circuit
via AbortError), Lifecycle (Startup + Shutdown only)
- Recover guards every plugin entry point: Capabilities(), Install(),
Wrapper factory composition AND inner Handler, Lifecycle handlers
- namespacedWrap copies AbortError so a plugin's package-level sentinel
is never mutated across concurrent invocations
- Selector unknown-risk uniform: ByExactRisk / ByWrite / ByReadOnly never
match unannotated commands; safety-side hooks opt in via
ByWrite().Or(ByUnknownRisk())
Bootstrap orchestration (cmd/build.go + cmd/policy.go):
- InstallAll uses a staging Registrar + atomic commit
- FailClosed plugin install / Plugin.Restrict conflict / Startup handler
failure each install a structured envelope guard at every dispatch path
- walkGuard neutralises every cobra bypass we know of (PersistentPreRunE
first-wins, ValidateArgs, ParseFlags, legacyArgs, __complete /
__completeNoDesc, non-runnable groups, required-arg subcommands)
- cmd/root.go::Execute calls hook.Emit(Shutdown, runErr) after
rootCmd.Execute; isCompletionCommand skips both __complete and
__completeNoDesc so Tab completion never triggers Shutdown handlers
Capabilities consistency:
- Restricts=true must declare FailurePolicy=FailClosed
- RequiredCLIVersion (semver constraint) is validated against build.Version;
a malformed constraint is treated as untrusted-config and aborts
unconditionally, regardless of FailurePolicy (DEV builds included)
JSON envelope contract:
- error.type closed enum: pruning / strict_mode / hook / plugin_install /
plugin_conflict / plugin_lifecycle
- reason_code closed enums per type, all referenced by structured tests
Bootstrap surfaces (new user commands):
- lark-cli config policy show -- JSON view of the active Rule + source
- lark-cli config policy validate -- parse + schema + glob check, no apply
Coverage:
- extension/platform: every public type has a unit test
- internal/{pruning,hook,platformhost,policydecision,cmdmeta}: full coverage
of denial guard isolation, AbortError sentinel safety, observer panic
safety, lifecycle error/panic typing, staging atomic rollback
- cmd/plugin_integration_test.go: end-to-end through buildInternal with
synthetic and real command trees
- cmd/install_guard_test.go: walkGuard covers auth / config / __complete /
__completeNoDesc / non-runnable parents
* fix(pruning): deny stub must override Args + PersistentPreRunE
The pruning denyStub and the strict-mode stub previously only swapped
RunE plus Hidden + DisableFlagParsing. Cobra's dispatch order means
several pre-RunE gates can fire BEFORE the stub's RunE ever runs:
1. Args validator: shortcut commands often declare cobra.NoArgs.
With DisableFlagParsing=true the user's `--doc xxx --mode append`
looks like positional args, so ValidateArgs surfaces a usage
error instead of the pruning / strict_mode envelope. Observer
hooks also miss the dispatch entirely.
2. Parent PersistentPreRunE: cmd/auth/auth.go declares a
PersistentPreRunE that returns external_provider when env
credentials are set. Cobra's "first PersistentPreRunE wins
walking up from the leaf" then short-circuits with
external_provider instead of the leaf's denial envelope.
Both stubs now also set:
- Args = cobra.ArbitraryArgs (bypass gate 1)
- PersistentPreRunE = no-op leaf hook (bypass gate 2)
- PreRunE / PreRun / PersistentPreRun = nil (defensive)
Effect: dispatch reaches the wrapped RunE, observers fire, the real
pruning / strict_mode envelope is emitted regardless of credential
provider or flag count.
Adds regression tests covering both gates on both stub paths.
* fix(config): policy subcommand bypasses parent's credential check
cmd/config/config.go::NewCmdConfig declares a PersistentPreRunE that
calls f.RequireBuiltinCredentialProvider; with env credentials set,
it returns external_provider for every config subcommand.
`config policy show` and `config policy validate` are READ-ONLY
diagnostic commands -- they inspect or parse the user-layer rule
without touching credentials. They MUST work regardless of which
credential provider is active, otherwise users on env-credential
deployments cannot debug their policy.
Same shape as the codex C11/C13 fix: install a no-op leaf-level
PersistentPreRunE on the `policy` group so cobra's "first walking up
from leaf" rule picks ours over the config parent's.
Regression caught by divergent e2e (F1-F6 all returned external_provider
before this fix; all pass after). Adds a unit test pinning the
PersistentPreRunE override.
* feat(shortcuts): tag service groups with cmdmeta.Domain
RegisterShortcutsWithContext now calls cmdmeta.SetDomain on each
service-level cobra.Command (im, docs, drive, calendar, ...) so the
business-domain axis is actually populated on every shortcut leaf via
parent-chain inheritance.
Before this change, platform.ByDomain("docs") never matched any
command: the domain annotation was unset across the entire shortcut
tree, so the selector's d != "" guard always failed and risk-style
selectors silently degraded to no-op.
The SetDomain call is placed AFTER the create-or-reuse branch so it
fires whether the service command was freshly created here or had
already been added by cmd/service/service.go's OpenAPI auto-
registration (which runs first and creates im, drive, calendar, etc.).
Without this placement only pure-shortcut services like docs would
have been tagged.
Adds a regression test asserting:
- service-group cobra.Command carries the cmdmeta.domain annotation
- leaf shortcuts inherit the domain via parent-chain walk
* feat(diagnostic): add unconditionally allowed command paths for introspection
* feat(plugins): add diagnostic command to inspect installed plugins and their contributions
* fix(cli): surface unknown_subcommand error instead of silent help fallback
When a user passed an unknown subcommand or shortcut (e.g. `lark-cli drive
+bogus`), cobra returned `flag.ErrHelp` for the non-runnable group command,
printed the parent help, and exited 0. AI agents couldn't distinguish a
typo from an intentional help request.
Install a tree-wide guard that attaches a RunE to every group command
without its own Run/RunE. The RunE forwards no-args invocations to help
(preserving prior behavior) and emits a structured unknown_subcommand
ExitError (exit 2) listing available subcommands when args are present.
* refactor(envelope): rename error.type pruning/strict_mode to command_denied
The envelope's `type` field was leaking implementation terms ("pruning",
"strict_mode") that describe enforcement mechanism rather than the user-
facing semantic. It also duplicated `detail.layer`, and forced consumers
to branch on two values for the same conceptual error ("a command was
denied by policy").
Collapse both into a single semantic type "command_denied". The
enforcement layer ("pruning" / "strict_mode") is preserved in
`detail.layer` so debugging and per-layer diagnostics still work.
* feat(platform): fail closed on unannotated/invalid risk when a Rule is active
The pruning engine used to treat any command without a risk annotation as
ALLOW even when a Rule with MaxRisk was set, and would silently skip the
MaxRisk comparison whenever the command's risk string was outside the
closed taxonomy. Both gaps let an unannotated or typo'd write command
slip past an "agent read-only" pruning rule.
Engine now denies before any other axis when a Rule is registered:
- reason_code "risk_not_annotated" for commands with no risk
- reason_code "risk_invalid" for commands whose risk is outside
the read | write | high-risk-write
taxonomy (e.g. typo "wrtie")
Main-flow is preserved: a nil Rule still returns Allowed=true
unconditionally, so a CLI with no pruning plugin behaves identically to
before. ByUnknownRisk() is removed from the public surface since the
Unknown state is no longer reachable through risk-based selectors when
any Rule is active; safety-side widening composition is no longer needed.
* chore(config): hide diagnostic policy/plugins commands from --help
`config policy show`, `config policy validate`, and `config plugins show`
are local-introspection-only commands kept behind the pruning
diagnostic whitelist so operators can always inspect why a command was
denied. They do not need to surface in `--help` for AI agents and were
contributing to help noise.
Hide the `policy` and `plugins` parent groups and both `show` /
`validate` leaves. Commands remain callable by exact name and continue
to bypass user-layer pruning via diagnosticPaths.
* style: gofmt
* fix(platform): nil Selector honours None contract; reject multi-doc policy yaml
- selector.go: And/Or/Not now treat nil Selector as None() per godoc,
preventing runtime panic when composed selectors are invoked.
- schema.go: Parse rejects multi-document YAML input so a stray '---'
separator can't silently drop trailing policy constraints.
* chore: go mod tidy
* feat(extension/platform): plugin SDK with policy engine, hooks, and Builder
Introduces extension/platform — the in-process plugin SDK external
Go forks of lark-cli use to extend or restrict the command surface.
Plugins compile in via blank import; there is no dynamic loading
and no RPC isolation.
Public SDK (extension/platform):
- Plugin interface (Name / Version / Capabilities / Install).
- Registrar verbs: Observe, Wrap, On, Restrict.
- Hook types: Observer (side-effect, panic-safe, fires Before/After
RunE), Wrapper (middleware, may short-circuit via AbortError),
LifecycleHandler (Startup / Shutdown), Selector with nil-safe
And/Or/Not composition.
- Risk / Identity are defined string types with closed taxonomies;
ParseRisk / ParseIdentity convert raw strings with the
absent-vs-invalid distinction the engine relies on.
- Builder ergonomic constructor (NewPlugin().Observer().Wrap()
...MustBuild()) that enforces name/hookName grammar, hookName
uniqueness, and the Restrict ↔ FailClosed pairing regardless of
call order.
- Invocation is a read-only interface; the framework's concrete
invocation type lives in internal/hook so plugins cannot
fabricate denial / strict-mode / identity state. Args() returns
a defensive copy on every call so hook mutation cannot leak
into the original RunE.
- CommandDeniedError + AbortError carry structured fields for the
closed `command_denied` / `hook` envelope contract.
- ResetForTesting gated behind //go:build testing.
- README + godoc examples (Observer / Wrapper / Restrict) + two
runnable example forks (audit-observer, readonly-policy).
Host (internal/platform, internal/hook, internal/cmdpolicy):
- InstallAll: staged plugin registration with atomic commit, panic
isolation, FailOpen / FailClosed semantics, RequiredCLIVersion
semver check, single-Restrict invariant, duplicate-plugin-name
detection.
- hook.Install wraps every runnable cmd.RunE with:
Before observers (panic-safe) → denial guard → composed Wrap
chain → original RunE → After observers (always fire, even on
err). Denied commands physically bypass the Wrap chain so a
plugin Wrapper cannot suppress or rewrite a denial; observers
still see the attempt for audit.
- Recover shim around plugin Wrappers converts panics (including
the factory call) into a structured `hook` envelope with
reason_code=panic; namespacing shim attributes AbortError to
the namespaced hook name.
- cmdpolicy (renamed from internal/pruning) is the user-layer
command policy engine: walks the cobra tree, evaluates each
runnable command against a Rule's four-axis filter (Allow /
Deny / MaxRisk / Identities), produces parent-group aggregate
denials, and installs denyStubs. Rule.AllowUnannotated opts out
of the unannotated-deny gate for gradual adoption; risk_invalid
typos always deny with an edit-distance "did you mean"
suggestion.
- Strict-mode stub in cmd/prune.go composes the shared
detail.* / wrapped CommandDeniedError shape via cmdpolicy
helpers (BuildDenialError / CommandDeniedFromDenial /
DenialDetailMap), so command_denied envelopes from strict-mode
and user-layer policy carry the same closed-enum fields
(detail.layer / reason_code / policy_source). The historical
short Message + independent Hint are preserved unchanged.
- cmdpolicy/yaml: structural parsing of ~/.lark-cli/policy.yml
with KnownFields strict mode, including allow_unannotated.
- `config policy show` / `config policy validate` and the plugin
inventory diagnostic surface the resolved Rule (allow,
deny, max_risk, identities, allow_unannotated) and the hook
contributions per plugin.
Envelope contract (docs/extension/reason-codes.md):
- error.type is a closed set: command_denied, hook, plugin_install,
plugin_conflict, plugin_lifecycle.
- reason_code is a closed enum per error.type, dispatched on by
external agents and CI integrations.
- detail.layer = "policy" | "strict_mode" attributes the rejection.
Build / CI:
- Makefile unit-test / vet / coverage and ci.yml fast-gate +
unit-test + coverage now pass -tags testing so register_testing.go
is visible; ./extension/... is in the package list so the SDK's
own tests actually run.
- fmt-check and examples-build Makefile targets.
- bmatcuk/doublestar/v4 added as a direct dependency for `**` glob
matching in Rule.Allow / Rule.Deny.
Author-facing material:
- docs/extension/ (quickstart, plugin-author-guide, reason-codes)
is provided in the working tree but kept out of git tracking
per repo convention (.gitignore covers docs/).
Change-Id: I3b8ecc2923bd54c2dff19e5dce8a0855a6f9e703
* feat(extension/platform): plugin SDK with policy engine, hooks, and Builder
Introduces extension/platform — the in-process plugin SDK external
Go forks of lark-cli use to extend or restrict the command surface.
Plugins compile in via blank import; there is no dynamic loading
and no RPC isolation.
Public SDK (extension/platform):
- Plugin interface (Name / Version / Capabilities / Install).
- Registrar verbs: Observe, Wrap, On, Restrict.
- Hook types: Observer (side-effect, panic-safe, fires Before/After
RunE), Wrapper (middleware, may short-circuit via AbortError),
LifecycleHandler (Startup / Shutdown), Selector with nil-safe
And/Or/Not composition.
- Risk / Identity are defined string types with closed taxonomies;
ParseRisk / ParseIdentity convert raw strings with the
absent-vs-invalid distinction the engine relies on.
- Builder ergonomic constructor (NewPlugin().Observer().Wrap()
...MustBuild()) that enforces name/hookName grammar, hookName
uniqueness, and the Restrict ↔ FailClosed pairing regardless of
call order.
- Invocation is a read-only interface; the framework's concrete
invocation type lives in internal/hook so plugins cannot
fabricate denial / strict-mode / identity state. Args() returns
a defensive copy on every call so hook mutation cannot leak
into the original RunE.
- CommandDeniedError + AbortError carry structured fields for the
closed `command_denied` / `hook` envelope contract.
- ResetForTesting gated behind //go:build testing.
- README + godoc examples (Observer / Wrapper / Restrict) + two
runnable example forks (audit-observer, readonly-policy).
Host (internal/platform, internal/hook, internal/cmdpolicy):
- InstallAll: staged plugin registration with atomic commit, panic
isolation, FailOpen / FailClosed semantics, RequiredCLIVersion
semver check, single-Restrict invariant, duplicate-plugin-name
detection.
- hook.Install wraps every runnable cmd.RunE with:
Before observers (panic-safe) → denial guard → composed Wrap
chain → original RunE → After observers (always fire, even on
err). Denied commands physically bypass the Wrap chain so a
plugin Wrapper cannot suppress or rewrite a denial; observers
still see the attempt for audit.
- Recover shim around plugin Wrappers converts panics (including
the factory call) into a structured `hook` envelope with
reason_code=panic; namespacing shim attributes AbortError to
the namespaced hook name.
- cmdpolicy (renamed from internal/pruning) is the user-layer
command policy engine: walks the cobra tree, evaluates each
runnable command against a Rule's four-axis filter (Allow /
Deny / MaxRisk / Identities), produces parent-group aggregate
denials, and installs denyStubs. Rule.AllowUnannotated opts out
of the unannotated-deny gate for gradual adoption; risk_invalid
typos always deny with an edit-distance "did you mean"
suggestion.
- Strict-mode stub in cmd/prune.go composes the shared
detail.* / wrapped CommandDeniedError shape via cmdpolicy
helpers (BuildDenialError / CommandDeniedFromDenial /
DenialDetailMap), so command_denied envelopes from strict-mode
and user-layer policy carry the same closed-enum fields
(detail.layer / reason_code / policy_source). The historical
short Message + independent Hint are preserved unchanged.
- cmdpolicy/yaml: structural parsing of ~/.lark-cli/policy.yml
with KnownFields strict mode, including allow_unannotated.
- `config policy show` / `config policy validate` and the plugin
inventory diagnostic surface the resolved Rule (allow,
deny, max_risk, identities, allow_unannotated) and the hook
contributions per plugin.
Envelope contract (docs/extension/reason-codes.md):
- error.type is a closed set: command_denied, hook, plugin_install,
plugin_conflict, plugin_lifecycle.
- reason_code is a closed enum per error.type, dispatched on by
external agents and CI integrations.
- detail.layer = "policy" | "strict_mode" attributes the rejection.
Build / CI:
- Makefile unit-test / vet / coverage and ci.yml fast-gate +
unit-test + coverage now pass -tags testing so register_testing.go
is visible; ./extension/... is in the package list so the SDK's
own tests actually run.
- fmt-check and examples-build Makefile targets.
- bmatcuk/doublestar/v4 added as a direct dependency for `**` glob
matching in Rule.Allow / Rule.Deny.
Author-facing material:
- docs/extension/ (quickstart, plugin-author-guide, reason-codes)
is provided in the working tree but kept out of git tracking
per repo convention (.gitignore covers docs/).
Change-Id: I3b8ecc2923bd54c2dff19e5dce8a0855a6f9e703
* refactor(policy): remove validate command and update diagnostics
* fix(extension/platform): address PR review must-fix items
- cmdpolicy: skip AnnotationPureGroup commands in EvaluateAll,
aggregateParents, and hasRunnableDescendant so user-layer policy
no longer blocks `<group> --help` after the unknown-subcommand
guard attaches RunE to every parent
- cmd/root: tag guarded parent groups with AnnotationPureGroup
- extension/platform: drop `//go:build testing` from register_testing.go
so `go test ./...` works without an extra build tag
- extension/platform/README: inline reason_code reference, fix plugin
lifecycle diagram order (init/Register precede RegisteredPlugins)
- cmd/platform_bootstrap: route userPolicyPath through
core.GetBaseConfigDir so LARKSUITE_CLI_CONFIG_DIR is honoured
- cmdpolicy: add RedactHomeDir helper, fold base config dir and
$HOME prefixes for config policy show + resolver errors
- internal/platform: reject unrecognised FailurePolicy values with
invalid_capability instead of silently fail-open
- cmd/config: surface diagnostic policy/plugins commands in
`config --help` Long text
- CHANGELOG: document command_denied error.type rename and
unknown_subcommand exit-2 behavior change
* fix(extension/platform): address CodeRabbit review comments + CI gofmt
- hook/install: propagate wrapper-injected ctx to invokeOriginal so
RunE/Run see context values added by upstream Wrappers
- hook/testing: SetStderrForTesting returns a restore func; tests now
defer it via t.Cleanup to avoid cross-test sink leakage
- cmdpolicy/active: deep-copy ActivePolicy.Rule on SetActive/GetActive
so callers can't mutate the stored global through shared slices
- platform/inventory: deep-copy Inventory + nested Plugins / HookEntry
/ RuleView slices on SetActiveInventory / GetActiveInventory
- platform/staging: Restrict clones the plugin-supplied Rule before
retaining it so the plugin can't mutate it after Install returns
- platform/version: reject RequiredCLIVersion with more than three
numeric components instead of silently truncating 1.2.3.4 to 1.2.3
- cmd/platform_bootstrap: clear cmdpolicy.SetActive on yaml resolver
error so config policy show doesn't surface a stale rule
- cmd/platform_bootstrap_test: tmpHome pins LARKSUITE_CLI_CONFIG_DIR
so host env can't bleed into the policy test fixtures
- cmdpolicy/apply: installDenyStub returns bool; Apply count no longer
over-reports when strict-mode short-circuits the install
- cmdpolicy/engine: aggregateParents now returns the runnable hybrid's
own denial status when all children are placeholder branches
- cmdpolicy/resolver_test: use t.TempDir()-rooted missing path instead
of hardcoded /nonexistent for hermetic missing-file assertion
- cmd/config/plugins: empty-inventory branch emits total: 0 so the
JSON schema stays stable across populated/empty cases
- cmd/platform_guards_test: select leaf by RunE != nil (not Runnable)
so the test doesn't nil-deref on Run-only commands
- gofmt run on previously committed cmdpolicy/path*.go (CI fast-gate)
* fix(cmdpolicy): replace filepath.Abs with filepath.Clean for lint policy
The depguard / forbidigo rule blocks filepath.Abs in internal/ on the
grounds that it accesses the filesystem (Getwd) directly. Switch
RedactHomeDir + foldPrefix to operate on filepath.Clean strings; real
callers pass already-absolute paths (resolver builds yamlPath via
filepath.Join on the absolute config root), so the redaction outcome
is unchanged for production inputs. Relative inputs fall through to
the unchanged branch — filepath.Rel rejects the mixed-absoluteness
case with an error, which the foldPrefix helper already treats as
"not a hit".
* refactor(cmdpolicy): pure Resolve + drop path redaction & verbose comments
- Resolve becomes a pure function; I/O moves to LoadYAMLPolicy so
precedence selection can be unit-tested without vfs mocks
- ActivePolicy drops YAMLPath; config policy show JSON loses yaml_path
and yaml_shadowed (and the TOCTOU stat that surfaced them)
- RedactHomeDir and path_test.go removed: the home-dir folding was only
earning its keep through the now-deleted yaml_path field
- cmd/build.go bootstrap block trimmed from 71 to 39 lines by cutting
PR-rationale comments; one note kept for the fail-CLOSED-vs-fail-OPEN
business rule
- cmd/config/config.go: parent Long no longer hard-codes hidden command
hints, matching their Hidden:true intent
Change-Id: Icfbb818ce3ef523c63286bfbed34c49be08ed6a2
* refactor(platform): drop StrictMode/Identity from Invocation interface
These two accessors were documented in the public SDK as "After observers
always see ok=true" but the framework never plumbed values to them, so they
always returned ("", false). Zero internal/example/test callers; a plugin
author trusting the doc would silently get wrong behaviour.
Identity is also fundamentally unsuited for Before observers (per-command
identity resolves inside RunE via f.AuthFor, after Before fires). StrictMode
is a global value better placed on a Framework/Environment interface than
per-Invocation. Removing is non-breaking now (no callers); adding later is
non-breaking too.
Change-Id: Ice200543e9bca3bda759ad98a6e34a56df69e915
* fix(prune): preserve original metadata on strict-mode denial stubs
strictModeStubFrom built a fresh *cobra.Command from scratch, dropping
the original command's annotations (risk_level, lark:supportedIdentities,
cmdmeta.domain) and help text. cobraCommandView is a live proxy walking
parent annotations, so after the Remove+Add replacement, audit observers
firing on a strict-mode-denied command saw Cmd().Risk()=("",false) and
Cmd().Identities()=nil -- breaking the first-class use case for
audit/compliance plugins.
Copy child.Annotations into the stub (stamping the denial annotations on
top) and propagate Short/Long for help-text parity with
cmdpolicy/apply.go::installDenyStub, which preserves these by virtue of
mutating in place.
Regression test asserts risk_level / supportedIdentities / Short / Long
all survive replacement, alongside the denial annotations.
Change-Id: I19810a34575996344b63e839066888c154d69335
* chore(platform): align docs with implementation; fold home in yaml warnings
Followup cleanup to the previous three refactor commits, addressing review
fallout where public docs / examples / contract notes still pointed at
deleted symbols or unimplemented designs:
- cmd/build.go: Build() docstring now mentions the plugin install + Startup
emit side effects; Shutdown only fires on Execute path
- extension/platform/doc.go, lifecycle.go, invocation.go: drop references
to the deleted StrictMode/Identity methods, restore minimal Godoc on
Cmd/Args/Started
- extension/platform/view.go, cmd/platform_bootstrap.go,
internal/hook/install.go: rewrite "snapshot before pruning" promise to
match the actual contract (live view + strict-mode stub metadata
preservation)
- cmd/platform_guards_test.go: stubInvocation drops the two old methods
- cmd/platform_bootstrap.go: redactHome() last-mile folds $HOME -> ~ in
warnPolicyError so an os.PathError carrying the absolute policy path
does not leak the user's home dir to stderr / agent / CI logs
- examples/readonly-policy/README.md: drop yaml_path from the sample
`config policy show` envelope (the field was removed in 52cbb92)
Change-Id: I2874cc2cf9225dfa44a9c07b2449149181b387cb
* chore(build): drop vestigial -tags testing from Makefile and CI
The `testing` build tag was introduced in 461e3c6 to gate
extension/platform/register_testing.go (ResetForTesting); PR review
0efee93 then dropped the //go:build testing directive from that file
so downstream `go test ./...` would work without the tag, but never
cleaned the matching tag references out of Makefile and ci.yml.
The result: 8 places passing -tags testing for a tag that nothing in
the repo actually gates, plus a Makefile comment that confidently
claims a gate exists. Net behaviour is identical to omitting the flag;
the only effect is misleading developers into believing there is a
test-only surface separation.
Drop the flag from vet / unit-test / lint / coverage / deadcode (head
+ base worktree) and remove the misleading comment. ResetForTesting's
public-API exposure was the conscious trade-off taken in 0efee93 and
is left untouched.
Change-Id: If0cd78c87d4aec2a2533419fe75b01aae6b165fd
* feat(cmdpolicy): enrich denial Reason with attempted value + rule constraint
The envelope reason for command_denied previously told the caller WHAT
axis failed but not the concrete values on each side, so an AI agent
reading the envelope could not tell which command identity / risk /
path was attempted vs. which the rule permits. The natural temptation
was then to recommend modifying the rule -- exactly the wrong nudge,
since policy exists to prevent the agent from rewriting its own limits.
Each Reason now carries both the attempted value and the rule's
constraint:
identity_mismatch:
"command supports identities [user]; rule allows [bot]"
domain_not_allowed:
"command path \"drive/+upload\" not in allow list [docs/** contact/**]"
command_denylisted:
"command path \"docs/+delete-doc\" matched deny pattern \"docs/+delete-*\""
risk_too_high / write_not_allowed:
"command risk \"high-risk-write\" exceeds rule max_risk \"write\""
risk_not_annotated:
"command has no risk_level annotation; rule denies unannotated commands"
(drops the prescriptive "set allow_unannotated=true" hint -- that
belongs in docs, not in the engine's denial path)
Adds firstMatch() helper so command_denylisted can name the specific
glob that fired; matchesAny() now wraps firstMatch.
Regression test pins the substring contract per reason_code so future
"comment cleanup" cannot silently strip the values out again.
Change-Id: I17c7cc9411f58e3e43ade5e1ce875f3b7fe3e5ea
* fix(cmdpolicy): gofmt engine_test.go
CI fast-gate flagged the test added in 2eb0c2b as unformatted. Local
make unit-test had it cached; should have run `make vet` (which runs
gofmt-equivalent check via fmt-check) before pushing. Trivial 3-line
indent fix.
Change-Id: I42297ae59f607b97b32e976c9ec1c9ec4ab7de21
* feat(cmd): annotate risk_level on all hand-written cobra commands
Without this, any non-empty user-layer policy.yml (default
allow_unannotated=false) denies these commands with reason_code
risk_not_annotated -- bricking auth login, config init, profile use
etc. on first contact with a policy.
cmdpolicy/engine evaluation now resolves to the intended axis (deny
list / allow list / max_risk / identities) instead of failing closed
on the unannotated gate. Policy authors can write `max_risk: write`
or `allow: [auth/** config/** ...]` to express real intent.
Classification:
read auth status/check/list/scopes, config show /
policy show / plugins show, doctor, completion,
schema, profile list, event list/status/schema/
consume
write auth login/logout, config init/bind/remove/
default-as/strict-mode, profile add/remove/
rename/use, event stop/_bus, api (raw transit)
high-risk-write update (replaces the CLI binary; failure can
leave the install broken)
Notes:
- api standalone is conservatively `write`; per-call risk is unknown
at parse time (raw transit), so static gating only enforces the
write-class minimum.
- event _bus is the hidden IPC daemon forked by consume; standalone
invocation by users is not expected, but the annotation keeps
policy evaluation consistent with the other event subcommands.
- The two diagnostic-allowlisted commands (config policy show /
plugins show) still bypass the engine via diagnosticPaths; the
read annotation is for consistency with surrounding leaves.
---------
Co-authored-by: liangshuo-1 <266696938+liangshuo-1@users.noreply.github.com>
The skill doc claimed wiki list/copy shortcuts default to --as user, but
the CLI --as default is `auto` (no --as commonly resolves to bot, listing
the app's spaces instead of the user's). Running `wiki +space-list`
without --as therefore returns app-scoped data, contradicting the doc.
Following the established lark-mail convention (concise user-centric
guidance, not a precedence essay):
- add a short "优先使用 user 身份" section to SKILL.md
- fix the --as rows in lark-wiki-space-list / node-list / node-copy
references to show the real `auto` default and steer to --as user
Change-Id: I539f8d622c1bbad57f8a64c2fc7b7ecc0dfe2116
* fix(drive): preserve parent token on nested overwrite
Ensure drive +push overwrite requests for nested files keep parent_node aligned with the actual remote parent folder and report parent resolution failures explicitly.
* test(drive): cover nested overwrite push workflow
Add a live drive +push workflow case for overwriting a nested remote file so the PR parent-token fix is exercised against the real backend and verified to converge via +status.
* feat(doc): add width/height params to buildBatchUpdateData
Extend buildBatchUpdateData signature with width and height int params.
When mediaType is "image" and either dimension is positive, the value is
included in the replace_image payload. Existing call sites pass 0, 0.
* feat(doc): add --width/--height flags with validation to docs +media-insert
* feat(doc): add aspect-ratio auto-calculation helpers
Add computeMissingDimension (pure ratio math) and detectImageDimensions
(header-only image.DecodeConfig) with PNG/JPEG/GIF blank-import decoders,
plus imageDimensions struct; drive with two new TDD tests.
* feat(doc): wire --width/--height into Execute with aspect-ratio calculation
* feat(doc): add best-effort dimension computation to DryRun
* docs: add --width/--height to docs +media-insert SKILL.md
* fix: add SafeInputPath validation to detectImageDimensionsFromPath
* fix: guard computeMissingDimension against division by zero and add rounding
* fix: add dimension upper bound, fix err variable reuse in Execute
* refactor: use early-return guard for zero native dimensions per review
* fix: add pixels unit to dimension validation error messages
* fix: surface dimension detection failures in dry-run to match Execute behavior
* fix: move dimension detection before upload to fail fast
* fix: restore withRollbackWarning on dimension detection errors in Execute
Dimension detection runs after the placeholder block is created (Step 2),
so failures must clean up the block to avoid leaving an empty placeholder
in the document.
Introduce three new wiki shortcuts that wrap the corresponding raw APIs
with structured flags, formatted output, my_library alias handling, and
unified envelope shape, replacing the bare `lark-cli wiki spaces list`
/ `wiki nodes list` / `wiki nodes copy` flows for the common cases.
Shortcuts
- wiki +space-list (read, scopes: wiki:space:retrieve):
lists wiki spaces. Default fetches a single page; --page-all walks
every page capped by --page-limit (default 10, 0 = unlimited).
Supports --page-size / --page-token / --format json|pretty|table|csv|ndjson.
Output: {spaces, has_more, page_token} + Meta.Count. Pretty mode
distinguishes "no spaces" from "empty page with has_more" and hints
the caller to resume.
- wiki +node-list (read, scopes: wiki:node:retrieve):
lists nodes in a space or under a parent. Same pagination + format
story as +space-list. Accepts the my_library alias for --space-id
with --as user (resolved via a shared resolveMyLibrarySpaceID helper
extracted from +node-create); rejects my_library upfront for --as bot.
- wiki +node-copy (high-risk-write, scopes: wiki:node:copy):
copies a node into a target space or parent. --target-space-id and
--target-parent-node-token are mutually exclusive. Risk is marked
high-risk-write to match the upstream API's danger: true flag, so the
framework requires --yes. Source is preserved; subtree is copied.
Both list shortcuts pick the narrowest scope the upstream API accepts.
The framework's preflight (internal/auth/scope.go MissingScopes) does
exact-string scope matching, so declaring the broader wiki:wiki:readonly
form would wrongly reject tokens that carry only the per-API scope —
which the API itself accepts — and emit a misleading missing-scope hint.
Shared changes
- shortcuts/wiki/wiki_node_create.go: factor out resolveMyLibrarySpaceID
so +node-list and +node-create share one my_library resolution path.
- shortcuts/wiki/shortcuts.go: register the three new shortcuts.
- skills/lark-wiki/SKILL.md and references/lark-wiki-{space,node-list,
node-copy}.md: documentation for the new shortcuts.
Tooling
- scripts/check-doc-tokens.sh + Makefile gitleaks target:
pre-commit check that scans skill reference docs for realistic-looking
Lark token values without the _EXAMPLE_TOKEN placeholder convention,
preventing gitleaks false positives.
- .gitleaks.toml: allowlist tuning.
- .gitignore: ignore .tmp/.
Tests
- shortcuts/wiki/wiki_list_copy_test.go: unit tests covering registry
membership, declared-narrow-scope pinning, flag validation (page-size
range, page-limit >= 0, target flag exclusivity, my_library + bot
rejection), auto-pagination merging, --page-limit truncation
surfacing next cursor, --page-token single-page mode, empty-slice
serialisation, has_more hint pretty rendering, my_library user-path
resolution, +node-copy copy-to-space / copy-to-parent + body shape,
pretty rendering, and the high-risk-write --yes gate.
- tests/cli_e2e/wiki/wiki_shortcut_workflow_test.go: live end-to-end
workflow exercising the shortcut layer against a real tenant.
Reuses an existing my_library node as a host so the test never adds
to the top-layer quota; the copy is placed under the same host node.
- tests/cli_e2e/wiki/coverage.md: shortcut coverage entries added.
Minor cleanups
- skills/lark-doc/references/lark-doc-search.md and
skills/lark-minutes/references/lark-minutes-search.md: replace
realistic-looking example ou_ tokens with _EXAMPLE_ placeholders so
scripts/check-doc-tokens.sh passes.
Change-Id: I9efb0557f477d369d7f26a09c1e154d4ab15b253
Co-authored-by: liujinkun <liujinkun@bytedance.com>
* fix(selfupdate): use LookPath instead of Executable for binary verification (fixes#836)
VerifyBinary was using vfs.Executable() to find the binary to run --version against.
On Linux with global npm install, this returns the inode of the running binary (old version),
not the newly installed one that sits behind npm's bin symlink.
Switch to exec.LookPath("lark-cli") which resolves the PATH entry and follows npm's
bin symlink to the correct newly installed version, matching what the user actually runs.
* test(selfupdate): add LookPath-based tests for VerifyBinary
Add TestVerifyBinaryLookPath, TestVerifyBinaryLookPathNotFound, and
TestVerifyBinaryEmptyOutput. Expose execLookPath variable so tests can
inject a mock LookPath and cover the full VerifyBinary execution path
including version parsing and error branches.
* test(selfupdate): add os/exec import and isolate config dir in VerifyBinary tests
CodeRabbit feedback:
- Add missing os/exec import for execLookPath variable
- Add t.Setenv(LARKSUITE_CLI_CONFIG_DIR, ...) to each new test for config isolation
* test(selfupdate): extract execLookPath to separate lookpath.go
Move the execLookPath variable declaration to its own file so it is
accessible to updater.go without the test-only import cycle.
* fix(selfupdate): remove unused os/exec import from test file
* fix(selfupdate): gofmt + fold lookpath hook and restore version fences
- Move execLookPath into updater.go (drops redundant lookpath.go)
- Document package-level mock: no t.Parallel()
- Extend TestVerifyBinaryLookPath with exact-match regressions (0.0, 12.1.0 vs 2.1.0)
Co-authored-by: CatfishGG <catfishgg@users.noreply.github.com>
* fix(registry): wait for background meta refresh before test reset
TestComputeMinimumScopeSet can start doBackgroundRefresh via Init() while
the next test's resetInit() mutates package-level globals the goroutine
still reads (e.g. remoteMetaURL / configuredBrand), causing data races under
-race in the coverage job.
Track the refresh goroutine with a WaitGroup and drain it at the start of
resetInit() in tests.
* docs: rewrite lark-shared update section to recommend lark-cli update
Change-Id: Ie043b1a32675dcd041f9123503fcccb791cccd07
* feat: add command field to _notice JSON for AI agents
Change-Id: I04b069880f7dca8db384ba8a6919e5682c0382be
* feat: demote npm install to fallback with skills-not-synced warning
Change-Id: If21c3ef6cd1818b28f5578078a04c3627128c6d0
* fix: address CodeRabbit review — guard type assertions, remove npm fallback from SKILL.md
- Add t.Fatalf guards before type-asserting notice sub-maps in
TestSetupNotices_BothUpdateAndSkills to prevent nil-panic on
unexpected shapes.
- Remove the npm fallback section from SKILL.md entirely so AI agents
only see `lark-cli update` as the update path.
- Strip remaining npm mentions from the "重要" note.
Change-Id: Ieb124763b918093e1dcae06f5ea7428dbc248d5f
* fix: add npx skills add hint alongside npm fallback in update paths
When npm is shown as a fallback (manual update path and rollback hint),
append the npx skills add command so users know how to sync skills
separately.
Change-Id: I454172be51073d35def635613a23ad35ba68b5fb
Add im +chat-list shortcut wrapping GET /open-apis/im/v1/chats (previously not exposed via lark-cli).
Add --exclude-muted to both +chat-search and +chat-list: client-side filter that calls POST /open-apis/im/v1/chat_user_setting/batch_get_mute_status after each page and drops is_muted=true chats.
Introduce shortcuts/im/mute_filter.go with pure helpers and an orchestrator (MaybeApplyMuteFilter) shared by both shortcuts.
Change-Id: I22221ac5835667f58cbd40b34de75825d2445d1c
Adds --chat-mode group|topic to lark-cli im +chat-create so users and AI agents can create 话题群 (topic chats) directly via the CLI. Without this, requests to create a topic chat silently fall back to a normal conversation group. Default remains group; chat_mode is now always emitted in the POST /open-apis/im/v1/chats request body.
Change-Id: I79385e2e8606f84e3f27de240d1b41037bf51261
`lark-cli auth login --scope "a,b"` previously sent the raw comma-joined
string to the device authorization endpoint, which treats it as a single
malformed scope and fails with:
device authorization failed: The provided scope list contains invalid
or malformed scopes.
OAuth 2.0 (RFC 6749 §3.3) requires space-delimited scopes on the wire,
but commas are the more natural separator for users typing on a shell
(quoting whitespace is awkward, especially for AI-agent generated
commands). Accept both: split on commas/whitespace, trim, dedupe, then
re-join with single spaces.
Also adds unit tests covering single, comma, space, mixed, dedupe, and
trailing-separator inputs.
Co-authored-by: aj <2072584+meijing0114@users.noreply.github.com>
Five tests in cmd/update mocked SkillsUpdateOverride to return success
and let runSkillsAndStamp call WriteStamp, but did not isolate
LARKSUITE_CLI_CONFIG_DIR. Each run clobbered the real
~/.lark-cli/skills.stamp with the mock version ("2.0.0" or "1.0.0"),
causing skillscheck to fire a misleading drift notice on every
subsequent lark-cli invocation.
Add t.Setenv("LARKSUITE_CLI_CONFIG_DIR", t.TempDir()) at the top of:
- TestUpdateNpm_JSON
- TestUpdateNpm_Human
- TestUpdateForce_JSON
- TestUpdateDevVersion_JSON
- TestUpdateWindows_NpmSuccess_JSON
Scope is limited to tests that mock SkillsUpdateOverride to success;
tests that invoke real npx are pre-existing and out of scope here.
Change-Id: I7a78a6c70f276b51333253acc115e0109c01a851
OpenClaw stores secret file paths in user-authored ~/-relative form so
the configuration stays portable across machines. lark-cli config bind
previously rejected these as non-absolute, blocking users until they
rewrote the OpenClaw config with literal absolute paths.
bind now resolves ~ to the OpenClaw home directory (OPENCLAW_HOME if
set, otherwise the OS home) before the path audit runs, mirroring how
OpenClaw itself reads the same field. Cwd-relative paths and other
unsafe locations are still rejected as before.
Adds shortcuts/mail/flag_suggest.go (~120 LOC) implementing a cobra
FlagErrorFunc hook for the mail subcommand tree. On 'unknown flag: --X'
or 'unknown shorthand flag: "X" in -X', it collects flags from the
current command via cmd.Flags().VisitAll, runs bidirectional prefix
match + Levenshtein DP (threshold=max(1,len/3+1), cap 4), and returns
top-5 candidates inside the existing ErrorEnvelope JSON:
error.type = "unknown_flag"
error.detail.{unknown, command_path, candidates}
error.detail.candidates[*] = {flag, shorthand, distance, reason}
Exit code stays 1 (ExitAPI), not ExitValidation - no breaking change for
CI/agent scripts that check non-zero exit. stderr switches from plain
'Error: unknown flag: --X' to JSON envelope, aligning with the existing
'errors = JSON envelope on stderr' convention; mail unknown-flag was the
last gap.
Scope is strictly the mail subcommand tree: shortcuts/register.go gains
a single 'if service == "mail" { mail.InstallOnMail(svc) }' branch
after the existing Mount loop. Other domains (calendar / im / api /
auth / ...) keep cobra's default FlagErrorFunc and unchanged plain-text
stderr behavior.
Covers:
- shortcuts/mail/flag_suggest.go (new, ~120 LOC)
- shortcuts/mail/flag_suggest_test.go (new, 12 table-driven tests)
- shortcuts/register.go (+3 lines after mail Mount loop)
No changes to cmd/root.go or internal/output/* - ErrDetail.Detail is
already interface{}, handleRootError already routes *ExitError via
WriteErrorEnvelope.
* feat(vc): agent join meeting basic shortcuts structure
Change-Id: Ic5d64067eb48670fa6636841cd00cbfa9b0bf3e7
* docs: add skill references for vc +meeting-join and +meeting-leave
* feat(vc): add meeting events shortcut
Add vc +meeting-events for bot meeting activity queries with page-all pagination support and tested pretty/json output.
* feat(vc): refine meeting events pagination and output
* test: add unit tests for vc +meeting-join and +meeting-leave shortcuts
* feat(vc): improve meeting events pretty timeline
* feat(vc): refine meeting events pretty output
* docs(skill): add vc meeting events shortcut guide
* docs(skill): clarify vc meeting events output guidance
* docs: clarify participant-snapshot vs meeting-events routing
* refactor: split lark-vc-agent from lark-vc
* docs: drop nonexistent workflow skill reference and fix identity
* docs: fix cross-links in lark-vc-agent references after split
* fix(vc): send meeting join password at top level
* docs: rewrite lark-vc-agent description in user-facing language
* docs: tighten lark-vc-agent description to descriptive neutral tone
* fix: use Chinese quotes in vc/vc-agent description YAML frontmatter
* docs: downgrade dry-run from mandatory to optional for vc-agent writes
* docs: clarify pretty vs json format choice by processing depth
* docs: systematic review of lark-vc-agent SKILL for clarity and precision
* feat(vc): print meeting event page token in pretty output
* docs(skill): refine vc agent meeting guidance
* revert: restore CRITICAL banner in lark-vc-agent to match repo convention
* docs: replace inaccurate no-replay warning with real social-cost risk
* docs: tighten meeting-join risk warning to single sentence
* docs: tighten vc-agent references - remove redundancy and fix vague wording
* Revert "docs: tighten vc-agent references - remove redundancy and fix vague wording"
This reverts commit 9845fc40622c65b0811da1c9ae4902434377f33e.
* docs(skill): refine vc meeting events paging guidance
* fix(vc): keep meeting event count aligned with events list
* docs(skill): tighten vc agent meeting events workflow
* refactor(vc): simplify meeting events pagination
* docs(skill): tighten vc agent meeting guidance
* docs(skill): require reading shared docs for meeting summaries
* chore(env): switch default feishu endpoints to pre
* fix(env): use feishu accounts host
* docs(vc): use explicit date in recording example
* revert(env): remove default ppe request header
* chore(env): switch default feishu endpoints to pre
* docs(skill): guide users to early-bird group on agent meeting gray miss
Teach the lark-vc-agent skill to recognize OAPI's new gray-miss signal for
the three agent meeting commands (`+meeting-join`, `+meeting-leave`,
`+meeting-events`) and route the user to the early-bird group instead of
treating it as a permission error.
When CLI stderr JSON returns `error.code=20017 / ErrNotInGray`, the agent
renders the fixed early-bird invite link
`https://go.larkoffice.com/join-chat/2f4nb0e1-fe00-4f67-bed7-25beaf533fbd`.
The user manual is intentionally not surfaced yet.
Scope-related errors still follow the existing `auth login --scope` flow
with no early-bird copy mixed in. lark-shared and other skills are not
touched, so the guidance stays scoped to the agent meeting commands only.
* chore(env): switch endpoints to boe for agent meeting gray testing
* chore(vc-agent): update gray guide and boe endpoints
* docs(vc-agent): refine gray guidance flow
* docs(vc-agent): centralize gray guidance
* fix(ci): stabilize vc output and skill frontmatter
* fix(vc): address review feedback
---------
Co-authored-by: zhaolei.vc <zhaolei.vc@bytedance.com>
Co-authored-by: renaocheng <renaocheng@bytedance.com>
Remove the cold-start _notice.skills that fires whenever
~/.lark-cli/skills.stamp is missing. The stamp is written
exclusively by `lark-cli update`, so users who installed skills via
`npx skills add larksuite/cli -g` (the documented path) saw the
notice on every run despite a fully populated ~/.agents/skills/.
The version-drift notice (stamp != binary) is preserved unchanged
for users who opted into tracking by running `lark-cli update`.
- internal/skillscheck/check.go: Init returns silently on empty stamp
- internal/skillscheck/notice.go: drop dead cold-start branch in Message;
Current field is now guaranteed non-empty
- tests updated in skillscheck package + cmd/root_integration_test.go
to assert the new contract
No new files, no env vars, no JSON schema change. The _notice.skills
shape stays {current, target, message} — only the cold-start message
string is no longer possible.
The +chat-search row in lark-im SKILL.md described the search as
"by keyword and/or member open_ids", which doesn't match the real
flag names (--query, --member-ids). Naming them inline avoids
agents guessing --keyword from the prose, matching the style
already used by +chat-messages-list.
Change-Id: Ife8668d9b13ee66711bc4e81a7b2bcc7f05d9586
Add IM flag shortcut commands to lark-cli, enabling users to create, list, and cancel bookmarks on messages and threads via +flag-create, +flag-list, and +flag-cancel.
Change-Id: I8f87f0eadf83fb59b024a3b9fe67b23d363abe0a
- Assemble applinks via net/url to ensure proper encoding
- Normalize message position values across more numeric types
- Avoid leaking null message_app_link; assemble when missing
- Update unit tests to assert URL semantics and cover edge cases
Change-Id: Ic473cb563c8a648c4f6677c32b25b9f371a0f84e
Adds a new top-level safety section "数据真实性与操作合规" to the
lark-mail skill via the canonical generation pipeline:
- skill-template/domains/mail.md (source) — adds the section to the
domain introduction file that gen-skills.py renders into SKILL.md.
- skills/lark-mail/SKILL.md (regenerated product) — produced by
`make gen-skills project=mail` from larksuite-cli-registry against
the modified mail.md source.
Why both files: skills/lark-mail/SKILL.md is auto-generated from
skill-template/domains/mail.md + registry-conf/skill-meta.yaml +
output/from_meta/mail.json. Editing only SKILL.md would be reverted on
the next `make gen-skills` run because SKILL.md has no AUTO-GENERATED
markers and falls into the "no markers -> overwrite whole file" branch
in scripts/gen-skills.py.
The section adds 3 hard constraints on agent behavior:
- empty result is a valid answer; do not fabricate IDs or placeholders
- explicit action preview before destructive write operations
(delete / trash / batch_trash / cancel_scheduled_send / rules.*)
- reversible modifications (label / read state / folder move) are
exempt from the preview requirement
Addresses recurring evaluation failures (c03/c04/c06/c09/c14/c19~c24/c40)
where the agent fabricated IDs or auto-executed destructive operations.
The --as flag displayed (default "bot"), (default "user"), or
(default "auto") in help text, but ResolveAs() never uses the cobra
default — it resolves identity via credential config and auto-detect.
The displayed default misled users into thinking a fixed identity was
used when --as was omitted.
Set cobra default to empty string so no (default ...) suffix appears.
Also remove "auto" from visible options since --as auto is equivalent
to omitting --as entirely.
Change-Id: I51ba550a6697eb3675a29f5cee4d0010e0a1cc16
Users who install or upgrade lark-cli via make install, go install, or
direct binary download end up with a binary but no AI agent skills,
degrading agent UX. This PR adds a startup-time skills version drift
notice (injected into JSON envelope _notice.skills, mirroring the
existing _notice.update pattern) and unifies lark-cli update's skills
sync across all three branches (npm / manual / already-latest) with
stamp-based dedup, so any explicit update invocation keeps skills in
sync regardless of how the binary was installed.
Changes:
- new internal/skillscheck package: notice (StaleNotice + atomic
pending), stamp (~/.lark-cli/skills.stamp), skip (CI / DEV /
non-release / LARKSUITE_CLI_NO_SKILLS_NOTIFIER opt-out), check
(synchronous Init)
- cmd/root.go: rename setupUpdateNotice -> setupNotices, compose
output.PendingNotice returning {update?, skills?}; capture
build.Version locally before spawning the async update goroutine
- cmd/update/update.go: add runSkillsAndStamp helper with stamp-based
dedup; rewire the three branches through shared applySkillsResult /
emitSkillsTextHints helpers; add skills_status block to --check JSON
output as a pure report (no side effects)
- internal/update: export IsRelease(version) bool / IsCIEnv() bool
for cross-package reuse; refresh UpdateInfo.Message to append
', run: lark-cli update' so both notices recommend the same fix
- AGENTS.md: add Notification Opt-Outs section documenting
LARKSUITE_CLI_NO_UPDATE_NOTIFIER and LARKSUITE_CLI_NO_SKILLS_NOTIFIER
- internal/binding/types.go: bump default exec-provider timeout from
5s to 10s (out-of-scope flake fix for TestResolveExecRef_JSONResponse
under heavy parallel test load)
AI agents running inside OpenClaw / Hermes were routinely creating a parallel
app via `config init --new` instead of binding to the agent's existing app,
because every "not configured" hint and several deny errors hard-coded
`config init` regardless of workspace. Once bound, the same agents could
silently grant themselves user identity (impersonation) without the user
ever seeing a risk message in chat.
Changes:
- Introduce `core.NotConfiguredError` / `NoActiveProfileError` /
`reconfigureHint` helpers that branch on `CurrentWorkspace()`. In agent
workspaces they point at `lark-cli config bind --help` (a help page, not
a ready-to-run command) so AI must read the binding workflow and confirm
identity preset with the user before acting. In local terminals they
preserve the previous `config init --new` guidance.
- Migrate every `config init` hint that should be workspace-aware:
RequireConfigForProfile, default credential provider, credential provider
fallback, secret-resolve mismatch, config show, strict-mode entry-point
errors, default-as, profile use/rename/remove, auth list, doctor's
config_file check (which now also wraps the OS-level "no such file"
noise into the user-shaped "not configured" message).
- Refuse `config init` when run inside an OpenClaw / Hermes workspace by
default; add `--force-init` for the rare case the user genuinely wants
a parallel app. Without this guard, hint fixes were undone the moment
AI ignored them.
- Rewrite the strict-mode deny errors in cmd/auth/login.go, cmd/prune.go,
and internal/cmdutil/factory.go. The previous "AI agents are strictly
prohibited from modifying this setting" terminated AI reasoning while
providing no real gate. New errors point at `config strict-mode --help`
with the legitimate confirmation flow and explicitly note that switching
does NOT require re-bind. Integration test envelopes updated.
- Tighten `config bind --help` and `config strict-mode --help` to encode
the user-confirmation discipline directly: identity preset semantics
(bot-only vs user-default), "DO NOT switch without explicit user
confirmation", and a cross-reference clarifying that `config bind` is
for changing the underlying app while `config strict-mode` is the
policy-only switch (resolves an ambiguity an audit run found).
- Surface user-identity (impersonation) risk at every config write that
newly grants it, by reusing the canonical IdentityEscalationMessage
string from bind_messages.go:
- `noticeUserDefaultRisk` fires on flag-mode bind landing on
user-default, including the first-time case `warnIdentityEscalation`
misses (it requires a previous bot lock).
- `setStrictMode` warns when transitioning bot → user or bot → off
(newly permits user identity); stays quiet on narrowing changes
and on off → user (off already permitted user).
- Add tests: notconfigured_test.go (workspace branches),
init_guard_test.go (refuse + --force-init bypass), bind_warning_test.go
(user-default warning fires; bot-only does not), strict_mode_warning_test.go
(5 transitions covering both warn and no-warn paths).
Two follow-ups intentionally deferred: the keychain master-key hint at
internal/keychain/keychain.go:42 still suggests `config init` because the
keychain package can't import core (would be circular); fixing requires
either parameterizing the hint via callback or extracting workspace into
its own package. The lark-shared skill doc still tells AI to run
`config init` for first-time setup; updating the skill is in scope for
a follow-up PR.
Change-Id: I02273e044d9e061d211ceaa4f3ed5a3fb28325b3
* fix(auth): handle missing scopes and device flow improvements
* fix: remove redundant error return in login scope handler
* test(auth): rename test for zero interval default case
* fix: increase device code polling timeout from 180 to 600 seconds
* feat(base): support batch record get and delete
* fix(base): address batch record PR feedback
* docs(base): refine record skill routing
* refactor(base): use batch record get and delete only
* refactor(base): share record selection normalization
* docs(base): clarify record get field projection help
* feat(drive): pre-flight per-text-element byte limit for +add-comment
The open-platform comment API returns an opaque [1069302] Invalid or
missing parameters whenever a single reply_elements[i] text exceeds
its implicit byte budget. The error does not name which element failed
or that length is the cause, so callers resort to binary-search
debugging.
Empirically: Chinese text up to ~80 chars (~240 bytes) lands; ~130
chars (~390 bytes) fails. Set the pre-flight limit to 300 bytes which
sits safely inside the known-good zone.
- parseCommentReplyElements now rejects any text element whose UTF-8
byte length exceeds 300, with an ExitError naming the element index
(#N, 1-based) and both the rune and byte counts, plus an ErrWithHint
recommending the correct remediation (split into multiple text
elements — the comment UI renders them as one contiguous comment).
- The previous 1000-rune check is removed: it was too lenient (a
Chinese text under that cap would still fail server-side).
- skills/lark-drive/references/lark-drive-add-comment.md documents
the per-element limit and the correct split pattern so agents
avoid constructing oversized single elements upstream.
Addresses Case 12 in the 踩坑列表 doc.
* fix(drive): correct +add-comment hint to match actual escape coverage
`escapeCommentText` only expands `<` and `>` (each → 4 bytes via
`<` / `>`); `&` is intentionally left as-is. Both the over-limit
hint and the inline comment in `parseCommentReplyElements` previously
claimed `&` was also escaped, with a "4-5 bytes each" range that
implicitly assumed `&` (5 bytes) — a string of 300 `&` chars
would actually fit in the budget, but a user reading the hint would
think otherwise and pre-emptively split it.
Code:
- Hint string ends with `Note: '<' and '>' are HTML-escaped and
counted in their escaped form (4 bytes each).` (was: included `&`
and "4-5 bytes")
- Inline comment above the budget check now matches:
`escapeCommentText only expands '<' and '>' (each becomes 4 bytes:
< / >); '&' is intentionally left as-is.`
Tests (regression):
- New `300 ampersands accepted (escapeCommentText leaves '&' as-is)`
subtest pins that 300 `&` chars stay within budget. Without the fix
this also passed (function was always correct), but the hint was
lying — the test pins the budget contract loud and clear.
- New `TestParseCommentReplyElementsHintMatchesEscape` asserts the
hint string itself: must mention `'<' and '>'` / `4 bytes`, must NOT
mention `'&'` / `&` / `4-5 bytes`. Catches a future drift if
`escapeCommentText` is changed without updating the hint, or
vice-versa.
The skill md (`skills/lark-drive/references/lark-drive-add-comment.md`)
already had the right wording (`每个 < 或 > 占 4 字节`), so it was the
in-Go strings that drifted; this commit aligns code with doc.
* fix(drive): rewrite +add-comment length cap to match real server behavior
The original PR set a 300-byte per-element pre-flight check, justified
by the empirical pattern "~80 Chinese chars succeeds, ~130 fails". A
fresh round of probing the live `/open-apis/drive/v1/files/{token}/
new_comments` endpoint with a real docx shows that pattern does not
reproduce, and the actual contract is very different:
- 10000 ASCII / 10000 Chinese / 10000 '<' (escaped to 40000 bytes)
in a single text element: all OK
- 10001 of any of the above in a single text element: [1069302]
- 5000 + 5000 across two text elements (total 10000): OK
- 5000 + 5001 across two text elements (total 10001): [1069302]
- 4000 + 4000 + 4000 across three (total 12000): [1069302]
Two consequences:
1. The cap is *10000 runes total across all reply_elements text*, not
300 bytes per element. The old check rejected legitimate input
anywhere from ~100 to 10000 Chinese chars (≈100x too aggressive).
2. The hint that recommended "split the content across multiple
{\"type\":\"text\",\"text\":\"...\"} elements" was actively wrong —
splitting doesn't bypass a total cap. A user told to split a
10001-char message into 5000+5001 hits the same opaque [1069302].
This commit:
- Replaces `maxCommentTextElementBytes = 300` with
`maxCommentTotalRunes = 10000`. The constant's doc comment records
the probe matrix above so future maintainers know how it was
derived.
- Switches the measurement from `len(escapeCommentText(input.Text))`
to `utf8.RuneCountInString(input.Text)`. Server counts raw runes;
byte width and post-escape form are irrelevant. The escape itself
still happens — `<` and `>` still get rendered literally — but it
no longer participates in the length check.
- Tracks a running `totalRunes` across the whole reply_elements array
and bails at the first element that pushes the cumulative total
over the 10000-rune budget, with index reporting that points at the
offending element.
- Rewrites the over-cap hint to (a) name the actual 10000-rune budget,
(b) explicitly say splitting does NOT help, (c) drop the wrong
"comment UI still renders them as one contiguous comment" framing
that implied splitting was a workaround.
- Adds a `TestParseCommentReplyElementsHintForbidsSplitAdvice`
watchdog that fails if any future drift puts the discredited split
advice back into the hint.
Tests: 11 cases on TestParseCommentReplyElementsTextLength covering
single-element boundary (ASCII / Chinese / angle brackets at exactly
10000 and at 10001), multi-element total cap (5000+5000 OK, 5000+5001
rejected with index pointing at element #2), early-element-overshoot
indexing (first element at 10001 reports index #1, not the trailing
element), and mention_user not double-counting toward the cap.
Skill md updated: removes the 300-byte / "split into multiple
elements" advice; documents the 10000-rune total cap with a note that
the schema currently advertises 1-1000 chars and is out of date,
plus a procedure for re-probing if the server-side limit ever moves.
Manual API verification: rebuilt binary and posted comments at
boundary lengths — all OK cases (100 / 5000 / 10000 chars, 5000+5000
split) accepted by server; over-cap cases (10001 / 10100 single, and
5000+5001 split) rejected by the new pre-flight before reaching the
network.
---------
Co-authored-by: fangshuyu <fangshuyu@bytedance.com>
* feat(doc): expand callout type= shorthand into background-color and border-color
When users write <callout type="warning" emoji="📝"> without an explicit
background-color, the Feishu doc renders the block with no color. This
commit adds fixCalloutType() which maps the semantic type= attribute to
the corresponding background-color/border-color pair accepted by create-doc.
- warning → light-yellow/yellow
- info/note → light-blue/blue
- tip/success/check → light-green/green
- error/danger → light-red/red
- caution → light-orange/orange
- important → light-purple/purple
Explicit background-color or border-color attributes are always preserved.
The fix is applied via prepareMarkdownForCreate() in both +create and
+update paths, and also inside fixExportedMarkdown() for round-trip fidelity.
* refactor(doc): replace silent callout type→color injection with hint output
Per reviewer feedback (SunPeiYang996), silently rewriting user Markdown is
the wrong layer for this adaptation. The type→color mapping is not part of
the Feishu spec, and covert transforms make debugging harder.
Replace fixCalloutType() (which rewrote the Markdown) with WarnCalloutType()
which leaves the Markdown unchanged and instead writes a hint line to stderr
for each callout tag that has type= but no background-color, telling the user
the recommended explicit attributes to add:
hint: callout type="warning" has no background-color; consider: background-color="light-yellow" border-color="yellow"
Also fixes CodeRabbit feedback: the type= regex now accepts both single-quoted
and double-quoted attribute values (type='warning' and type="warning").
* fix(doc): harden background-color detection in WarnCalloutType
CodeRabbit flagged that the previous strings.Contains(attrs,
"background-color=") check missed forms like 'background-color =
"light-red"' with whitespace around the equals sign. Replace with a
regex that tolerates optional whitespace, and add a regression test.
* fix(doc): close real review gaps left over after rebase
PR #467's review thread had three substantive comments
(`fangshuyu-768`, 2026-04-21) that the prior reply messages claimed
were fixed in commit 7d4b556 — but that commit no longer exists on the
branch (lost in a rebase / squash), and the head still ships the
original buggy code. This commit makes the fixes real.
Three behavior fixes in shortcuts/doc/markdown_fix.go:
1. (#5) Tighten the type= and background-color= regex anchors. \b sits
at any word/non-word boundary, and `-` is a non-word char, so
`\btype=` also matched the suffix of `data-type=` — a tag like
`<callout data-type="warning">` would emit a bogus light-yellow
hint. Switched both regexes to `(?:^|\s)…` so a real attribute
separator is required. The same anchor on background-color closes
the symmetric case where a `data-background-color=` attribute
would silently suppress the real hint.
2. (#4) WarnCalloutType is now a fence-aware line walker. Previously
the regex ran over the entire markdown body, so a callout sample
inside a documentation code fence (```markdown … ```) would
generate a phantom stderr hint every time the docs mentioned the
feature. The walker tracks fence state via the existing
codeFenceOpenMarker / isCodeFenceClose helpers from
docs_update_check.go, which handle both backtick and tilde fences
per CommonMark §4.5.
3. (#3) Drop the ReplaceAllStringFunc-as-iterator pattern. The
previous code routed callout iteration through a rewrite primitive
whose rebuilt-string return value was discarded, then ran the same
regex a second time inside the callback to recover the capture
groups. New scanCalloutTagsForWarning helper uses
FindAllStringSubmatch — one pass, no thrown-away allocation,
intent matches the surface (read-only scan, not a mutator).
Tests: 5 new TestWarnCalloutType subtests pin each contract:
- data-type attribute does not trigger hint (#5)
- data-background-color does not suppress hint (#5, symmetric)
- callout inside backtick fence emits no hint (#4)
- callout inside tilde fence emits no hint (#4)
- callout after fence close still emits hint (#4, fence-state reset)
All 14 TestWarnCalloutType cases pass; go vet / golangci-lint
--new-from-rev=origin/main both clean.
* feat(base): add record read SOP guidance
1. Add a unified lark-base record read SOP for get/search/list routing, field projection, temporary view querying, pagination, matrix result binding, and link field reads.
2. Inline command-focused parameter guidance into +record-get, +record-search, and +record-list help, including examples, JSON shape, view scope, projection, and limit constraints.
3. Preserve base shortcut flag order in help output and add tests covering record read help guidance.
4. Remove the single-method record read skill references in favor of the unified SOP.
* test(base): remove stale record list fixture
* fix(base): scan record markdown output
* fix(base): fallback record markdown output
* fix(base): unify base token wording in shortcuts and skills
* feat(drive): add +pull shortcut to mirror a Drive folder onto local
Adds `drive +pull`, a one-way Drive → local mirror command. It
recursively lists --folder-token, downloads each type=file entry
into --local-dir at the matching relative path, and optionally
deletes local files absent from the remote (mirror semantics).
Implementation notes:
- Listing recurses through subfolders with the standard 200-page
pagination loop. Online docs (docx, sheet, bitable, mindnote,
slides) and shortcuts are skipped since there is no equivalent
local binary to write back. Folder tree is reproduced under
--local-dir, with parent directories auto-created by FileIO.Save.
- Per-file --if-exists=overwrite (default) | skip controls how
pre-existing local files are treated; the framework's enum guard
rejects any other value.
- --delete-local is the only destructive flag and is bound to --yes
in Validate: --delete-local without --yes is rejected upfront so
no listing or download even runs. --delete-local --yes performs
downloads first, then walks --local-dir and removes regular files
not present in the remote map. This matches the spec doc's
"high-risk-write" intent for --delete-local without making the
default pull path require confirmation.
- --local-dir is funneled through validate.SafeLocalFlagPath so
errors reference --local-dir instead of the framework default
--file. FileIO().Stat then enforces existence and IsDir.
- Scopes: drive:drive.metadata:readonly + drive:file:download. The
broader drive:drive is disabled by enterprise policy in some
tenants.
- Listing helper (drivePullListRemote) is duplicated locally rather
than reused from drive_status.go because that change is still in
open PR #692; once it merges, both can be lifted into a shared
drive package helper. TODO marker is left in the code.
Tests cover six unit scenarios (happy-path with nested subfolder +
docx skipping, --if-exists=skip, --delete-local rejection without
--yes, --delete-local --yes deletes orphans, absolute-path
rejection, bad enum) and four E2E dry-run scenarios (request shape,
absolute path rejection, --delete-local --yes guard, missing
required flag).
* docs(skills): document drive +pull in lark-drive skill
Adds references/lark-drive-pull.md covering parameters, output schema
(summary + per-item action breakdown), the type=file scoping rule,
the --if-exists policy matrix, and the --delete-local + --yes safety
contract. Calls out the network-traffic caveat (pull is full-download,
unlike +status which only fetches when both sides have the file) and
the cwd boundary on --local-dir.
Wires +pull into the Shortcuts table in SKILL.md.
* fix(drive): walk +pull on canonical absolute root to close symlink/.. escape
Same root cause as the +status fix: --local-dir was validated through
SafeLocalFlagPath but the walk used the user-supplied raw string.
SafeLocalFlagPath returns the original value (the canonical form is
discarded), and SafeInputPath itself relies on filepath.Clean for
normalization, which shrinks "link/.." to "." purely as string
manipulation. The kernel then resolves "link/.." through the symlink
target's parent at walk time, putting the traversal outside cwd.
For +pull the bug is more dangerous than for +status because it
travels through --delete-local --yes — a raw walk would let the
delete pass land on files outside cwd.
Fix:
- In Execute, resolve --local-dir via validate.SafeInputPath to get a
canonical absolute path, and resolve "." the same way for cwd.
- Convert the resolved root back to a cwd-relative form
(filepath.Rel) for download targets so FileIO.Save's existing
SafeOutputPath check (which rejects absolute paths) still applies.
- For --delete-local, walk the canonical absolute root, then delete
via the absolute path. Both values come from the validated
safeRoot, so kernel path resolution cannot redirect a delete to a
file outside the canonical subtree.
- drivePullWalkLocal now returns absolute paths instead of rel paths;
the caller computes the rel_path via filepath.Rel against safeRoot
for output / remote-set membership checks.
Adds TestDrivePullDeleteLocalDoesNotEscapeViaSymlinkParentRef as a
regression: it stages an "escape" sibling directory containing a
sentinel file, adds a "link" symlink in cwd pointing into it, and
runs +pull --delete-local --yes against an empty remote with
--local-dir "link/..". The sentinel must survive (proving --delete
did not escape) and the in-cwd file must be removed (proving the
walk did run).
* test(drive): pin walker / download behavior on +pull symlink corner cases
Adds three regressions on top of the canonical-root walk fix:
- TestDrivePullSkipsSymlinkInsideRoot: a child symlink inside the
validated root pointing to a sibling temp dir. Under
--delete-local --yes with an empty remote, the sentinel inside the
target must survive (walker did not follow the child symlink) and
the in-cwd file must be deleted (walker did run).
- TestDrivePullSurvivesCircularSymlinkInsideRoot: a child symlink
pointing at one of its ancestors. The walk must terminate so the
test does not hang on the per-test timeout.
- TestDrivePullDownloadDoesNotEscapeViaSymlinkParentRef: pins the
download half of the fix. With --local-dir "link/.." the canonical
root resolves to cwd, so the remote file must land in cwd, not
inside the symlink target's parent. The preexisting sentinel inside
the escape directory must remain untouched.
* fix(drive): +pull --delete-local must not unlink local files shadowed by online docs
CodeRabbit (PR #696) flagged that the --delete-local pass treated any
local path missing from `remoteFiles` as orphaned, but `remoteFiles` only
records type=file entries. If Drive held a docx/sheet/shortcut at the
same rel_path as a local file, the local file would be unlinked even
though Drive still owned that path.
drivePullListRemote now returns two views:
- files: rel_path -> file_token, type=file only (download/skip set)
- allPaths: every entry's rel_path regardless of type
The download loop continues to consume `files`; the --delete-local pass
consults `allPaths`, so an online-doc shadow of a local filename keeps
the local file safe.
Also routes the local walk and the delete through the vfs abstraction
(vfs.ReadDir + vfs.Remove) instead of filepath.WalkDir + os.Remove.
This drops the //nolint:forbidigo justifications and lines up with how
internal/keychain and internal/registry already do filesystem I/O. The
recursive vfs.ReadDir walker preserves the same "do not follow child
symlinks" semantics that filepath.WalkDir gave us, so the canonical-root
escape protections in 240b772 stay intact.
Adds TestDrivePullDeleteLocalPreservesLocalFileShadowedByOnlineDoc as a
direct regression: Drive serves keep.txt (file) plus notes.docx (docx),
local has both keep.txt and a hand-edited notes.docx; --delete-local
--yes must download keep.txt, leave notes.docx untouched, and report
deleted_local=0.
* fix(drive): count +pull delete failures in summary.failed
CodeRabbit (PR #696) flagged that both delete_failed branches in the
--delete-local pass appended an item but left the `failed` counter at
zero, so the JSON summary could legitimately report `"failed": 0` after
a partially-failed mirror. Increment failed in both branches (the
filepath.Rel error path and the vfs.Remove error path) so summary.failed
reflects every item flagged delete_failed in items[].
Adds TestDrivePullDeleteLocalCountsFailureInSummary, which forces
vfs.Remove to fail by chmod-ing the local dir 0o555 right before the
run and restoring 0o755 in t.Cleanup so t.TempDir teardown still works.
* fix(drive): swap +pull walk/remove back to filepath/os to satisfy depguard
The previous fix-up commits used vfs.ReadDir + vfs.Remove inside the
+pull shortcut, which depguard's "shortcuts-no-vfs" rule rejects:
shortcuts cannot import internal/vfs directly. CI lint failed on the
import line.
Restore the same pattern used in drive_status.go and the prior +pull
walker:
- filepath.WalkDir to enumerate files under the canonical absolute
root, gated by //nolint:forbidigo with a comment explaining why.
- os.Remove for the actual delete, also gated by //nolint:forbidigo.
The canonical-root safety still holds: validate.SafeInputPath bounds
the walk root inside cwd before WalkDir runs, and WalkDir's default
"do not follow child symlinks" policy is preserved. The two earlier
fixes (drivePullListRemote returning allPaths so online-doc shadows
do not look orphaned, and incrementing failed on delete_failed) stay
in place.
`go test ./shortcuts/drive/...` and `golangci-lint run
--new-from-rev=origin/main` are both clean.
* fix(drive): record remote folder rel_path in +pull allPaths
Follow-up to 45fe4e3. The folder branch in drivePullListRemote merged
descendant rel_paths into allPaths but never recorded the folder's own
rel_path, so a local regular file with the same name as a remote
folder still looked orphaned and got unlinked under --delete-local.
Adds the missing allPaths[rel] for the folder case and a regression:
TestDrivePullDeleteLocalPreservesLocalFileShadowedByRemoteFolder
stages a Drive containing a folder named shadow alongside a
downloadable file, with the local side holding a regular file named
shadow; --delete-local --yes must download keep.txt and leave the
shadow file untouched.
* fix(drive): +pull pagination + dir/file conflict + skill doc symlink claim
Codex review on PR #696 surfaced three issues; addressed in one go:
1. drivePullListRemote only honored next_page_token. The shared
common.PaginationMeta helper accepts both page_token and
next_page_token; switched +pull over so a backend reply using
page_token no longer makes the lister stop at page 1 (which would
silently drop later remote files from both download and
--delete-local).
2. --if-exists=skip swallowed mirror conflicts. The skip/overwrite
branch only checked Stat success, so a local directory shadowing a
remote regular file was reported as action=skipped. Now Stat's
IsDir() is checked first; the conflict surfaces as action=failed
with a message naming the directory, under both --if-exists=skip
and --if-exists=overwrite, and increments summary.failed.
3. Skill doc told callers to soft-link the target into cwd if they
wanted to pull from outside cwd. That is wrong: SafeInputPath
evaluates symlinks before the cwd check, so a symlink pointing
out-of-tree is rejected. Replaced the bogus shortcut with the
actually viable options (switch the agent working directory,
physically move/copy the target, or skip the comparison).
Two new regressions:
- TestDrivePullSurfacesDirectoryFileMirrorConflict — table test over
both policies asserting failed=1, no skipped, action=failed, plus
the 'is a directory' hint in the error message.
- TestDrivePullPaginationHandlesPageTokenField — first page returns
page_token (not next_page_token) with has_more=true; asserts both
pages are fetched and both files land on disk.
* fix(drive): +pull exits non-zero on item failures; gate --delete-local
Two PR-696 review fixes:
- Item-level failures (download error, dir/file conflict, delete error)
now surface as a structured partial_failure ExitError instead of a
success envelope with summary.failed > 0. Exit code becomes non-zero
and error.detail still carries the {summary, items[]} payload, so
AI / script callers can detect the failure via the exit code without
reaching into the JSON body.
- A failed download pass now skips the --delete-local walk entirely.
Previously +pull would continue removing local-only files even when
the download phase had partially failed, leaving the mirror in a
half-synced state (some Drive files missing locally AND some
local-only files unlinked). Re-runs after fixing the download error
recover cleanly.
Skill doc / shortcut description / flag desc updated to call the
operation a one-way file-level mirror, since --delete-local only
unlinks regular files and does not prune empty local directories left
behind by remote folder deletes (true directory-level mirroring is
explicitly out of scope).
Tests: existing dir/file-conflict and delete-failure cases now assert
the partial_failure ExitError shape; new test covers the
"download fails => --delete-local skipped" gating contract.
* refactor(drive): consolidate folder-listing helpers into listRemoteFolder
Closes the post-#692 / post-#709 TODO that lived in drive_pull.go (and
the matching note in drive_push.go): both #692 and #709 are now on main,
so the three near-identical recursive Drive folder listers can collapse
into one.
New shared helper in shortcuts/drive/list_remote.go:
driveRemoteEntry { FileToken, Type, RelPath }
listRemoteFolder(ctx, runtime, folderToken, relBase) -> map[rel]entry
Returns one entry per Drive item (every type), keyed by rel_path.
Subfolders are descended into and the folder's own entry is recorded so
callers can reason about "this rel_path is occupied by a folder"
without re-listing. Pagination via common.PaginationMeta is unchanged.
Each shortcut now derives its own per-shortcut view from the unified
listing:
- drive_status.go: collapses to remoteFiles (Type=="file" -> token) for
the content-hash diff.
- drive_pull.go: derives remoteFiles (Type=="file") for the download
set, plus remotePaths (every rel_path) as the --delete-local guard.
- drive_push.go: derives remoteFiles (Type=="file") for upload /
overwrite / orphan-delete, plus remoteFolders (Type=="folder") for
the create_folder cache. drivePushRemoteEntry was a duplicate of
driveRemoteEntry's first two fields and is dropped; the few call
sites that read .FileToken keep working unchanged.
Per-shortcut copies removed:
- drive_status.go: listRemoteForStatus, joinRelStatus,
driveStatusListPageSize/FileType/FolderType
- drive_pull.go: drivePullListRemote, drivePullJoinRel,
drivePullListPageSize/FileType/FolderType
- drive_push.go: drivePushListRemote, drivePushJoinRel,
drivePushListPageSize/FileType/FolderType, drivePushRemoteEntry
drive_push_test.go's TestDrivePushHelpersRelPath is retargeted at the
shared joinRelDrive; the docstrings on the same-name-conflict tests
were tweaked to refer to "the remoteFiles view" instead of the
just-removed drivePushListRemote.
Net diff: +1 new file, -207 net lines across the four touched files.
All existing unit + e2e dry-run tests pass without behavioral change;
the rel_path / pagination / type-filter contracts each shortcut depends
on are preserved by construction.
* feat(cmdutil): support @file for --params/--data (issue #705)
Inline JSON values for --params/--data are mangled by Windows
PowerShell 5's CommandLineToArgvW. Stdin (-) was the only escape
hatch but supports just one flag at a time.
Extend ResolveInput to accept @<path> (read JSON from a file) and
@@... (escape for a literal @-prefixed value), mirroring the
shortcuts framework's resolveInputFlags semantics. With this, both
--params and --data can be sourced from files in the same call,
sidestepping shell quoting on every platform.
- internal/cmdutil/resolve.go: add @path / @@ handling, trim file
content like stdin does, error on empty path or empty file
- internal/cmdutil/resolve_test.go: cover file read, whitespace
trim, missing file, empty path, empty content, @@ escape, plus
ParseJSONMap / ParseOptionalBody integration through @file
- cmd/api/api.go, cmd/service/service.go: update --params/--data
help text to mention @file
Change-Id: I366aa0f5783fbec6f05403f7f542505098a98c82
* refactor(cmdutil): route @file through fileio.FileIO abstraction
The first cut of @file support called os.ReadFile directly inside
ResolveInput, bypassing the codebase's fileio.FileIO abstraction
(SafeInputPath validation, pluggable provider). That diverged from
how every other file-reading path works: BuildFormdata for --file
uploads and the shortcuts framework's resolveInputFlags both go
through fileio.FileIO.Open with explicit fileio.ErrPathValidation
handling.
Re-route @file through the same path:
- ResolveInput, ParseJSONMap, ParseOptionalBody now take a
fileio.FileIO; @path uses fileIO.Open which goes through
SafeInputPath (control-char rejection, abs-path rejection,
symlink-escape check) — same security posture as --file
- cmd/api and cmd/service callsites pass
Factory.ResolveFileIO(ctx); the upload path now reuses the
resolved fileIO instead of resolving twice
- Path-validation errors surface as
`--params: invalid file path "...": ...` distinct from
`--params: cannot read file "...": ...` for genuine I/O errors
- Nil fileIO with an @path returns a clear
"file input (@path) is not available" error
- Tests use localfileio.LocalFileIO with TestChdir(t, dir),
matching the existing fileupload_test.go pattern; absolute-path
rejection and nil-fileIO are covered
This makes the feature behave identically under any FileIO
provider (including server mode) instead of being silently bound
to the local filesystem.
Change-Id: I878c4e8fb03f43f1f19afad75ec3af9cdab7a7f9
* refactor(cmdutil): share at-file input handling
Change-Id: I92a6eb6ea8fd02054bf8f4925cd81807449d5e51
* feat(drive): add +push shortcut for one-way local → Drive mirror
Mirrors a local directory onto a Drive folder: walks --local-dir,
recursively lists --folder-token, mirrors local subdirectory structure
(including empty dirs) onto Drive via create_folder, and for each
rel_path uploads new files, overwrites already-present files, or skips
them per --if-exists. With --delete-remote --yes, any Drive type=file
entry absent locally is removed; Lark native cloud docs (docx/sheet/
bitable/mindnote/slides) and shortcuts are never overwritten or deleted.
Overwrite hits POST /open-apis/drive/v1/files/upload_all with the
existing file_token in the form body and the response's `version` is
propagated to items[].version, mirroring the markdown +overwrite
contract. Files >20MB fall back to the 3-step
upload_prepare/upload_part/upload_finish path with a single shared fd
reused via io.NewSectionReader per block.
Output is a {summary, items[]} envelope; items[].action is one of
uploaded / overwritten / skipped / folder_created / deleted_remote /
failed / delete_failed.
--delete-remote is bound to --yes upfront in Validate, same pattern as
+pull's --delete-local: a stray flag never silently deletes anything.
Path safety reuses the canonical-root walk + SafeInputPath mechanics
from the sibling +status / +pull commands.
Scopes: drive:drive.metadata:readonly + drive:file:upload +
space:folder:create. space:document:delete is intentionally NOT in the
default set — the framework's pre-flight scope check would otherwise
block plain pushes and dry-runs for callers that haven't granted delete;
--delete-remote --yes relies on the runtime DELETE call to surface
missing_scope. The skill ref calls out the scope so users running
mirror sync can grant it upfront.
13 unit tests cover the upload/overwrite/skip/delete matrix, online-doc
protection, same-name conflict between local file and native cloud doc,
empty-directory mirroring, multipart, scope/path validation, and helper
correctness. 4 dry-run e2e tests pin the request shape.
* fix(drive +push): address review — failure semantics, default skip, scope pre-check, mirror wording
- Item-level failures now bump the exit code via output.ErrBare(ExitAPI)
while keeping the structured items[] envelope on stdout. The
--delete-remote phase is skipped entirely when any upload / overwrite /
folder step fails, so a partial upload never proceeds to delete remote
orphans (a half-synced state).
- Default --if-exists flipped from "overwrite" to the safer "skip": the
upload_all overwrite-version protocol field is still rolling out, so
the default no longer fails a first push against a pre-populated
folder. Callers must opt into "overwrite" explicitly.
- --delete-remote --yes now triggers a conditional space:document:delete
scope pre-check in Validate via the new RuntimeContext.EnsureScopes
helper, so a missing grant fails the run before any upload — instead
of after the upload phase, which would leave orphans uncleaned.
- Description, Tips and skill doc rewritten to call this a file-level
mirror (not a directory mirror): the command does not remove
remote-only directories or close gaps in directory structure that
exists only on Drive.
Tests:
- new TestDrivePushDefaultsToSkipForExistingRemote pins the new default
- new TestDrivePushSkipsDeleteAfterUploadFailure pins the half-sync
guard and the non-zero exit on item-level failure
- new TestDrivePushExitsZeroOnCleanRun pins the inverse
- existing tests that relied on the old overwrite default now opt in
explicitly with --if-exists=overwrite
- TestDrivePushOverwriteWithoutVersionFails updated to assert
*output.ExitError with Code=ExitAPI
- new TestDrive_PushDryRunAcceptsDeleteRemoteWithYes (e2e) symmetric to
the existing reject-without-yes test, pinning that EnsureScopes is a
silent no-op when the resolver has no scope metadata
* fix(drive +push): close remaining CodeRabbit comments
Three small follow-ups on the +push review thread that were still
open after the earlier failure-semantics / default-skip / scope
pre-check fix:
- drivePushUploadAll now extracts data.file_token before checking
larkCode, and surfaces the returned token on the partial-success
path (non-zero code + non-empty file_token). Without this, a backend
response where bytes already landed but code != 0 would force the
caller to fall back to entry.FileToken and silently lose the actual
Drive token, defeating the overwrite-error token-stability handling
in Execute.
- TestDrivePushOverwriteWithoutVersionFails switched from "tok_keep"
to "tok_keep_new" in the upload_all stub and now asserts that the
returned token (not entry.FileToken) lands in items[].file_token —
pins the contract that a regression to the fallback branch would
otherwise pass silently.
- New TestDrivePushOverwritePartialSuccessSurfacesReturnedToken pins
the new partial-success branch end-to-end.
- drive_push_dryrun_test.go: tightened the three Validate / cobra
rejections from `exit != 0` to exact codes — `exit == 2` for the
two Validate-stage rejections (--local-dir absolute,
--delete-remote without --yes), `exit == 1` for the cobra
required-flag check (--folder-token missing). Locks in failure
classification so a regression that misroutes the error layer
doesn't slip through.
* feat(drive): add +status shortcut for content-hash diff
Adds `drive +status`, a read-only diff primitive that walks --local-dir,
recursively lists --folder-token, and reports four buckets — new_local,
new_remote, modified, unchanged — by SHA-256 content hash.
Implementation notes:
- Drive's list/metas APIs do not expose a content hash, so files
present on both sides are downloaded via DoAPIStream and hashed in
memory (sha256 + io.Copy, no disk write). Files only on one side are
not fetched. The command stays Risk: "read".
- Only Drive entries with type=file participate. Online docs (docx,
sheet, bitable, mindnote, slides) and shortcuts are skipped — there
is no equivalent local binary to hash against.
- --local-dir is funneled through the framework's
validate.SafeLocalFlagPath helper so that absolute paths and any ..
that escapes cwd are rejected with --local-dir in the error message
(rather than the internal default --file). FileIO().Stat() then
enforces existence and the IsDir check.
- Local walk uses filepath.WalkDir behind a //nolint:forbidigo comment.
The runtime FileIO interface has no walker today and shortcuts can't
import internal/vfs; SafeInputPath has already bounded the walk root
inside cwd, so the bare walk is acceptable until a runtime-level
walker lands.
- Scopes: drive:drive.metadata:readonly (list folders) +
drive:file:download (fetch files for hashing). The broader
drive:drive scope is disabled by enterprise policy in some tenants;
this narrower pair was verified end-to-end.
Tests cover the four-bucket categorization with a nested subfolder and
docx/shortcut filtering, plus validation errors for missing local-dir,
non-directory local-dir, and absolute-path local-dir.
* docs(skills): document drive +status in lark-drive skill
Adds references/lark-drive-status.md covering parameters, output
schema, the type=file scoping rule, and the network-traffic caveat
(hash is streamed in memory, but bytes still cross the wire).
Notes that --local-dir is bounded to cwd by the CLI's path validation,
and that when a user wants to compare a directory outside cwd the
agent should ask the user to relocate or to switch the agent's working
directory rather than `cd`-ing on its own.
Wires +status into the Shortcuts table in SKILL.md.
* test(drive): cover --folder-token validation and add +status dry-run E2E
Addresses two CodeRabbit review comments on PR #692:
- Adds TestDriveStatusRejectsEmptyFolderToken and
TestDriveStatusRejectsMalformedFolderToken so the Validate-stage
required-check and the ResourceName format guard for --folder-token
are exercised, not just --local-dir.
- Adds tests/cli_e2e/drive/drive_status_dryrun_test.go which drives
the real binary in dry-run mode and asserts:
* the request shape (GET /open-apis/drive/v1/files with
folder_token in the dry-run envelope), plus the description text,
* --local-dir absolute paths are rejected by Validate (which still
runs under --dry-run) with --local-dir surfaced in the message,
* cobra's required-flag enforcement rejects a missing
--folder-token before any custom validation.
* fix(drive): walk +status on canonical absolute root to close symlink/.. escape
Reported in PR review: --local-dir was validated through
SafeLocalFlagPath, but the actual walk used the user-supplied raw
string. SafeLocalFlagPath returns the original value (it only checks
the path through SafeInputPath and discards the canonical form), and
SafeInputPath itself relies on filepath.Clean for path normalization.
filepath.Clean shrinks "link/.." to "." purely as string manipulation,
so the validator sees a path inside cwd. The kernel, however, resolves
"link/.." through the symlink target's parent — which is outside cwd
and is what filepath.WalkDir actually traverses.
Fix: in Execute, resolve --local-dir via validate.SafeInputPath to
get the canonical absolute path (this one fully evaluates symlinks
across the entire path), and walk that path. Each absolute walk hit
is converted to a cwd-relative form via filepath.Rel against
validate.SafeInputPath(".") so FileIO.Open's existing SafeInputPath
guard (which rejects absolute paths) still applies.
Adds TestDriveStatusDoesNotEscapeViaSymlinkParentRef as a regression:
it stages an "escape" sibling directory containing a sentinel file,
adds a "link" symlink in cwd pointing into the escape directory, and
runs +status with --local-dir "link/..". Without this fix, the raw
walk visits the sentinel and leaks it into new_local; with the fix,
the walk stays inside the canonical cwd.
A standalone repro confirms the underlying behavior: raw
filepath.WalkDir("link/..", ...) traversed dozens of unrelated files
in the kernel-resolved parent directory; walking the canonical root
visits only the legitimate cwd contents.
* test(drive): pin walker behavior on child / circular symlinks for +status
Adds two corner-case regressions to back up the canonical-root walk fix:
- TestDriveStatusSkipsSymlinkInsideRoot: a child symlink under
--local-dir that points to a sibling temp dir outside cwd. WalkDir's
default policy must report it as a non-regular entry so the callback
skips it, and the sentinel inside the target must not surface in
new_local. This pins the contract our caller relies on (walk
declines to follow child symlinks even when the canonical root
resolves cleanly).
- TestDriveStatusSurvivesCircularSymlinkInsideRoot: a child symlink
pointing back at one of its ancestors. The walk must terminate and
surface the legitimate sibling file; if WalkDir ever followed the
loop, the per-test timeout would catch it.
* fix(drive): close +status review gaps from Codex (pagination, doc, live E2E)
Three independent fixes flagged on PR #692:
1. Route the recursive Drive folder listing through common.PaginationMeta
instead of reading next_page_token directly. The shared helper accepts
both page_token and next_page_token, matching what okr/im already do
and keeping +status safe against a backend field rename. Adds
TestDriveStatusPaginatesRemoteListing, which serves a 2-page response
where page 1 advertises the cursor as next_page_token and page 2 as
page_token; either spelling alone would silently drop one page.
2. The skill doc previously suggested "or symlink the target into cwd"
as a workaround for cwd-relative --local-dir. SafeInputPath calls
filepath.EvalSymlinks before checking isUnderDir(canonicalCwd), so
any symlink whose final target sits outside cwd still gets rejected
as `unsafe file path`. Rewrite the section so agents stop steering
users into a path that always errors out.
3. Add tests/cli_e2e/drive/drive_status_workflow_test.go — the live
E2E that AGENTS.md requires for new shortcuts. Seeds a real Drive
folder with three uploaded files (unchanged.txt, modified.txt,
remote-only.txt), seeds a local tree with matching/diverging
content plus a local-only.txt, runs +status, and asserts each of
the four buckets contains exactly the file we expect with the
right file_token. Cleanup of every uploaded file plus the parent
folder is registered through the existing best-effort cleanup
helpers. Coverage table bumped: drive +status moves to ✓ and the
denominator goes from 28→29 to account for the new shortcut.
Codex also flagged the local-side filepath.WalkDir as a vfs-bypass.
Investigated: the depguard rule shortcuts-no-vfs explicitly forbids
shortcuts from importing internal/vfs (see commit c1b0bed on the
+pull branch where the same migration was rejected by CI). The
filepath.WalkDir + nolint:forbidigo pattern in walkLocalForStatus is
the lint-required convention until FileIO grows a walker, so leaving
it as-is.
Support minutes +upload to generate a minute from an uploaded media file token.
Change-Id: I59c0719a39541134e395a23262aea7f387105715
Co-authored-by: calendar-assistant <calendar-assistant@users.noreply.github.com>
## Summary
Add explicit guidance on the parent `docs` command so agents pick the right
lark-doc API version. Without this, agents that have an older lark-doc skill
installed can mistakenly mix v2 flags into a v1 flow.
## Changes
- Add `--api-version` help flag and a Tips section to `docs` so `lark docs --help`
(and `--api-version v2`) explain when v2 should be used.
- Refresh the lark-doc skill references and `docs_fetch_v2` keyword flag
description for clarity.
- Add `shortcuts/register_test.go` covering the new docs help wiring.
## Test Plan
- [x] Unit tests pass (`go test ./shortcuts/...`)
- [x] Manual local verification confirms the `lark docs --help` and
`lark docs --help --api-version v2` commands work as expected
## Related Issues
- None
Change-Id: Id3b3196e6a069bb52f95a6fc679b8258313faf3d
The Windows extraction step relied on `powershell -Command Expand-Archive`,
which fails when:
- Microsoft.PowerShell.Archive (a script module) cannot be loaded due to
PSModulePath shadowing (Store-installed pwsh injecting WindowsApps
paths) or ExecutionPolicy Restricted (issue #603), or
- the temp directory contains characters that corrupt PowerShell string
parsing (e.g. a single quote in TEMP).
Switch to a two-tier extraction:
1. Primary: Add-Type System.IO.Compression.FileSystem +
[ZipFile]::ExtractToDirectory. Bypasses the PowerShell module system
entirely. .NET 4.5+, available on Win 8 / Server 2012 by default and
widely on Win 7 SP1.
2. Fallback: Expand-Archive -LiteralPath, kept for the rare host without
.NET 4.5 but with PS 5.0+ (e.g. Win 7 SP1 with WMF 5).
Both paths pass file paths through env vars ($env:LARK_CLI_ARCHIVE /
$env:LARK_CLI_DEST) so quoting / wildcard chars in the path can no longer
break command parsing. -LiteralPath ensures Expand-Archive treats the value
literally rather than as a wildcard pattern. $ErrorActionPreference='Stop'
makes non-terminating cmdlet errors propagate as non-zero exit codes.
Also drop `stdio: "ignore"` so the actual PowerShell error surfaces in the
postinstall log when both paths fail, instead of leaving users with
"Command failed: powershell ..." with no detail.
Verified on Windows 10 + PS 5.1:
- Reproduced #603 with shadow Microsoft.PowerShell.Archive +
Restricted ExecutionPolicy: original install.js fails, patched
install.js succeeds.
- Reproduced single-quote-in-TEMP path corruption: original fails,
patched succeeds.
- Fallback path verified end-to-end with primary forced to fail.
- Normal-environment install: no regression.
* feat(contact +search-user): add --queries multi-name fanout
Add --queries CSV flag to lark-cli contact +search-user for parallel
multi-name fanout (up to 20 entries, partial-failure tolerant).
Output shape in fanout mode:
- data.users[] rows carry matched_query (string)
- data.queries[] sidecar lists each input with {query, error?, has_more}
- top-level data.has_more removed (per-query in queries[])
- error is omitempty; absent on success
Single --query mode is byte-for-byte unchanged (regression-guarded).
--queries is mutually exclusive with --query and --user-ids; bool
filters propagate to every sub-request.
Workers run with WaitGroup + buffered semaphore + index-slot writes;
each has defer recover() converting panics to internal error: ... in
the sidecar (no stack to stderr). Pre-canceled context returns
context canceled without making the request.
All-failed exit propagates first failure's HTTP/API code via ErrAPI;
falls back to ExitInternal for transport/parse/panic/ctx-canceled
(avoids emitting code 0, which means success in the Lark protocol).
HTTP non-200 ErrMsg now includes truncated response body for diagnosis.
Drive-by: signature field is now omitempty (mostly empty in practice).
Infrastructure:
- internal/httpmock gains BodyFilter/OnMatch/Reusable/CapturedBodies
hooks to support concurrent stub-driven tests
- internal/output adds 'users' to knownArrayFields so CSV picks the
primary array correctly
Change-Id: I3c14195fb8e094ae150002d90c36a0e4a0cc97d0
* fix(config/init): use parseBrand(opts.Brand) instead of hardcoded BrandFeishu in --new mode
The --new flag was ignoring the --brand flag and always passing BrandFeishu
to runCreateAppFlow. Now it correctly uses parseBrand(opts.Brand) to
respect the user's --brand parameter (e.g., --brand lark for international).
Change-Id: I1d4d78b3d586142b0210e6ceaeeb467b14e9c1a1
Add --queries CSV flag to lark-cli contact +search-user for parallel
multi-name fanout (up to 20 entries, partial-failure tolerant).
Output shape in fanout mode:
- data.users[] rows carry matched_query (string)
- data.queries[] sidecar lists each input with {query, error?, has_more}
- top-level data.has_more removed (per-query in queries[])
- error is omitempty; absent on success
Single --query mode is byte-for-byte unchanged (regression-guarded).
--queries is mutually exclusive with --query and --user-ids; bool
filters propagate to every sub-request.
Workers run with WaitGroup + buffered semaphore + index-slot writes;
each has defer recover() converting panics to internal error: ... in
the sidecar (no stack to stderr). Pre-canceled context returns
context canceled without making the request.
All-failed exit propagates first failure's HTTP/API code via ErrAPI;
falls back to ExitInternal for transport/parse/panic/ctx-canceled
(avoids emitting code 0, which means success in the Lark protocol).
HTTP non-200 ErrMsg now includes truncated response body for diagnosis.
Drive-by: signature field is now omitempty (mostly empty in practice).
Infrastructure:
- internal/httpmock gains BodyFilter/OnMatch/Reusable/CapturedBodies
hooks to support concurrent stub-driven tests
- internal/output adds 'users' to knownArrayFields so CSV picks the
primary array correctly
Change-Id: I3c14195fb8e094ae150002d90c36a0e4a0cc97d0
* feat(install): enhance binary URL resolution with environment variable support
* fix(install): defer mirror resolution into install() to surface friendly errors
resolveMirrorUrl was called at module scope, so an invalid
LARK_CLI_DOWNLOAD_HOST (e.g. file://) threw before the try/catch in the
postinstall entrypoint, dumping a raw stack trace instead of the recovery
guidance with proxy/registry/host-override options.
Move resolution into install() via getMirrorUrl() so the throw is caught
and the user sees the actionable help text.
* fix(install): keep npmmirror fallback when npm_config_registry is set
resolveMirrorUrl returned a single URL, so any non-default
npm_config_registry replaced the npmmirror fallback entirely. Corporate
npm proxies (Verdaccio, Artifactory, Nexus) often only serve npm package
metadata and don't host /-/binary/<pkg>/..., turning previously-working
installs into 404s when GitHub is unreachable.
Switch to resolveMirrorUrls returning an ordered chain:
- LARK_CLI_DOWNLOAD_HOST set → [override] only (explicit user choice;
no silent leak to npmmirror).
- Otherwise → [derived_from_registry?, npmmirror_default]; npmmirror
is always the final entry, restoring the pre-PR safety net.
install() now walks [GITHUB_URL, ...mirrorUrls] and stops at the first
success.
* fix(install): skip GitHub when LARK_CLI_DOWNLOAD_HOST is set
The download loop unconditionally tried GITHUB_URL first, even when the
user explicitly named a download host. In locked-down networks, probing
github.com can trigger DLP / firewall alerts and contradicts the
explicit-override semantics ("use only this host, nothing else").
When LARK_CLI_DOWNLOAD_HOST is set, the chain is now just [override].
When it isn't, behavior is unchanged: [GITHUB_URL, derived?, npmmirror].
* refactor(install): drop LARK_CLI_DOWNLOAD_HOST env override
Issue #640 only asked for --registry to influence the binary download.
The LARK_CLI_DOWNLOAD_HOST escape hatch was added speculatively for
locked-down networks but is YAGNI — users in those environments already
have npm-level mirrors (--registry) or proxy controls (https_proxy).
Removing it shrinks the surface area:
- delete parseDownloadBase() and its strict https-only validation
- drop the install() branch that skipped GitHub on explicit override
- simplify failure-help message to two recovery options
Resolution chain becomes [GITHUB, derived_from_npm_config_registry?,
npmmirror_default]. The npmmirror tail still preserves the pre-PR safety
net when a corp registry doesn't actually serve /-/binary/<pkg>/...
End-to-end verified on Linux + Windows via real `npm install -g <tgz>`:
all four user scenarios pass, with the issue #640 path (--registry=
npmmirror + GitHub blocked) finishing in 2s on Linux / 6s on Windows.
* feat(ics): add RFC 5545 iCalendar generator and parser
Add shortcuts/mail/ics package:
- builder.go: generates METHOD:REQUEST ICS with VEVENT, ORGANIZER,
ATTENDEE, DTSTART/DTEND with timezone, UID, and X-LARK-MAIL-DRAFT
- parser.go: parses ICS into ParsedEvent struct, detects IsLarkDraft
- Handles CN quoting, control-char sanitization, email validation,
line folding per RFC 5545, and TZID edge cases
Change-Id: I01d13285a57a5a4de50891c54d655efa8423c3c1
* feat(mail): support calendar events in emails
- Add --event-summary/start/end/location flags to +send, +reply,
+reply-all, +forward, +draft-create
- Build ICS and embed as text/calendar in multipart/alternative
- Validate event time range and enforce --event/--send-time mutual
exclusion (extracted into validateEventSendTimeExclusion)
- CalendarBody() in emlbuilder places ICS correctly
- Exclude BCC from ATTENDEE list
Change-Id: Icf9e49ababebc4e8fcf36760ab613c64938c2744
* feat(mail): X-LARK-MAIL-DRAFT and read-only calendar guard
- ics.Build() writes X-LARK-MAIL-DRAFT:TRUE so Feishu client
recognizes CLI-created calendar events as editable
- ics.ParseEvent() detects IsLarkDraft field
- +draft-edit rejects --set-event-* on calendars without
X-LARK-MAIL-DRAFT marker (read-only after send)
- Export FindPartByMediaType from draft package for cross-package use
- Add set_calendar/remove_calendar patch ops with full test coverage
Change-Id: I7d547a4b40880e8d4ee3fecf68864d6ea89e66cd
* feat(mail): forward preserves original calendar ICS
When forwarding an email that contains a calendar event (body_calendar),
pass through the original ICS bytes as text/calendar part if no new
--event-* flags are specified.
Change-Id: I67d2e82604eaf969cee8c7e0bedcf32198d12d57
* docs(mail): document calendar invitation feature
- Add --event-* params to +send, +reply, +reply-all, +forward,
+draft-create, +draft-edit reference docs
- Add calendar_event output section to +message reference
- Add calendar invitation workflow to skill-template/domains/mail.md
- Regenerate SKILL.md via gen-skills
Change-Id: Iccacd06990d91e1cf3beb896d5b772d27e5e29ff
* fix(mail): reject --set-event-start/end/location without --set-event-summary
Change-Id: Icb651ff28ede526ff96b22e7b304b7bdea86d01f
Co-Authored-By: AI
* fix(mail): include --event-location in validateEventFlags; fix stale comment
Change-Id: I2f47016b6bfa11957dfe2c8c499cf36737efba53
Co-Authored-By: AI
* fix(mail): clear stale headers when wrapping single-leaf body in multipart/alternative
Change-Id: I29fe883c9151570f7939d372523b128cbea0b1ed
Co-Authored-By: AI
* fix(mail): add method=REQUEST to text/calendar MIME part created by set_calendar
Change-Id: I4d23674e20e4c42adab36385ff5ee8bb6d97625d
Co-Authored-By: AI
* fix(mail): use post-edit recipients for ICS attendees when --set-to combined with --set-event-*
Change-Id: I659e06635dd043f798d2f2e90d7dbca6e13d7f3d
Co-Authored-By: AI
* fix(mail): cover add_recipient/remove_recipient in ICS attendee resolution
Extract effectiveRecipients() to replay all three recipient op types
(set_recipients, add_recipient, remove_recipient) before building the
ICS for set_calendar, so patch-file recipient changes are reflected in
ATTENDEE metadata.
Change-Id: I3a7a55f96df8fac7d924a4dbeedd5b3d0d9d443c
Co-Authored-By: AI
* fix(mail): derive method= from ICS body in writeCalendarPart instead of hardcoding REQUEST
Passthrough ICS (e.g. forwarded METHOD:CANCEL) previously emitted a
Content-Type with method=REQUEST, disagreeing with the body. Now
extractICSMethod() scans the ICS for METHOD: and falls back to REQUEST
when absent, keeping existing behavior for our own generated ICS.
Change-Id: I4bf6c3755a189a436c2d26b082372d9f838f4051
Co-Authored-By: AI
* fix(mail): normalize calendar_event start/end to UTC in output
Callers expect RFC 3339 UTC strings; source ICS with TZID offsets
previously emitted +08:00 instead of Z.
Change-Id: I88bd4b925f8fc3b4f569e41712ae58ab50d94a2f
Co-Authored-By: AI
* fix(mail): make ICS parser case-insensitive and handle parameterized property names
RFC 5545 §3.1 allows any case and optional parameters on all property
names. Unify UID/SUMMARY/LOCATION/DTSTART/etc. to compare via
strings.ToUpper(name) and add HasPrefix checks for the NAME; form,
consistent with how ORGANIZER and ATTENDEE were already handled.
Change-Id: I7dc642dd210a3256f2189a901a2d9518ea284815
Co-Authored-By: AI
* docs(base): align base skills and view config contracts
1. Rework the lark-base source-of-truth docs around canonical field, cell, record and view payload shapes.
2. Refresh view, workflow, lookup and related references against current openapi behavior and remove stale or broken guidance.
3. Remove dead array-wrapper handling from view sort/group setters and add unit plus dry-run e2e coverage for object-only input.
* docs(base): drop view config code changes from doc refactor
1. Revert the temporary Base view config Go and test adjustments so this PR only keeps lark-base skill and reference updates.
2. Preserve the documentation contract changes while leaving runtime behavior unchanged from the pre-refactor implementation.
* docs(base): revert temporary view config code cleanup
1. Restore the pre-refactor Base view config Go paths and related unit tests so this PR keeps runtime behavior unchanged.
2. Leave the lark-base skill and reference updates in place as the only intended product change in this branch.
* docs(base): fix progress color typo
* docs(base): trim padding in reference docs
1. Remove obviously excessive alignment spaces from base reference examples and operator lists.
2. Shorten a few overlong separator rows in the formula guide to reduce low-value formatting noise.
3. Keep the changes scoped to four lark-base reference files without changing documented behavior.
* docs(base): clarify field description guidance
* test: isolate dry-run e2e config state
* chore: update data-query prompt
* docs(base): simplify formula filter guidance
* docs(base): drop stage field mention from data query
* revert: keep e2e changes scoped to base docs
* docs(base): clarify dashboard field type wording
* docs(base): trim number filter operators
`im chats link` is registered as a regular service method (no
`risk: high-risk-write` annotation), so the framework does not register
the `--yes` flag on it. Setting `Yes: true` on the e2e Request makes
the runner append `--yes`, which cobra rejects with `unknown flag:
--yes` before the request is ever issued — the rest of the assertions
then fall through with empty stdout.
The flag was added in #633 alongside the risk-tiering rollout that
covered other workflows that genuinely flipped to high-risk-write.
For chats link the API call (creating a chat share link with a
configurable validity period) is not destructive and was never
re-classified, so the line is just leftover from that pass. Drop it
to restore the e2e green; if we ever decide to gate share-link
creation behind confirmation we can re-add it together with the
metadata flip.
Change-Id: Ieb094407a7f0fa18cd130a9d80c7146274b5ecc7
* feat(contact +search-user): add search filters and richer profile fields
- Filter results by chat history, employment status, tenant boundary,
or enterprise email presence; keyword is now optional so filter-only
queries ("list all my external contacts") work end-to-end.
- Each result now carries multilingual names, contact email, activation
state, whether you've chatted with them, tenant context, user
signature, and a hit-highlight line that surfaces the matched segment
and the user's department path.
- Always-empty legacy columns and fields the new backend no longer
returns are dropped.
- Also fixes the contact +get-user skill doc, which previously
instructed callers to pass --table (a flag that never existed); now
correctly documents --format table and the full --format enum.
* refactor(lark-contact): clean up search-user code, tighten skill docs
- contact_search_user.go / _test.go: simplify and clarify
- SKILL.md: focus description on user-facing trigger scenarios;
rework decision table; trim notes to load-bearing constraints
- references/lark-contact-search-user.md: add flag table covering
all four bool filters; add multi-filter examples; clean up
output field contract (drop server <h> tag implementation detail)
- references/lark-contact-get-user.md: trim to two real use cases
(self via user identity; full profile of others via bot identity);
point user-mode-by-id users to +search-user instead
- .golangci.yml: replace package-level deny on net/http with a
symbol-level forbidigo rule. Constants (http.MethodPost,
http.StatusOK) and helpers (http.StatusText) were never the
intent; only Client / NewRequest / Get / Post / Do etc. are now
blocked in shortcuts/, matching the rule's actual purpose
Change-Id: Ic42043d3f4c1b675800e48229c7ba2e970da26fe
* fix(contact +search-user): align query limit and reject empty user-ids
API rejects queries longer than 50 characters; local cap was 64 runes,
producing confusing "passed local validation but server-rejected"
behaviour. Lower the cap to 50 and rename the constant accordingly.
Also reject --user-ids inputs that parse to zero entries (",,,",
" , , ", ","): SplitCSV silently dropped empty segments, so the
shortcut sent an empty body to the API and returned indeterminate
results.
Change-Id: Ib34fe897023e175bf4c657273bdb49a33d2f083b
---------
Co-authored-by: liangshuo-1 <266696938+liangshuo-1@users.noreply.github.com>
Add BuildResourceURL helper and wire it into doc/sheets/drive/base/wiki
create paths so callers always receive a clickable link, even when the
backend response (MCP degraded path or upstream OpenAPI) returns an
empty URL field. The fallback uses the brand-standard host
(www.feishu.cn / www.larksuite.com), which redirects to the tenant
domain.
Affected entries:
- docs +create v1 / v2
- sheets +create
- drive +create-folder / +import / +upload (newly exposes url)
- wiki +node-create (newly exposes url)
drive +create-shortcut is intentionally skipped because the URL form
depends on the underlying file kind, which the shortcut payload does
not carry.
* feat(risk): implement confirmation for high-risk write operations
* feat(risk): streamline confirmation for high-risk write operations
* feat(risk): document approval protocol for high-risk write operations
* feat(risk): refine confirmation protocol for high-risk write operations
* feat(risk): remove redundant variable declaration in risk test
* feat(risk): add 'Yes' flag to various test cases for confirmation
The previous default (atomic.Bool zero-value = enabled) meant any
*cobra.Command built without first calling configureFlagCompletions
leaked into cobra's package-global flagCompletionFunctions map. Bench
runs (scripts/bench_build) showed hundreds of KB and thousands of
objects retained per Build call.
Flip the semantics so the zero-value matches the safe default:
- Rename internal var to flagCompletionsEnabled (zero = disabled).
- Rename public API to SetFlagCompletionsEnabled / FlagCompletionsEnabled.
- Update call sites in cmd/root.go and scripts/bench_build/main.go.
- Add cmd.TestBuild_DefaultNoCompletionLeak: asserts that, with no
setter call at all, repeated cmd.Build invocations stay under 50 KB
and 500 objects per build (observed: ~0.7 KB, 3 objs/build). This
closes the gap that let the wrong default ship — every previous
test explicitly Set the switch before exercising it.
Change-Id: Ifefb04af5fd45eea9676a344a64ad071b6a4cd1a
* feat(event): add event subscription & consume system with orphan bus detection
Introduces end-to-end Feishu event consumption via a new `lark-cli event`
command family. Users can subscribe to and consume real-time events
(IM messages, chat/member lifecycle, reactions, ...) in a forked bus
daemon architecture with orphan detection, reflected + overrideable JSON
schemas, and AI-friendly `--json` / `--jq` output.
Commands
--------
- `event list [--json]` list subscribable EventKeys
- `event schema <key>` Parameters + Output Schema + auth info
- `event consume <key>` foreground blocking consume; SIGINT/SIGTERM
/stdin-EOF shutdown; `--max-events` /
`--timeout` bounded; `--jq` projection;
`--output-dir` spool; `--param` KV inputs
- `event status [--fail-on-orphan] [--json]` bus daemon health
- `event stop [--all] [--force] [--json]` stop bus daemon(s)
- `event _bus` (hidden) forked daemon entrypoint
Architecture
------------
- Bus daemon (internal/event/bus): per-AppID forked process that holds
the Feishu long-poll connection and fans events out to 1..N local
consumers over an IPC socket. Drop-oldest backpressure, TOCTOU-safe
cleanup via AcquireCleanupLock, idle-timeout self-shutdown, graceful
SIGTERM.
- Consume client (internal/event/consume): fork+dial the daemon,
handshake, remote preflight (HTTP /open-apis/event/v1/connection),
JQ projection, sequence-gap detection, health probe. Bounded
execution (`--max-events` / `--timeout`) for AI/script usage.
- Wire protocol (internal/event/protocol): newline-delimited JSON
frames with 1 MB size cap and 5 s write deadlines. Hello / HelloAck /
PreShutdownCheck / Shutdown / StatusQuery control messages.
- Orphan detection (internal/event/busdiscover): OS process-table scan
(ps on Unix, PowerShell on Windows) with two-gate cmdline filter
(lark-cli + event _bus) that naturally rejects pid-reused unrelated
processes.
- Transport (internal/event/transport): Unix socket on darwin/linux,
Windows named pipe on windows.
- Schema system (internal/event, internal/event/schemas): SchemaDef with
mutually-exclusive Native (framework wraps V2 envelope) or Custom
(zero-touch) specs. Reflection reads `desc` / `enum` / `kind` struct
tags, with array elements diving into `items`. FieldOverrides overlay
engine addresses paths via JSON Pointer (including `/*` array
wildcard) and runs post-reflect, post-envelope. Lint guards orphan
override paths.
- IM events (events/im): 11 keys — receive / read / recalled, chat and
member lifecycle, reactions — all with per-field open_id / union_id /
user_id / chat_id / message_id / timestamp_ms format annotations.
Robustness
----------
- Bus idle-timer race fix: re-check live conn count under lock before
honoring the tick; Stop+drain before Reset per timer contract.
- Protocol frame cap: replace `br.ReadBytes('\n')` with `ReadFrame` that
rejects frames > MaxFrameBytes (1 MB). Closes a DoS path where any
local peer could grow the reader's buffer unbounded.
- Control-message writes gated by WriteTimeout (5 s) so a wedged peer
kernel buffer can't stall writers indefinitely.
- Consume signal goroutine: `signal.Stop` + `ctx.Done` select, no leak
across repeated invocations in the same process.
- JQ pre-flight compile so bad expressions fail before the bus fork and
any server-side PreConsume side effects.
- `f.NewAPIClient`'s `*core.ConfigError` now passes through unwrapped
so the actionable "run lark-cli config init" hint reaches the user.
Subprocess / AI contract
------------------------
- `event consume` emits `[event] ready event_key=<key>` on stderr once
the bus handshake completes and events will flow. Parent processes
block-read stderr until this line before reading stdout — no `sleep`
fallback needed.
- All list-like commands have `--json` for structured consumption.
- Skill docs in `skills/lark-event/` (SKILL.md + references/) brief AI
agents on the command surface, JQ against Output Schema, bounded
execution, and subprocess lifecycle.
Testing
-------
Unit tests across bus/hub, consume loop, protocol codec, dedup,
registry, transport (Unix + Windows), schema reflection, field
overrides, pointer resolver. Integration tests cover fork startup,
shutdown, orphan detection, probe, stdin EOF, preflight, bounded
execution, and Windows busdiscover PowerShell compatibility.
Change-Id: Ib69d6d8409b33b99790081e273d4b5b01b7dbf80
* fix(event): address CodeRabbit findings + lift patch coverage above 60%
CodeRabbit comments (PR #654)
-----------------------------
1. bus/dedup: IsDuplicate dropped legitimate (post-TTL) events after
cleanupExpired fired. The run-every-1000-inserts cleanup removed
TTL-expired IDs from the `seen` map but left them in the ring;
IsDuplicate's ring-scan fallback then rediscovered them and falsely
reported "duplicate", and bus.Publish silently dropped the event.
Removed the ring-scan branch — `seen` is the sole authority, the ring
only bounds map size via overflow eviction. New regression test
TestDedupFilter_TTLExpiryAfterCleanupRunRespected exercises the 10-
insert + cleanup path and guards the fix.
2. consume/remote_preflight: the decoder only read `data.online_instance_
cnt`. A non-zero business code with no data payload decoded to 0 and
callers treated it as "verified zero", forking a local bus that would
duplicate events. Added Code / Msg fields and promoted code != 0 into
an error so the caller distinguishes verified-zero from check-failed.
3. cmd/event/stop: swapped os.ReadDir / os.Stat to vfs.ReadDir / vfs.Stat
in discoverAppIDs per project guideline (enables test mocking). New
TestDiscoverAppIDs_* lifts discoverAppIDs from 0% to 100%.
4. cmd/event/appmeta_err: narrowed authURLPattern from
feishu.cn|feishu.net|larksuite.com|larkoffice.com to the two hosts
consoleScopeGrantURL actually produces. Kept the allowlist pinned to
ResolveEndpoints' output with a comment flagging the synchrony.
5. cmd/event/list: moved "No EventKeys registered." and "Use 'event
schema <key>' for details." hints to stderr so `event list | jq`
style pipelines don't ingest them as data.
6. cmd/event/schema: runSchema is a RunE entry point; swapped the bare
fmt.Errorf on resolveSchemaJSON failure to output.Errorf so AI
agents parse a structured error envelope.
Coverage bumps (patch ~50% -> ~60%)
-----------------------------------
internal/event/consume/loop_test.go: loop.go was 0% at patch time.
New tests cover consumeLoop end-to-end via net.Pipe (events -> sink,
max-events -> ctx.Done -> PreShutdownCheck/Ack), seq-gap warning,
jq filtering + early compile failure, isTerminalSinkError classifier.
Takes consumeLoop from 0% to ~74%.
internal/event/protocol/messages_test.go: all NewXxx constructors,
Encode/Decode roundtrip per message type, EncodeWithDeadline deadline
enforcement, ReadFrame MaxFrameBytes rejection + EOF propagation.
Takes protocol from 28% to ~86%.
Also bundles small UX polish:
- cmd/event/consume: --output-dir flag doc flags path-traversal behavior;
jq-validation failures now re-wrap with an event-specific hint
pointing at `event schema` for payload shape.
- internal/event/consume.validateParams: error now names the EventKey
and lists valid param names inline so AI callers recover without a
second `event schema` round-trip.
- skills/lark-event: description expanded to mention
listener/subscribe/consume synonyms + the IM scope set explicitly;
lark-event-im reference polished; obsolete lark-event-subscribe
reference removed.
Verified with go test -race -timeout 120s across ./cmd/event/...,
./events/..., ./internal/event/...; gofmt clean; go vet clean.
Change-Id: I3837b8645ea1d7529c9a8fd4c2bbfa965ae1b519
* test(event): cover format helpers + cobra factories
Adds cmd/event/format_helpers_test.go covering the pure output helpers
and factory wire-ups that RunE-level tests would need a live bus to
exercise:
- writeStopJSON: shape assertions + nil → [] (scripts expecting
.results | length must not see null).
- writeStopText: stdout vs stderr routing — stopped / no-bus lines to
stdout, refused / errored lines to stderr.
- busState.String: all three discriminator values.
- humanizeDuration: each bucket boundary (seconds / minutes / hours / days).
- writeStatusText: covers stateNotRunning / stateRunning (with consumer
table) / stateOrphan (with kill hint).
- writeStatusJSON: orphan entry carries suggested_action + issue;
running entry must NOT carry those fields (hint-leak guard for
scripts that key on issue != "").
- exitForOrphan: flag-off never errors; flag-on errors iff any orphan
is present, with ExitValidation code.
- NewCmdConsume / NewCmdStatus / NewCmdStop / NewCmdList / NewCmdBus:
flag registration + RunE presence, so review catches flag-name drift.
NewCmdBus check also pins Hidden=true.
Lifts cmd/event coverage 51.7% → 61.1%; aggregate event-package
coverage crosses the 60% codecov patch threshold (62% locally).
Change-Id: I9ecf3d905a8f9607b9441ee8a61e746496e2be63
* fix(event): address lint + deadcode CI failures
4 golangci-lint findings + 1 deadcode finding flagged on PR #654.
lint
----
1. cmd/event/stop.go:86 (ineffassign): `targets := []string{}` is
overwritten by both branches of the `if o.all` below, so the empty-
slice initializer is dead. Switched to `var targets []string`.
2. cmd/event/consume.go nilerr: the user-identity scope preflight
swallows a non-nil ResolveToken error and returns nil. This is
intentional — a missing/expired user token must not block consume;
the bus handshake will surface the real auth error with actionable
hints. Added `//nolint:nilerr` with a 4-line comment pinning the
reasoning.
3. events/im/message_receive.go:62 nilerr: malformed JSON payload
returns the original bytes + nil so consumers still see the event
(the WARN breadcrumb lives in the outer loop). Added
`//nolint:nilerr` with a one-line comment.
4. internal/event/schemas/fromtype_test.go:26 unused: `unexportedStr`
is a reflection-test fixture — its presence (not value) exercises
the FromType skip-unexported path verified at the "unexported
field should not be in schema" assertion. Added `//nolint:unused`
and a 4-line comment pointing at the guarded assertion.
deadcode
--------
5. internal/event/testutil/testutil.go: NewTCPFake has no callers in
the repo. Removed the constructor plus the `inner == nil` TCP-mode
branches from Listen / Dial / Cleanup. FakeTransport now only
supports the wrapped-overlay mode (NewWrappedFake), which is the
one every existing test uses. Doc comment simplified accordingly.
Verified locally: go test -race -timeout 120s across ./cmd/event/...,
./events/..., ./internal/event/... all green; gofmt clean; go vet
clean.
Change-Id: Ie8a2270827a0bde6b8159ab70aaf5c1e9ca7d5b9
* fix(event): drop stale enum + simplify protocol test type helper
- events/im/message_receive.go: dropped the `enum` tag on
ImMessageReceiveOutput.MessageType. convertlib registers many more
message types than the old 11-item list (video / location /
calendar / todo / vote / hongbao / merge_forward / folder / ...),
so a partial enum would tell AI consumers that valid values like
"video" are invalid and produce false-negative JQ filters.
- internal/event/protocol/messages_test.go: collapsed the
typeOf → reflectTypeName → stringType chain in
TestEncode_DecodeRoundtripAllTypes to a single fmt.Sprintf("%T", v).
The hand-maintained type switch silently returned "<unknown>" for
any new message type, which would have let future Decode bugs slip
past the roundtrip assertion. Also removed a dead `cases` table at
the top of TestConstructors_PinTypeField left over from an earlier
refactor.
Change-Id: I831e96f8417e80637596030d652a559de0d33122
* docs(event): polish skill docs + rename root_path_hint to jq_root_path
- skills/lark-event/SKILL.md, lark-event-im.md: translated to English,
reorganized around a top-level "Core commands" table, scenario
recipes tightened.
- cmd/event/schema.go: renamed the writeSchemaJSON hint field
RootPathHint / "root_path_hint" -> JQRootPath / "jq_root_path" to
make its purpose (a jq path prefix) obvious at the call site; no
external consumer depends on the old name yet.
Change-Id: I00c14061ca33caedc0975bfeadc4b26d3dcd314d
* chore(event): strip excessive comments
Change-Id: I8f44f36f5dbdba3ef95dfc67069dc796232f91ec
* fix(event): dedup self-eviction race + protocol oversized-frame test
dedup: in IsDuplicate, the ring-slot eviction step deleted seen[id] even
when ring[pos] equalled the freshly-recorded id (post-TTL reinsertion
landing on its own historical slot). Net result: ring still held id but
seen did not, so the next IsDuplicate(id) returned false and the
duplicate was delivered. Skip the delete when old == eventID. New
TestDedupFilter_SelfEvictionPreservesFreshEntry pins the invariant by
pre-loading the ring slot and asserting the second call still reports
duplicate.
protocol: TestReadFrame_RejectsOversized used strings.Contains feeding
t.Logf, so any non-nil error passed — including a future regression
that returned io.ErrUnexpectedEOF while silently keeping the buffer
unbounded. Promoted MaxFrameBytes overflow to a sentinel
ErrFrameTooLarge and the test now asserts via errors.Is.
Change-Id: I50281dad392152b0ca083fd30c38eb0695e63bd3
* docs(event): clarify .content shape per message_type + add sender filter recipe
Change-Id: I619fd15c1a362e42e6602fd3e3316bbc75eddc5e
* fix(event): replace cmdline-regex bus discovery with PID file + close concurrent fork race
Bus discovery previously walked the OS process table and parsed `--profile cli_*` from
cmdline; the regex rejected any non-cli_ profile name (D-03a). Replace with per-AppID
bus.pid + bus.alive.lock under events/<AppID>/, probed via try-lock. AppID round-trips
through the directory name, so the profile-vs-AppID confusion is gone by construction.
Also fix B-07 (two consumers each fork an independent bus, halving event delivery):
- forkBus holds bus.fork.lock until child is dial-able, not just until cmd.Start
- bus daemon takes alive.lock before binding the socket; cleanup-TOCTOU race can no
longer leave two listeners on different inodes
status.go renders an orphan with PID=0 distinctly (live bus but pid file unreadable)
so we never print "Action: kill 0".
Change-Id: I3bf0a6cf1d91fb274ac5a6df83d66896aafb291f
* style(event): gofmt bus.go
Trailing blank line introduced when appending acquireAliveLock helper.
Change-Id: I4ae1b4a4363dc6c89dcbd6a170f4563117490ba3
* fix(event): swap os.Remove/Rename for vfs.* and silence forbidigo on internal diagnostics
golangci-lint forbidigo blocks os.* in internal/. Switch the pid-file write to vfs.Remove/vfs.Rename and add a nolint marker on the two stderr diagnostics in busdiscover, matching the existing pattern in consume/*.
Change-Id: Ia6768be62aefeb8ca40f991d3130a78ef2ec0ea5
* fix(event): cross-platform --all + clean SIGPIPE shutdown for consume
- stop --all: replace bus.sock-file probe with busdiscover lock-based
scan; previously skipped Windows entirely (named-pipe transport, no
socket on disk) and misidentified Unix stale sockets as live. Same
win for `event status` (shares discoverAppIDs).
- consume: ignore SIGPIPE so a closed stdout pipe (e.g. `... | head -n 1`)
surfaces as EPIPE error and reaches the existing isTerminalSinkError
cleanup path (log "output pipe closed", lastForKey query, hub
unregister), instead of being killed by Go's default fd 1/2 SIGPIPE
handler with exit 141 and zero deferred cleanup.
Build-tagged: real on unix, no-op on windows (no SIGPIPE there).
Change-Id: I453b19f05c489fd9d5c1a9ba3bdc35e127c15b83
* docs(event): translate IM EventKey descriptions and field tags to English
Aligns with the rest of the codebase (titles, struct names, README) which
are already in English. Surfaces in `event list` / `event schema` and is
also consumed by AI agents.
- events/im/message_receive.go: 11 desc tags on ImMessageReceiveOutput
- events/im/native.go: 10 description fields on Native EventKeys
- events/im/register.go: im.message.receive_v1 Description
Change-Id: I6f46950b4793f137e0129c1f06019a3419195443
* docs(event): drop misleading AuthTypes[0] auto-default claim
The KeyDefinition comment and SKILL.md flag table both stated that
`--as auto` resolves to `AuthTypes[0]`. It does not — ResolveAs goes
through global rules (config default_as / credential hint / `bot`
fallback) without consulting the EventKey. AuthTypes is only used by
CheckIdentity as a post-resolve whitelist.
Reword the field comment to plain whitelist semantics and have SKILL.md
defer `--as` documentation to lark-shared.
Change-Id: Ia5d3d3790aed05813a0fa72d6b43518224e2055b
* revert(comments): restore original comments on 3rd-party files
e61482a stripped comments across 105 files. Restore the four files
authored by others (cmd/build.go, shortcuts/common/{types,runner}.go,
shortcuts/event/subscribe.go) to their pre-strip state so unrelated
documentation isn't churned in this PR.
Change-Id: Ie2527b06bfaf5b3861b0b9dff1e19bbfe7dde456
* fix(e2e/wiki): pass obj_type when deleting wiki nodes in cleanup
The wiki node DELETE endpoint now rejects requests without obj_type
(API error 99992402: "obj_type is required"), causing TestWiki_NodeWorkflow
cleanup to fail on every run. Forward the obj_type from the create/copy
response into the delete query params so cleanup succeeds.
* fix(e2e/wiki): delete cleanup wiki nodes via drive v1 endpoint
The wiki v2 DELETE /spaces/{space_id}/nodes/{node_token} endpoint is
undocumented and rejects requests with `obj_type is required` even when
obj_type is forwarded as a query parameter (see actions run #25005966144).
Switch cleanup to the documented path: delete the underlying drive file
via DELETE /drive/v1/files/{obj_token}?type=<obj_type>, which removes the
backing document and the wiki node in one call.
Change-Id: Ieb93b1f92ea758d8b80bcfdd4f20b2be8f35a0bd
* fix(e2e/wiki): pass obj_type to wiki delete in body, not query
Previous attempts:
- query (?obj_type=docx) → API still rejects with 99992402 obj_type
required (the wiki delete-node endpoint reads it from the body, not
the query string).
- drive v1 fallback → bot identity does not have drive write scope and
returns 1061004 forbidden, so we cannot reuse drive's delete API for
the cleanup helpers.
Add example commands for file types declared in the supported-conversions
table but absent from the command examples section: .docx/.doc, .txt,
.html, .xls -> sheet, and .csv -> sheet.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(strict-mode): reject explicit --as instead of silently overriding it
ResolveAs checked strict mode before the --as flag, so `--as bot` under strict=user
was silently rewritten to user. Reorder so explicit --as is returned as-is and CheckStrictMode rejects the conflict (exit=2). Implicit paths (--as auto / unset) are still forced by
strict mode.
* fix(strict-mode): fix CI
Project management for Lark/Feishu is provided by the standalone
meegle-cli (https://github.com/larksuite/meegle-cli), which requires a
separate install. Surface it in the Features table so users can
discover the capability without expecting it to ship inside lark-cli.
Add --at-chatter-ids flag to shortcuts/im/im_messages_search.go that
passes filter.at_chatter_ids to the search API, restricting results to
messages that @mention any of the given user open_ids. Messages that
* feat(pagination): preserve pagination state on truncation and natural end
* feat(pagination): drop page_token from merged output to reflect aggregate view
Expose doc_wiki/search v2 under the drive domain via explicit flags
(--query, --edited-since, --commented-since, --opened-since,
--created-since, --mine, --creator-ids, --doc-types, --folder-tokens,
--space-ids, ...) instead of a nested JSON filter, so natural-language
queries from AI agents map 1:1 to discrete flags.
Time handling:
- my_edit_time and my_comment_time are snapped to the hour (floor/ceil)
with a stderr notice, since those fields are aggregated at hour
granularity server-side. create_time passes through as-is.
- open_time has a server-side 3-month cap per request. When
--opened-since / --opened-until span exceeds 90 days, the CLI narrows
the request to the most recent 90-day slice and emits a stderr notice
listing every remaining slice's --opened-* values so the agent can
re-invoke for older ranges. Spans over 365 days are rejected up front
to bound runaway slicing.
Flag ergonomics:
- --doc-types accepts mixed case; values are normalized to upper case
before validation and before being sent to the server.
- --sort default is translated to the server enum DEFAULT_TYPE (every
other sort value upper-cases 1:1).
Error hints:
- Lark code 99992351 (referenced open_id outside the app's contact
visibility) is enriched with a +search-specific hint that
distinguishes API scope from contact visibility and points at
--creator-ids / --sharer-ids as the likely source.
Skill docs:
- new reference at skills/lark-drive/references/lark-drive-search.md,
including the open_time slicing protocol and the paginate-within-
slice-before-switching agent playbook.
- lark-drive/SKILL.md routes resource-discovery to drive +search.
- lark-doc/SKILL.md and lark-doc-search.md mark docs +search as
deprecated and point users at drive +search.
Change-Id: I36d620045809b448446d4fdbdfa923b05794da19
* feat(mail): add +share-to-chat shortcut to share emails as IM cards
Two-step API (create share token → send card) wrapped in a single
shortcut. Supports message-id/thread-id, five receive-id-type variants
(chat_id, open_id, user_id, union_id, email), and dry-run mode.
Change-Id: Ic7b8c01c0d25fef262f35be92555f1fd019bd679
Co-Authored-By: AI
* fix(mail): regenerate SKILL.md from skill-template instead of manual edit
Add missing safety rule 8 (draft link rule) to skill-template/domains/mail.md
so it survives regeneration. SKILL.md is now produced by `make gen-skills`
in the registry repo rather than hand-edited.
Change-Id: I9cf3605deae8b6de2042e40819fedc304967e78e
Co-Authored-By: AI
* fix(mail): add docstrings and use real validation path in tests
- Add Go doc comments to exported symbols for docstring coverage
- Rewrite tests to exercise MailShareToChat.Validate via RuntimeContext
instead of duplicating validation logic
- Replace hand-rolled containsStr with strings.Contains
- Add httpmock stubs for execute and error path tests
Change-Id: Ic781494f61e9e844224185844bce7b0c48e8e200
Co-Authored-By: AI
* test(mail): add dry-run E2E test for +share-to-chat
Validate request shape (method, URL, mailbox path) under --dry-run
with fake credentials. Covers message-id, thread-id, and custom
mailbox variants.
Change-Id: Iae87bf141cbe4f312d3e9b1fca4ba175052c5c35
Co-Authored-By: AI
* fix(mail): include request body and params in dry-run output
DryRun now mirrors Execute: the share-token POST shows message_id or
thread_id, and the send POST shows receive_id_type and receive_id.
E2E test updated to assert these fields. Also fix strconv.Itoa usage.
Change-Id: I00f8770fd5a12b7354986c5e5077f97cfe5d6653
* style(mail): gofmt dry-run test file
Change-Id: I47dc6a9a47252dcfb7853737f88dfdaef65a0ae7
* test(mail): assert exact API call count in dry-run test
Change-Id: I9f4a1a183b55d03f5248eb4adddfddb08037ca95
End-to-end RFC 3798 Message Disposition Notification support, covering
both sides of the receipt flow — requesting a receipt when composing, and
responding to one (send or decline) when reading.
Request side (compose)
- New --request-receipt flag on +send / +reply / +reply-all / +forward /
+draft-create / +draft-edit. When set, the outgoing EML carries a
Disposition-Notification-To header (RFC 3798) addressed to the resolved
sender. Recipient mail clients may prompt the user, auto-send a receipt,
or silently ignore — delivery is not guaranteed.
- requireSenderForRequestReceipt gates the flag against a controlled
sender address resolved BEFORE the orig.headTo fallback in +reply /
+reply-all / +forward, so the DNT cannot silently land on someone else
in CC / shared-mailbox flows.
Response side
- +send-receipt: build a system-templated reply for messages carrying the
READ_RECEIPT_REQUEST label (-607). Subject / recipient / sent / read
time layout matches the Lark client; body is non-customizable — receipt
bodies are system templates by industry convention; free-form notes
belong in +reply. Risk:"high-risk-write" + --yes required.
- +decline-receipt: clear READ_RECEIPT_REQUEST without sending anything
(mirrors the client's "不发送" / "Don't send" button). Idempotent on
re-run; Risk:"write" — no --yes needed.
Read-path hints
- +message / +messages / +thread emit a stderr hint when surfacing a
mail carrying READ_RECEIPT_REQUEST, exposing BOTH response paths
(+send-receipt --yes / +decline-receipt) so agents present a real
choice to the user instead of silently auto-sending.
Guard rails
- +send / +reply / +reply-all / +forward stay draft-by-default and
require --confirm-send to send, gated by a dynamic scope check for
mail:user_mailbox.message:send (absent from the default scope set so
draft-only flows don't need the sensitive permission).
- All header-bound user input (sender / display name / recipient /
subject) goes through CR/LF rejection plus Bidi / zero-width / line-
separator guards, mirroring emlbuilder.validateHeaderValue, to block
header injection and visual spoofing.
- Hint output strips terminal control characters (CR, LF) from any
untrusted field embedded into the user-visible suggestion.
Backend coupling
- Outgoing receipt EML carries the private header
X-Lark-Read-Receipt-Mail: 1. The data-access backend parses it into
BodyExtra.IsReadReceiptMail; DraftSend then applies READ_RECEIPT_SENT
(-608) and clears READ_RECEIPT_REQUEST (-607) from the original
message, closing the client-side banner.
- en receipts require backend TCC SubjectPrefixListForAdvancedSearch to
include "Read Receipt:" for conversation-view aggregation; zh prefix
("已读回执:") is already configured.
Docs: new reference pages for +send-receipt / +decline-receipt;
--request-receipt noted on each compose-side reference; SKILL.md
workflow (section 9) describes the full privacy-safe decision tree on
both sides.
Tests cover emlbuilder DispositionNotificationTo / IsReadReceiptMail
helpers, receiptMetaLabels (zh / en), buildReceiptSubject, text and HTML
body generators (with HTML escaping and Bidi guards), header-injection
defenses, sender-resolution gating (CC-only / shared-mailbox regression),
hint emission paths, and the full +send-receipt / +decline-receipt happy
+ idempotent paths via httpmock.
Update im +chat-messages-list to request only thread root messages from /open-apis/im/v1/messages by default. This aligns the shortcut request shape with topic-group usage and makes the intended API behavior explicit in both runtime params and dry-run output.
Change-Id: I3901b27e70b0e4db506ff199eb03c96fcf98671d
Adds an `--api-version v2` path to the docs shortcuts, backed by the
`docs_ai/v1/documents` OpenAPI. DocxXML is the default document format
and Markdown is available as an alternative. Content input is unified
across the three shortcuts via `--content` + `--doc-format`. The v1
(MCP) path is preserved for backward compatibility and now prints a
deprecation notice on use.
Shortcuts:
- `docs +create --api-version v2`: create a document from XML or
Markdown, with optional `--parent-token` or `--parent-position`.
Bot identity continues to auto-grant the current CLI user
full_access on the new document.
- `docs +fetch --api-version v2`: adds `--detail simple|with-ids|full`
for export granularity and `--scope full|outline|range|keyword|section`
for partial reads, along with `--context-before` / `--context-after`,
`--max-depth`, and `--revision-id`.
- `docs +update --api-version v2`: introduces structured operations
via `--command`: `str_replace`, `block_delete`, `block_insert_after`,
`block_copy_insert_after`, `block_replace`, `block_move_after`,
`overwrite`, `append`.
Framework support in `shortcuts/common`:
- `OutRaw` / `OutFormatRaw` emit the JSON envelope with HTML escaping
disabled so XML/HTML document bodies are preserved verbatim.
- New `Shortcut.PostMount` hook runs after a cobra.Command is fully
configured; used here to install a version-aware help function
that hides flags belonging to the inactive `--api-version`.
Also refreshes the lark-doc skill pack (SKILL.md, create/fetch/update
references, new lark-doc-xml and lark-doc-md references, style and
workflow guides), README examples, and downstream skill call sites
(lark-drive, lark-vc, lark-whiteboard, lark-workflow-meeting-summary,
lark-event).
Change-Id: Ide2d86b190a4e21095ae29096e7fb00031d80489
Give each AI Agent (OpenClaw, Hermes) its own lark-cli workspace so
its Feishu calls don't overwrite the developer's local config or
collide with other Agents.
lark-cli config bind [--source openclaw|hermes] [--app-id <id>]
[--identity bot-only|user-default] [--force]
Key capabilities:
- Source auto-detected from OPENCLAW_* / HERMES_* env signals; config
written to ~/.lark-cli/<agent>/, isolated per Agent.
- Two identity presets: 'bot-only' (flag-mode default) and
'user-default'. Flag mode rejects silent bot→user escalation
without --force; TUI prompts are exempt.
- Agent-friendly stdout JSON with 'identity' + 'message' for
next-step branching.
- 'config show' and 'doctor' expose the bound 'workspace'.
- OpenClaw SecretRef resolution: plain / ${VAR} / file:+JSON Pointer
/ exec:.
* refactor: make install.js side-effect-free on require
Change-Id: I5444e3f34642d7c0740b6422a70ca6921a85e363
* feat: add getExpectedChecksum with unit tests
Change-Id: I87548be25d30c384e743da17b1d161b9d9f0ea87
* feat: add verifyChecksum with unit tests
Change-Id: Ifc2067bf1b824b02257dba7b53716fbe18d0f6b6
* feat: harden download with host allowlist and checksum verification
Change-Id: I2580782866049f1f62a2597e86b7bf59d0e50925
* ci: bundle checksums.txt in npm package for install verification
Change-Id: I2d7c44d9d5b9075158f63c0f8cf66c1e0abe3d8d
* ci: use triggering tag and verify checksums.txt presence in release workflow
Address CodeRabbit review: use GITHUB_REF_NAME instead of parsing
package.json to avoid version drift, and add explicit file check to
fail loudly if checksums.txt is missing or empty.
Change-Id: I8a5658412b6afc338ad2a642baba146cceafd0fc
* feat: streaming hash, allowlist tests, and malformed-line coverage
- verifyChecksum: switch from readFileSync to streaming 64KB chunks
to avoid loading entire archive (10-100MB) into memory
- Export and test assertAllowedHost: 7 cases covering allowed hosts,
rejection, case normalization, port handling, invalid URL
- Add ALLOWED_HOSTS comment clarifying it only gates initial URL
- Add getExpectedChecksum tests for malformed/tab-separated lines
Change-Id: Ida639def89c242b3b261a76effae08fd414a10dc
* refactor(base): enforce field-map record upsert input
1. Reject top-level fields wrappers in base +record-upsert input and keep request bodies as field maps.
2. Replace record-upsert tests with Map<FieldNameOrID, CellValue> input and assert the outgoing body has no fields wrapper.
3. Consolidate Base record value documentation around lark-base-cell-value and update record command references.
* refactor(base): use common record JSON parsing for upsert
1. Remove the dedicated record-upsert parser and restore the shared record JSON object validation path.
2. Keep record-upsert dry-run and execution as raw JSON object passthrough.
3. Drop the test assertion that rejected a top-level fields key for record-upsert.
* docs(base): refine record cell value guidance
1. Align record CellValue examples with live behavior for date, URL, user, link, select, numeric styles, and readonly fields.
2. Remove misleading user_id_type and execution identity prompts from record-writing guidance.
3. Keep record JSON file input guidance generic and avoid documenting environment-specific stdin or path limits.
Introduces `lark-cli slides +replace-slide`, a shortcut over the
native `xml_presentation.slide.replace` API for element-level editing
of existing Lark Slides pages. Callers pass a JSON array of parts and
the CLI handles URL resolution, XML hygiene, client-side validation,
and 3350001 hint enrichment.
Why a dedicated shortcut
The native API has three sharp edges every caller hits:
1. URL formats. Users have /slides/<token> or /wiki/<token> URLs, not
bare xml_presentation_id.
2. Undocumented XML hygiene. `block_replace` requires id=<block_id> on
the replacement root; <shape> requires <content/>. Missing either
returns a catch-all 3350001 with no guidance.
3. 3350001 is a catch-all on the backend with no actionable message.
Code
shortcuts/slides/slides_replace_slide.go (new)
- Flags: --presentation (bare token | /slides/ URL | /wiki/ URL),
--slide-id, --parts (JSON array, max 200), --revision-id (-1 for
current, specific number for optimistic locking), --tid,
--as user|bot.
- Validation (pre-API): [1,200] item cap; action restricted to
block_replace / block_insert (str_replace rejected); per-action
required fields (block_id for block_replace, insertion for
block_insert); per-field string type-assertion guards on the
decoded JSON so a numeric/bool payload fails fast with a targeted
error.
- XML hygiene:
* injects id="<block_id>" on block_replace replacement roots;
* auto-expands self-closing <shape/> and injects <content/> on
shapes for SML 2.0 compliance.
Dry-run surfaces injection errors and renders the same
path-encoded presentationID that Execute sends.
- On backend 3350001 attaches a generic common-causes checklist
(missing block_id / invalid XML / coords out of 960×540).
shortcuts/slides/helpers.go
- ensureXMLRootID: regex tightened to `(?:^|\s)id` so data-id and
xml:id are not matched as root id.
- ensureShapeHasContent: regex `<content(?:\s|/|>)` avoids false
positives like <contention/>; self-closing branch preserves
trailing siblings.
shortcuts/slides/shortcuts.go: register SlidesReplaceSlide.
Tests (package coverage 89.4%; parseReplaceParts and
injectBlockReplaceIDs both reach 100%)
- helpers_test.go: regex edge cases, id override semantics, content
auto-inject across self-closing and open-tag shapes.
- slides_replace_slide_test.go: parameter validation table, URL
resolution (slides / wiki), mixed block_replace + block_insert,
size boundaries, auto-inject behavior, 3350001 hint enrichment,
per-field type-assertion guards, whitespace-only --parts guard
(distinct from the `[]` "at least 1 item" path), replacement
without root element surfaces pre-flight instead of reaching the
backend, and a tight negative assertion that non-3350001 errors
get no slides-specific hint.
Docs (skills/lark-slides)
- SKILL.md: add +replace-slide to the Shortcuts table, register the
new xml_presentation.slide.get / .replace native endpoints,
update core rule 7 to prefer block-level replace over full-page
rebuild now that element-level editing exists, extend the error
table with 3350001 / 3350002 pointing at the replace-slide doc,
add "add image to existing slide via block_insert" as an explicit
Workflow step and symptom-table entry, and refresh the reference
index to include the three new docs below. The old "整页替换" 4-rule
checklist is retired — its one still-relevant guard (new <img>
avoiding overlap) is preserved in the symptom table.
- New references:
* lark-slides-replace-slide.md — flags, parts schema, auto-inject
notes, mixed-action support, 200-item cap, revision_id
semantics, error table, and a "合法根元素速查" cheatsheet for
the eight supported root elements (shape / line / polyline /
img / icon / table / td / chart) with minimal verified XML
snippets. Explicit unsupported list: video / audio / whiteboard
(these appear only as <undefined> export placeholders in SML 2.0).
* lark-slides-edit-workflows.md — recipe-style edit flows covering
the read → modify → write loop and the block_replace vs
block_insert decision tree.
* lark-slides-xml-presentation-slide-get.md — native read API with
block_id extraction examples.
- Fixes across existing references:
* replace / create / delete / presentations.get: add the .data
wrapper in return-value examples, correct jq paths.
* media-upload: fix jq path .file_token → .data.file_token.
* examples.md: annotate auto-inject behavior, replace the
incorrect failed_part_index example with the actual 3350001
error shape.
Empirical corrections (BOE-verified)
- revision_id: stale-but-existing values are accepted; only values
greater than current return 3350002.
- Wrong block_id returns 3350001, not a 200 with failed_part_index.
- Mixed block_replace + block_insert in one call is supported.
- Type-mismatched block_replace (e.g. shape id with a <td>
replacement) is silently accepted by the backend and may destroy
content; 3350001 specifically signals a missing block_id.
Unify lark-cli im +messages-search pagination flags to use int semantics consistently.
Previously, page-limit was registered as an int flag while page-size was still handled as a string flag and parsed manually. This led to inconsistent runtime behavior inside the same shortcut and allowed test helpers to drift from the real CLI flag registration.
Change-Id: Ic4876f4ca7f410a8fe3234e08e41b54ce26990d9
* feat: unify minute artifacts output to ./minutes/{minute_token}/
* fix: tighten path validation and batch-mode --output rejection
* style: translate comments to english and trim historical context
* style: translate leftover chinese comments in vc_notes
* refactor: address review findings across validate ordering, error types, JSON, tests
* fix: sanitize server-provided filename to prevent escape from artifact dir
* style: tighten flag help text for minutes/vc output flags
* docs: update minutes/vc skill docs for unified artifact layout
Add lark-cli wiki +delete-space to delete a knowledge space via
DELETE /open-apis/wiki/v2/spaces/:space_id. When the API returns an
async task_id, the shortcut polls /open-apis/wiki/v2/tasks/:task_id
with task_type=delete_space for a bounded window and emits a
next_command pointing to drive +task_result on timeout. A new
wiki_delete_space scenario is added to drive +task_result for resuming
timed-out deletes.
Change-Id: I75da52b617c206fb778a493ffaa200adf7920a27
* feat(doc): add --from-clipboard flag to docs +media-insert
Allow users to upload the current clipboard image directly to a Lark
document without saving to a local file first.
- New --from-clipboard bool flag (mutually exclusive with --file)
- shortcuts/doc/clipboard.go: readClipboardToTempFile() with per-OS impl
macOS — osascript (built-in, no extra deps)
Windows — PowerShell + System.Windows.Forms (built-in)
Linux — tries xclip / wl-paste / xsel in order; clear install hint
on failure
- No new Go dependencies, no Cgo
- Temp file is created before upload and removed via defer cleanup()
- --file changed from Required:true to optional; Validate enforces
exactly-one of --file / --from-clipboard
* fix(doc): fix clipboard image read on macOS for screenshots and browser-copied images
- Add TIFF fallback (macOS screenshots default to TIFF, not PNG)
- Add HTML base64 fallback (images copied from Feishu/browser embed data URI)
- Use current directory for temp file so FileIO path validation passes
* fix(doc): scan HTML/RTF/text clipboard formats for base64 image data URIs
Extend attempt-3 fallback to iterate all text-based clipboard formats
(HTML, RTF, UTF-8, plain text) rather than only HTML. Any format that
contains a "data:<mime>;base64,<data>" pattern is accepted, covering
images copied from Feishu, Chrome, Safari, and other apps that embed
base64 in non-HTML clipboard slots. Also handle URL-safe base64.
* test(doc): add unit tests for clipboard helpers to meet 60% coverage threshold
Cover decodeHex, hexVal, decodeOsascriptData, reBase64DataURI, and
extractBase64ImageFromClipboard (via fake osascript on PATH).
Package coverage: 57% → 61.2%.
* fix(doc): address CodeRabbit review comments on clipboard feature
- Extend reBase64DataURI regex to cover URL-safe base64 chars (-_) so
URL-safe payloads are matched before decoding is attempted
- Fix readClipboardLinux to continue to next tool when a found tool
returns empty output instead of failing immediately
- Guard fake-osascript test with runtime.GOOS == "darwin" skip
- Use os.PathListSeparator instead of hardcoded ":" in test PATH setup
* fix(doc): replace os.* temp-file clipboard path with in-memory streaming
Fixes forbidigo lint violations in shortcuts/doc: os.CreateTemp, os.Remove,
os.Stat, os.WriteFile are banned in shortcuts/; replaced with vfs.* equivalents
for sips TIFF→PNG conversion, and eliminated temp files entirely elsewhere by
having platform clipboard readers return []byte directly.
- readClipboardDarwin: osascript outputs hex literals decoded in Go (no file I/O)
- readClipboardWindows: PowerShell outputs base64 to stdout, decoded in Go
- readClipboardLinux: tool stdout bytes returned directly
- convertTIFFToPNGViaSips: still needs temp files — uses vfs.CreateTemp/Remove
- DriveMediaUploadAllConfig/DriveMediaMultipartUploadConfig: add Content io.Reader
field so in-memory clipboard bytes skip FileIO.Open() path
- Fix ineffassign in clipboard_test.go (scriptBody double-assignment)
- Update TestReadClipboardLinux_NoToolsReturnsError for new signature
* fix(doc): address CodeRabbit review comments on Linux clipboard path
- Update --from-clipboard flag description to list xclip, xsel and wl-paste
- Preserve last backend-specific error in readClipboardLinux so users see
a meaningful message when a tool is found but fails
- Validate PNG magic bytes for xsel output (xsel cannot negotiate MIME types)
- Add URL-safe base64 regression test for reBase64DataURI
* fix(doc): strip whitespace from base64 payload before decoding clipboard data URI
HTML and RTF clipboard content often line-wraps base64 at 76 characters.
FindSubmatch returns the raw wrapped token so direct decode would fail.
Normalize whitespace with strings.Fields before passing to base64.Decode.
* fix(doc): drop TIFF fallback and internal/vfs import on macOS clipboard
depguard rule shortcuts-no-vfs forbids shortcuts/ from importing
internal/vfs directly. The only caller was the sips TIFF→PNG
conversion, which was already a fragile best-effort fallback that
required temp files.
Remove the TIFF fallback entirely; the remaining two attempts cover
the real-world cases:
1. osascript → PNG hex literal — native screenshots and most apps
2. scan text clipboard formats for base64 data URI — Feishu/browsers
* test(doc): cover readClipboardLinux xsel PNG validation and dispatcher path
Added tests:
- TestReadClipboardLinux_XselRejectsNonPNG: fake xsel that returns plain
text is rejected by the PNG-magic check, preventing text from being
uploaded as an "image".
- TestHasPNGMagic: table-driven coverage of the PNG signature check.
- TestReadClipboardImageBytes_UnsupportedPlatform: exercises the shared
dispatcher post-processing and asserts the (nil, nil) invariant.
Raises clipboard.go diff coverage and brings the package from 61.6% to
63.8% overall.
* test: cover in-memory Content upload paths for clipboard feature
Adds unit tests for the new Content io.Reader branches introduced by
the clipboard feature:
- UploadDriveMediaAll with in-memory Content (drive_media_upload.go 87.5%)
- UploadDriveMediaMultipart with in-memory Content (84.6%)
- uploadDocMediaFile single-part and multipart with clipboard bytes
(doc_media_upload.go 0% -> 88.9%)
Adds TestNewRuntimeContextForAPI helper that wires Factory, context,
and bot identity so package tests can invoke DoAPI without mounting
the full cobra command tree.
* test: cover clipboard Validate/DryRun branches and testing helper
Adds unit tests for the clipboard-related Validate/DryRun paths that
Codecov patch-coverage was flagging as uncovered:
- Validate error when neither --file nor --from-clipboard is supplied
- Validate error when both are supplied (mutual exclusion)
- DryRun output contains <clipboard image> placeholder
- Self-test for TestNewRuntimeContextForAPI so shortcuts/common
sees coverage for the new helper (not just shortcuts/doc)
* test: cover Execute clipboard branch via injectable readClipboardImage
Makes readClipboardImageBytes swappable in tests by routing the call
through a package-level variable readClipboardImage. Tests inject a
synthetic PNG payload so the full Execute clipboard flow
(resolve → create block → upload in-memory bytes → bind) runs under
unit test without a real pasteboard.
Covers:
- TestDocMediaInsertExecuteFromClipboard: end-to-end happy path
- TestDocMediaInsertExecuteClipboardReadError: early-return on
readClipboardImage() failure
* ci: re-trigger pull_request workflow for PR #508
Previous push to 9dedb7a did not trigger the main CI workflow via
the pull_request event (only PR Labels ran). The workflow_dispatch
run I triggered manually lacks PR-scoped secrets so security and
e2e-live failed. An empty commit replays the pull_request event so
the full matrix (deadcode, license-header, security, e2e-live) runs
with proper context.
* test(doc): guard info.Size() behind err check to prevent nil-deref
CodeRabbit flagged that 't.Fatalf("... size=%d err=%v", info.Size(), err)'
evaluates info.Size() even when os.Stat returned (nil, err), which nil-derefs.
Split the check into two stages so the error-path t.Fatalf does not touch
info.
* fix(doc): address fangshuyu-768 review on clipboard PR
Seven code changes driven by review feedback:
1. clipboard.go: stop using CombinedOutput() on osascript / powershell.
Stdout is decoded, stderr is captured separately via cmd.Stderr and
surfaced in the terminal error message, so locale warnings or
AppleEvent permission prompts no longer pollute the hex/base64
payload or mask the real failure.
2. clipboard.go: validate decoded base64 data URI bytes against known
image magic headers (PNG/JPEG/GIF/WebP/BMP). A text clipboard that
happens to contain a literal 'data:image/...;base64,...' fragment
(documentation, tutorials, pasted HTML source) no longer silently
becomes an image upload.
3. clipboard.go: simplify the Linux 'no tool found' install hint to a
distro-agnostic phrasing instead of apt/yum only.
4. clipboard_test.go: delete the stale TestReadClipboardToTempFile_*
tests. They referenced a readClipboardToTempFile function that no
longer exists and only exercised os.CreateTemp/os.Remove. Replace
with TestReadClipboardImageBytes_EmptyResultReturnsError which
actually locks in the 'empty clipboard' → error contract of the
current API (Linux-only since mac/Windows need a real pasteboard).
5. doc_media_upload.go: introduce UploadDocMediaFileConfig struct so
uploadDocMediaFile takes a named config instead of 8 positional
params. Drops the //nolint:lll the old call site had to carry.
6. doc_media_insert.go: convert the clipboard upload call to the new
config struct and only set Config.Content when the clipboard branch
actually produced bytes — this also fixes a latent typed-nil bug
where a nil *bytes.Reader was being passed through an io.Reader
parameter, which tripped the 'if cfg.Content != nil' check in
UploadDriveMediaAll and crashed --file uploads.
7. shortcuts/common/testing.go: TestNewRuntimeContextForAPI now takes
the identity as an explicit core.Identity parameter instead of
hardcoding core.AsBot, and its self-test covers both AsBot and
AsUser. Existing call sites pass core.AsBot explicitly.
Also annotates DryRun output with an 'upload_size_note' when
--from-clipboard is set, since DryRun never reads the pasteboard and
can't predict whether the payload will take the single-part or
multipart path.
* fix(doc): capture line-wrapped base64 in clipboard data URI regex (#586)
HTML and RTF clipboard content commonly folds base64 payloads at
76 chars (standard MIME folding). The previous character class
[A-Za-z0-9+/\-_]+=* stopped at the first \n, so the downstream
strings.Fields normalisation was a no-op (nothing to strip) and
extractBase64ImageFromClipboard silently uploaded a truncated
payload whose 8-byte prefix happened to pass hasKnownImageMagic.
Extend the class to include \s so the Fields strip actually has
whitespace to remove before base64 decoding. Terminators (", <,
), ;) remain outside the class so the match still ends at the
URI boundary.
Add TestReBase64DataURI_LineWrapped covering \n, \r\n, and \t
folds, full round-trip byte-equality, and the terminator-boundary
invariant so any future regression trips a failing test.
* docs(skill): add clipboard-empty fallback guidance for +media-insert
When --from-clipboard returns 'no image data' (empty clipboard, non-image
content, or Linux without xclip/wl-paste/xsel), the agent must NOT silently
swallow the error. It should tell the user the clipboard had no image, ask
for a local file path, then retry the same insert command with --file.
Lists three anti-patterns (silent success, guessing a file path, pre-emptive
save-then-file workaround) that agents have been tempted into.
* docs(skill): user-stated source trumps clipboard/file heuristic
The heuristic table (prefer --from-clipboard when image is on the
clipboard) is a fallback for when the user is vague. If the user
explicitly says 'use the screenshot I just copied' → clipboard; if
they give a path → --file. Agent must not silently swap sources even
when the other looks 'better'.
---------
Co-authored-by: fangshuyu-768 <shuyufang768@outlook.com>
Allow drive +export to request bitable snapshots with --file-extension base and write them with a .base suffix.
Allow drive +import to accept .base files for bitable only, enforce the 20 MB size limit, and document the new examples and constraints.
Add unit tests for validation and size-limit coverage.
Change-Id: Ia13f5013913812df5fc600c43f90918de4ca6b39
Preserve fenced code blocks and balanced-parenthesis URLs when converting markdown to post elements. Add regression tests covering code-block URLs and wiki-style links.
Change-Id: I709a3daf3635402848c96b5122edfc67979ed1a4
When downloading message resources, the saved filename was always derived from
file_key (e.g. file_v2_abc123.xlsx), ignoring the original filename the
sender uploaded. This PR resolves filenames from the Content-Disposition
response header first, falling back to Content-Type-based extension inference
only when the header is absent.
Change-Id: I68b48cf428aa8aded4ad9d55fa042f9d68263c3a
Skill examples taught the pattern --markdown "## A\n\n- x\n- y",
which in bash double quotes is a literal backslash + n, not a
newline. lark-cli forwards the value byte-for-byte to MCP, so
the resulting Feishu doc renders "\n\n" as visible text. Agents
and users copy-pasting the examples reliably produced broken
docs.
Documentation-only fix (issue #580 Option 1, non-breaking):
- Replace 9 "...\n..." examples with multi-line quoted strings,
plus 1 single-quoted example that had the same bug inside
Markdown-block content
- Add a one-sentence warning callout at the top of each file
- Add a stdin/heredoc example in lark-doc-create.md for longer
content
- Leave existing $'...' ANSI-C examples untouched — those
already produce real newlines
No CLI behavior change. Byte-for-byte forwarding is standard
shell semantics; auto-interpreting \n (Option 2) would be a
breaking change and is intentionally not pursued.
Fixes#580
* feat(cmdutil): add X-Cli-Build header for CLI build classification
Adds X-Cli-Build (official / extended / unknown) so the gateway can distinguish official CLI from ISV-repackaged builds.
* test(cmdutil): lift coverage on build-kind classification
Extract classifyBuild as a pure helper so every branch (unknown / extended
main-path / extended credential / extended transport / extended fileio /
official) is reachable without mutating process-wide provider registries.
Also cover: isBuiltinProvider non-pointer values, BuildHeaderTransport
nil-Base fallback path, and fix the Name-spoof test so the test double
returns a value that actually mimics an ISV provider.
Coverage on PR-changed functions:
- classifyBuild: 100% (new)
- computeBuildKind: 61.5% -> 93.3%
- BuildHeaderTransport.RoundTrip: 80% -> 100%
Wrap the POST /drive/v1/permissions/:token/members/apply endpoint as a
user-only shortcut. --token accepts either a bare token or a document
URL, with type auto-inferred from the URL path (/docx/, /sheets/,
/base/, /bitable/, /file/, /wiki/, /doc/, /mindnote/, /minutes/,
/slides/); an explicit --type always wins. --perm is limited to view or
edit; full_access is rejected client-side to match the spec.
Classifier gains two domain-specific hints for the endpoint's newly
documented error codes: 1063006 (per-user-per-document quota of 5/day
reached) and 1063007 (document does not accept apply requests — covers
disallow-external-apply, already-has-access, and unsupported-type).
test(drive): add dry-run E2E for +apply-permission
Invoke the real CLI binary via clie2e.RunCmd under --dry-run and
parse the rendered request JSON with gjson to lock in method, URL
path (including the token segment), type query parameter (auto-inferred
for docx / sheet / slides URLs, taken from explicit --type for bare
tokens), perm body field, and remark presence/omission. A separate
test asserts --perm full_access is rejected by the enum validator
before reaching the server. Fake LARKSUITE_CLI_APP_ID / APP_SECRET /
BRAND are enough because dry-run short-circuits before any API call.
Update drive coverage.md to add a row and refresh metrics.
test(drive): isolate E2E dry-run subprocess from local CLI config
Set LARKSUITE_CLI_CONFIG_DIR to t.TempDir() in both +apply-permission
dry-run tests so the subprocess can't read a developer's real
credentials/profile instead of the fake env vars the tests inject.
test(drive): add E2E case that exercises URL inference override
Previous "bare token with explicit type wins over inference" row used a
bare token, which has no URL-derived type to override. Replace it with
a /docx/ URL + --type wiki combo that actually forces the explicit flag
to win over URL inference, and add a separate bare-token row to keep
the simpler path covered. Refresh coverage.md wording to match.
* fix(base): add default-table follow-up hint to base-create
* fix(base): route base-create hint to stderr
* fix(base): prefix base-create stderr tip
---------
Co-authored-by: kongenpei <kongenpei@users.noreply.github.com>
* fix: skip flag-completion registration outside completion path
Cobra keeps completion callbacks in a package-global map keyed by
*pflag.Flag with no removal path, so registrations made during Build()
outlive the command itself. Route all seven call sites through
cmdutil.RegisterFlagCompletion and enable registration only when the
invocation actually serves a __complete request.
Measured over 30 dropped Builds: ~202 KB / 2180 retained objects per
Build before, ~0 after.
Change-Id: I734d598a4c91a92c33b02e0f292f640cc0e224c6
The <<<<<<< HEAD marker was accidentally left in mail.md and SKILL.md
by commit cb301a3 (draft preview URL). Remove it.
Change-Id: I6e1d5c0c66761302a3c4ee1421a16961b666bd80
* feat(mail): add large attachment support via medias/upload API
When attachments would cause the EML to exceed the 25MB limit, they are
automatically uploaded to the mail attachment storage (medias/upload_all
with parent_type="email") and a download-link card is injected into the
HTML body, matching the desktop client's exportLargeFileArea style.
Key changes:
- Add classifyAttachments: EML-size-based splitting of normal vs oversized
- Add uploadLargeAttachments: upload via medias API with email MountPoint
- Add buildLargeAttachmentHTML: desktop-aligned card with CDN icons
- Add processLargeAttachments: unified entry point for all compose shortcuts
- Add LargeAttachmentHTML to emlbuilder.Builder for HTML block injection
- Fix 7bit line folding: use RFC 5322 limit (998) instead of incorrect 76
- Integrate into +draft-create, +forward, +reply, +reply-all
Known limitation: recipient access to large attachment links requires
backend support to register tokens with the draft (see progress doc).
Change-Id: If8d5938015cac8bc82de3ea3ff41022950f2571e
Co-Authored-By: AI
* refactor(mail): remove legacy size check, add 3GB limit, integrate +send
- Remove checkAttachmentSizeLimit (replaced by processLargeAttachments)
- Remove 25MB pre-check from validateComposeInlineAndAttachments so that
large files reach Execute where they are uploaded as large attachments
- Integrate processLargeAttachments into +send shortcut
- Add 3GB single file limit aligned with desktop client
- Clean up unused imports from helpers.go and helpers_test.go
Change-Id: Ie590ad2b58263c075f48b338959b8f5b3f912f85
Co-Authored-By: AI
* feat(mail): quote-aware HTML insertion, +draft-edit support, cleanup emlbuilder
- Add insertBeforeQuoteOrAppend: insert large attachment HTML before the
quote block (lark-mail-quote) instead of appending to body end, matching
desktop's exportLargeFileArea placement logic
- Add preprocessLargeAttachmentsForDraftEdit: intercept add_attachment
patch ops before draft.Apply, upload oversized files, inject HTML into
snapshot's HTML body Part directly. No changes to draft sub-package.
- Remove LargeAttachmentHTML field/setter/logic from emlbuilder — it was
business logic (quote-aware insertion) that doesn't belong in a generic
EML builder. processLargeAttachments now sets the full HTML body via
bld.HTMLBody() after merging the large attachment card at the right position.
- All compose shortcuts pass htmlBody to processLargeAttachments for
quote-aware insertion (composedHTMLBody for reply/forward, body for others).
Change-Id: If6e7ed7e77989ab9a8a41a93758f686d72ccf497
Co-Authored-By: AI
* fix(mail): align large attachment HTML IDs with desktop client
- Container ID: lark-mail-large-file-container → large-file-area (matching
desktop's MAIL_LARGE_FILE_CONTAINER constant)
- Item ID: lark-mail-large-file-item → large-file-item (matching
desktop's MAIL_LARGE_FILE_ITEM constant)
- Timestamp: truncate to 9 digits (matching TIMESTAMP_CUT_OUT_ID = 9)
- Refactor HTML generation to use template constants for readability
These IDs are used by the desktop client's BigAttachmentPlugin
([id^=large-file-area]) and the server's LargeFileRule to identify and
remove the HTML block when rendering the attachment card UI.
Change-Id: Ib5a77a1a3d60eeb3a05c585f2af0a5ddaacf887b
Co-Authored-By: AI
* docs(mail): document large attachment behavior in skill references
Update --attach parameter descriptions across all compose shortcuts
(+send, +reply, +reply-all, +forward, +draft-create, +draft-edit) to
describe automatic large attachment handling when EML exceeds 25 MB.
Change-Id: I8c30e390c127ea1119cb8c4b83ec636e41fbaf66
Co-Authored-By: AI
* fix(mail): pass signature-injected HTML to processLargeAttachments
When both --signature-id and large attachments are used, the htmlBody
passed to processLargeAttachments must include the already-injected
signature. Previously mail_send and mail_draft_create passed the
original body, causing processLargeAttachments to overwrite the
signature-injected HTML body when inserting the large attachment card.
Use composedHTMLBody variable (same pattern as reply/forward) to
capture the full processed HTML including signature.
Change-Id: I6be330776abca704b10cc3b8bfd5e20838e6e538
Co-Authored-By: AI
* fix(mail): skip draft.Apply when all ops consumed by large attachment preprocessing
When all patch ops are add_attachment targeting oversized files,
preprocessLargeAttachmentsForDraftEdit uploads them and removes the
ops from the patch. The resulting empty patch caused draft.Apply to
fail with "patch ops is required". Now skip Apply when no ops remain.
Change-Id: I8067a54b5f849fa519e8344a7eb10c48f58e54b8
Co-Authored-By: AI
* fix(mail): add X-Lms-Large-Attachment-Ids header in draft-edit large attachment flow
draft-edit's preprocessLargeAttachmentsForDraftEdit uploaded oversized files
and injected HTML cards but never wrote the X-Lms-Large-Attachment-Ids header
into the snapshot, so the mail server could not associate the attachments with
the draft. Merge new token IDs with any existing ones already in the snapshot.
Also extract the duplicated largeAttID struct and header name string into
package-level declarations.
Change-Id: Id256d948ec07e86296157436feefa3c2052af721
Co-Authored-By: AI
* fix(mail): i18n large attachment HTML text aligned with desktop client
Parameterize title and download text in large attachment HTML templates.
Chinese lang uses "来自Lark邮箱的超大附件"/"下载", others use
"Large file from Lark Mail"/"Download", matching desktop's i18n keys
Mail_Attachment_AttachmentFromFeishuMail and Mail_Attachment_Download.
Change-Id: I2aada8d52af41ae77dd7001d24d14e333f12066e
Co-Authored-By: AI
* fix(mail): insert large attachment card before quote wrapper, not inside nested quote
insertBeforeQuoteOrAppend matched id="lark-mail-quote" which can appear
deeply nested inside quoted content from previous replies in a thread.
This caused the card to be placed inside the quote area instead of before
it. Switch to matching the "history-quote-wrapper" class which is the
outermost quote container generated by the CLI.
Change-Id: I720b6d62d719613b411b7ed4b7820a1535bf14bd
Co-Authored-By: AI
* feat(mail): unify large attachment handling in +draft-edit with normal attachments
Extend +draft-edit so that large attachments behave like normal attachments
from the user's perspective: survive body edits, are listed in inspect
output, and are removed via the same remove_attachment op.
Code-wise:
- remove_attachment target now accepts token (for large attachments) in
addition to part_id / cid; priority part_id > cid > token.
- setBody / setReplyBody auto-preserve the large attachment card in the
HTML body, mirroring how normal attachments (MIME parts) survive body
edits. Detection checks only the user-authored region of the value so
cards inside an appended quote block (from the original quoted message)
are not mistaken for user-supplied cards.
- --inspect returns large_attachments_summary (token, filename, size) by
parsing the X-Lms-Large-Attachment-Ids header and the HTML card DOM.
- Well-known Lark HTML/header constants (LargeAttachmentIDsHeader,
LargeFileContainerIDPrefix, LargeFileItemID, LargeAttachmentTokenAttr)
moved to the draft package alongside QuoteWrapperClass; the mail package
consumes them.
- Shared helpers FindHTMLBodyPart and InsertBeforeQuoteOrAppend exported
from the draft package; mail package switched to consume them, removing
local duplicates.
Skill reference (lark-mail-draft-edit.md) updated: three locator fields by
attachment type, unified remove_attachment examples, set_body behavior.
Change-Id: Ic064d1a8df0edf1cef6069cd44ec2a7534cd2182
Co-Authored-By: AI
* fix(mail): place signature before large attachment card consistently
When inserting a signature into a draft that already has a large
attachment card, the signature was placed after the card, diverging from
the compose-time layout where the order is [user][sig][card][quote].
Root cause: insertSignatureOp split only at the quote block, so the
"user region" side inadvertently included the card.
Centralize signature placement in draft.PlaceSignatureBeforeSystemTail,
which splits at the earliest system-managed element (card or quote,
whichever comes first). Both edit-time insertSignatureOp and compose-time
injectSignatureIntoBody now share this single source of truth, removing
the duplicated HTML splicing logic.
Change-Id: I234bfebaaa31a32731ebbaa78c6596a72618b7c5
Co-Authored-By: AI
* fix(mail): auto-preserve signature in set_body and set_reply_body
Previously set_body / set_reply_body replaced the entire HTML body,
silently dropping the signature block. The "replace whole body" semantic
treated signature as user-authored content, which is inconsistent with
how attachments (normal + large) and quote blocks survive body edits —
signature is a system-managed element managed via insert_signature /
remove_signature ops.
Unify the mental model: body-edit ops replace user-authored content
only; signature, large attachment card, normal attachments, and (for
set_reply_body) quote block are all auto-preserved. Users can override
by including equivalents in value, or explicitly delete via dedicated
ops (remove_signature, remove_attachment).
- Add ExtractSignatureBlock helper (symmetric to RemoveSignatureHTML).
- Rename autoPreserveLargeAttachmentCard to
autoPreserveSystemManagedRegions; extract and inject both sig and card
from old body, respecting user-supplied equivalents in value's
user-authored region.
- Update skill doc and patch template notes to reflect the new
semantics consistently.
Change-Id: I96660d2ff06a6c9cdf1b86793c2d89cf9cb09ffe
Co-Authored-By: AI
* fix(mail): use brand-aware display name in large attachment card title
The title "Large file from Lark Mail" / "来自Lark邮箱的超大附件" hard-coded
"Lark" regardless of brand. The desktop client switches between
"Feishu"/"飞书" and "Lark" based on the APP_DISPLAY_NAME i18n
substitution.
Add brandDisplayName(brand, lang) helper:
- BrandLark → "Lark"
- BrandFeishu → "飞书" (zh) / "Feishu" (en)
Applied to title in buildLargeAttachmentHTML, aligning with the icon CDN
and download URL, which already branch on brand.
Change-Id: I06258b9982b6280a2230193d90a6a88884e10aa3
Co-Authored-By: AI
* style(mail): apply gofmt
CI fast-gate check flagged gofmt-unformatted files. Run gofmt -w on
touched mail files only.
Change-Id: Iec690dc63adfaa54b8f7c85ab5b3ca035476ddbd
* fix(mail): address review feedback on large attachment PR
- Strip <html><head><body> wrapper from xhtml.Render output in
removeLargeFileItemFromHTML to avoid polluting the HTML body
- Reject plain-text messages with oversized attachments instead of
silently losing the body content
- Fix attachment count limit in skill doc (100 → 250)
- Remove unused fio/attachFlag params from validateComposeInlineAndAttachments
- Add token escaping test for large attachment HTML builder
Change-Id: Ie589a1f1d204b0aeebc4486b16bb435041793ceb
Co-Authored-By: AI
* fix(mail): recognize server-format X-Lark-Large-Attachment header in draft-edit
When a draft with large attachments is created by the desktop client,
the server returns X-Lark-Large-Attachment (with file_key/file_name/
file_size fields) instead of the CLI-written X-Lms-Large-Attachment-Ids.
Previously CLI only recognized its own header, causing existing large
attachments to be silently dropped when the draft was edited.
- Parse both header formats via IsLargeAttachmentHeader and unified
largeAttHeaderEntry struct
- Convert server-format entries to CLI-format on save so the server
can process the update
- Fix inline attachment classification: require non-empty CID to
classify as inline image (large attachments may have is_inline=true
but no CID)
Change-Id: Ie7def4fc5923d2cf3446eedfbca4fd8cae44bfac
Co-Authored-By: AI
* fix(mail): skip large attachments in forward URL validation
Large attachments do not have download URLs since they are referenced
by token, not embedded in the EML. Validate only normal attachments
to avoid false "missing download URL" errors when forwarding messages
that contain expired or token-based large attachments.
Change-Id: Ibe3f45390cd3b3cbe6ddd15961dcda4f17aefe4f
Co-Authored-By: AI
* fix(mail): classify forwarded original attachments for large attachment upload
Previously, all original attachments were unconditionally embedded in
the EML before user attachments were processed for large attachment
upload. When original + user attachments together exceeded the 25 MB
EML limit, the build would fail.
Now all attachments (original + user-added) are classified together
via classifyAttachments. Original attachments that push the EML over
the limit are re-uploaded as large attachments with download cards,
matching the compose/reply flow behavior.
Also refactors uploadLargeAttachmentBytes to reuse the shared
common.UploadDriveMediaAll utility (via new Reader field on the
config struct) instead of duplicating the upload logic, and replaces
bare fmt.Errorf with output.ErrValidation for user input errors.
Change-Id: I98d4ad8960cd68e38765b05c94f7786d6a8444c8
Co-Authored-By: AI
* fix(mail): normalize large attachment header on draft edit to prevent loss
Server returns X-Lark-Large-Attachment header on draft readback, but only
recognizes X-Lms-Large-Attachment-Ids on write. Without normalization,
editing a draft with existing large attachments (e.g. adding a small
attachment) would send back the server-format header unchanged, causing
the server to drop the large attachment association.
Add normalizeLargeAttachmentHeader() at the entry of
preprocessLargeAttachmentsForDraftEdit to convert server-format headers
to CLI format before any processing or early return.
Change-Id: Id99a46f29015a32921bfb72a003f766c397787e1
Co-Authored-By: AI
* fix(mail): extract large attachment card from quote on forward
When forwarding a message that contains large attachments, the original
message's download card (large-file-area div) was left inside the
forward quote block. Extract it and place it in the main body area
(after signature, before quote), matching the desktop client behavior.
Change-Id: Iebede35cdf4ed0f65b72bce28ffb18af21ddf668
Co-Authored-By: AI
* fix(mail): use octet-stream for re-embedded attachments and file-based large upload on forward
- Use application/octet-stream instead of original content type when
re-embedding downloaded attachments in forward EML. Prevents the mail
server from treating image/* attachments as inline parts.
- Replace in-memory uploadLargeAttachmentBytes with temp-file-based
uploadLargeAttachments for oversized original attachments. This
enables multipart upload for files >20MB which the single-part API
does not support.
Change-Id: Ib02add5710e8b052e47b513ed3d9a688e0f98212
Co-Authored-By: AI
* fix(mail): address PR review — blocked extension bypass, index-based op filtering, plain-text draft guard
1. Move CheckBlockedExtension into statAttachmentFiles so oversized
attachments are validated before classification, covering compose,
draft-edit, and forward paths.
2. Replace path-based oversized op filtering with SourceIndex-based
filtering in preprocessLargeAttachmentsForDraftEdit to avoid
incorrectly removing duplicate-path normal ops.
3. Add HTML body preflight in preprocessLargeAttachmentsForDraftEdit
before uploading, so plain-text-only drafts fail early instead of
silently producing a draft with tokens but no download card.
Change-Id: Ib8771812f50a18f00a40e50149b028b8aaa101fe
Co-Authored-By: AI
* fix(mail): preserve original content type for normal forwarded attachments
The octet-stream override was only needed for the large attachment
upload path (to prevent image/* from being treated as inline by the
drive API). Normal attachments embedded in the EML should retain their
original MIME type so recipients can preview/open them correctly.
Change-Id: Ie40b7c362524a3b82255b58e9bcfd770eacfe911
Co-Authored-By: AI
* fix(mail): reconstruct missing large attachment HTML cards on draft edit
The server strips HTML download cards from the EML body when storing
drafts, so every draft read-back (regardless of creator) lacks them.
Add ensureLargeAttachmentCards which runs before header normalization,
compares server-format header tokens against existing HTML cards via
data-mail-token, and rebuilds only the missing ones. This ensures
external recipients see download links after draft-edit → send.
Also exports ParseLargeAttachmentSummariesFromHeader and
ParseLargeAttachmentItemsFromHTML from the draft package for
cross-package use.
Change-Id: I9cb0f47a9f4582909de24984d9a9f6e366521e62
Co-Authored-By: AI
* feat(mail): support large attachments in plain-text emails
Previously large attachments required an HTML body for the download card.
Now plain-text emails (--plain-text or text/plain-only drafts) get download
info appended as structured text (title + filename + size + URL), with
i18n and brand awareness matching the HTML card.
Changes:
- Add buildLargeAttachmentPlainText and injectLargeAttachmentTextIntoSnapshot
- Add FindTextBodyPart in draft/projection.go
- Update processLargeAttachments to accept textBody parameter
- Update ensureLargeAttachmentCards to handle text/plain body reconstruction
- Update preprocessLargeAttachmentsForDraftEdit to allow text/plain drafts
- Update all callers (send, draft-create, reply, reply-all, forward)
Change-Id: I3b375e2ff34697eeb73a3768ace6d577d1bead3e
Co-Authored-By: AI
* fix(mail): FindBodyPart skips attachment-disposition parts; update skill docs
FindHTMLBodyPart and FindTextBodyPart now skip parts with
Content-Disposition: attachment, preventing .txt/.html file attachments
from being mistakenly treated as the email body.
Also update all lark-mail skill reference docs to reflect that large
attachments now work in both HTML (download card) and plain-text
(download link text) modes.
Change-Id: I1e6da4fd614217dff61304212304b5fd80c8246c
Co-Authored-By: AI
* fix(mail): fix origIdx mismatch, predictable temp files, and attachment count on forward
- Use SourceIndex instead of linear origIdx counter so classifyAttachments
reordering does not cause content mismatch between normal/oversized loops
- Use os.CreateTemp for temp files instead of predictable names in CWD
- Include original large attachment count in totalCount limit check
Change-Id: Ide5dce14b1efc672687800d77c3853f15dfc191b
Co-Authored-By: AI
* fix(mail): use composed body size and source inline bytes in EML size estimation
estimateEMLBaseSize was using len(body) (raw --body flag) instead of the
actual composed body (which includes quotes, signatures, forward headers).
Source inline images downloaded from the original message were also not
counted. This could cause borderline attachments to be misclassified.
- Use len(composedHTMLBody) + len(composedTextBody) for body size
- Return total downloaded bytes from addInlineImagesToBuilder and pass
as extraBytes to estimateEMLBaseSize
- Fix applied to all compose shortcuts: send, draft-create, reply,
reply-all, forward
Change-Id: Ibe6c44e22d40ac51f0a4652d279e66bd92330723
Co-Authored-By: AI
* fix(mail): merge large attachment items into single container on draft edit
When draft-edit had both set_body and add_attachment (oversized), the
ensureLargeAttachmentCards and preprocessLargeAttachmentsForDraftEdit
each created independent large-file-area containers. The subsequent
set_body's autoPreserveSystemManagedRegions only captured the first
container via SplitAtLargeAttachment, discarding the second one.
Fix: injectLargeAttachmentHTMLIntoSnapshot now detects an existing
large-file-area container and appends new items inside it instead of
creating a new container, matching the desktop client's single-container
behavior.
Change-Id: I3d701683053842f1d7bdad34fc4b2ef26ede784e
Co-Authored-By: AI
* fix(mail): strip large attachment card from reply/reply-all quote
Reply and reply-all should not carry over the original email's large
attachment HTML card into the quoted block. Extract the shared
stripLargeAttachmentCard helper (also used by forward) that removes
the card from orig.bodyRaw before quote construction.
- Reply/reply-all: card is discarded (not re-inserted)
- Forward: card is moved to body area before the quote (unchanged)
Change-Id: I5399bb901c120206c7c045bed107f7d68be23bb1
Co-Authored-By: AI
* fix(mail): skip invalid attachments on forward instead of blocking
When forwarding a message with deleted/expired attachments, the forward
flow now automatically removes them instead of either blocking (normal
attachments) or silently including dead references (large attachments).
- Propagate failed_ids from fetchAttachmentURLs into composeSourceMessage
- Skip failed attachments in the forward download loop with a warning
- Remove corresponding large attachment HTML card items from the body
- Extend itemContainsToken to match server-generated href?token= format
Change-Id: I9c0096dcbe96f1d61caa0f6f0b2f8b738fdfa66b
Co-Authored-By: AI
* fix(mail): restore dry-run file preflight and reserve card overhead in classifier
1. Restore file existence and blocked-extension checks in
validateComposeInlineAndAttachments so --dry-run surfaces local
path errors before Execute.
2. Reserve 3KB per oversized file in classifyAttachments to account
for the HTML card / plain-text block injected after classification.
Change-Id: Ib48a75f86a50298413c1f9ab8226e583c0161a8c
Co-Authored-By: AI
* fix(mail): revert classifier overhead reserve for simplicity
The 3KB-per-oversized-file reserve in classifyAttachments addressed
a boundary case that is practically impossible to trigger (requires
Normal attachments to fill to within a few KB of 25MB). Remove it
to keep the classifier simple.
Change-Id: I5148f14ecca1a0dee677a1a2c60ec4efab160ea8
Co-Authored-By: AI
* style(mail): fix gofmt indentation in draft create tests
Change-Id: Ib41aa22f94144f2d47b12675d444aa43cb333a88
Co-Authored-By: AI
* fix(mail): remove temp files in forward, use in-memory upload instead
Replace os.CreateTemp/os.WriteFile/os.Remove with in-memory Data field
on attachmentFile, conforming to the project's forbidigo rule against
temp files in shortcuts. Also remove dead uploadLargeAttachmentBytes.
Change-Id: Ic26e4025eebfa1bac3948438ef185ff3e2f15abb
Co-Authored-By: AI
* test(mail): add tests for validateComposeInlineAndAttachments and fileTypeIcon
Covers all branches: inline+plain-text conflict, inline+non-HTML body,
missing file, blocked extension, valid pass-through, and all file type
icon mappings.
Change-Id: I8b81c1b34010a9ecb7153462a5524e3d7b171de2
Co-Authored-By: AI
* test(mail): improve coverage for large attachment and draft edit functions
Add tests for snapshotEMLBaseSize, flattenSnapshotParts, estimateEMLBaseSize,
normalizeLargeAttachmentHeader, processLargeAttachments error paths,
preprocessLargeAttachmentsForDraftEdit early-return paths, inject edge cases,
buildLargeAttachmentItems, statAttachmentFiles edge cases, and
prettyDraftAddresses.
Change-Id: Ie661e6ebea63512864d97e20135dd89cb9e9304e
Co-Authored-By: AI
* fix(docs): validate --selection-by-title format early
* fix(docs): reject multiline selection-by-title before prefix check
* chore: refresh CI against current main (no code change)
* test(doc): cover DocsUpdate.Validate integration for selection-by-title
codecov/patch was at 27.27% because the PR added three lines to the
Validate closure (the `if err := validateSelectionByTitle(selTitle); err
!= nil { return err }` block) but nothing in the test file exercised
that closure — only the helper function was tested directly.
TestDocsUpdateValidate now builds a bare RuntimeContext via
common.TestNewRuntimeContext, sets the relevant flags on a cobra
command, and calls DocsUpdate.Validate(ctx, rt) across five cases:
1. Heading-style selection-by-title passes — covers the happy path
through the new call site and the final `return nil`.
2. Plain-text title is rejected with heading-prefix guidance —
covers the new error branch.
3. Multi-line title is rejected as not a single heading line —
covers the other error branch inside the helper.
4. Invalid --mode is still rejected first — proves the new check
doesn't swallow pre-existing validation.
5. Conflicting --selection-with-ellipsis + --selection-by-title is
rejected at the mutual-exclusion check — same ordering contract.
Coverage profile confirms the three added production lines
(docs_update.go L65-67) are now hit: condition 3x, error branch 2x,
happy path via the closure's return nil 1x.
* refactor(cmd): split Execute into Build with IO/Keychain injection
Introduce a public cmd.Build entry point so external consumers (cli-server,
MCP server, other embedders) can assemble the full CLI command tree without
going through os.Args or the platform keychain. Build takes an
InvocationContext plus functional BuildOptions:
* WithIO(in, out, errOut) — inject custom streams; terminal detection
is derived from the input's underlying *os.File when present.
* WithKeychain(kc) — swap the credential store.
* HideProfile(bool) — registered later in cmd.HideProfile.
The existing Execute() keeps using the internal buildInternal (which
still returns the Factory so error handling can attribute exit codes),
and SetDefaultFS replaces the global VFS implementation at startup.
Hardening applied up front:
* cmdutil.NewIOStreams(in, out, errOut) centralizes terminal detection
so SystemIO() and WithIO share one path.
* cmdutil.NewDefault normalizes partial IOStreams — callers may pass
&IOStreams{Out: buf} without tripping nil-writer panics in the
RoundTripper warnings, Cobra, or the credential provider.
* Build guards against nil functional options.
* An API contract test (cmd/build_api_test.go) exercises Build +
WithIO + WithKeychain + HideProfile + SetDefaultFS so the public
surface is reachable by deadcode analysis.
Change-Id: I7c895e6019817401accbde2db3ef800da40ad319
* feat(schema): filter methods by strict mode in schema output
When strict mode is active, schema output now excludes methods that
are incompatible with the forced identity. This applies to both
pretty and JSON output formats at the resource and method levels.
Change-Id: I39647d5578466c3e23dc545bfb917ae075203ad7
* refactor: centralize strict-mode as flag registration
Change-Id: Iec11151c5002c2f58a8aa067d08747db2e4d2d8c
* fix(cmd): align strict-mode completion and build context; drop dead register shims
Thread a context.Context through RegisterShortcuts, RegisterServiceCommands,
and service.registerService/Resource/Method by introducing explicit
*WithContext variants. Pass that context into NewCmdServiceMethodWithContext
so shortcut and service command construction can honor cancellation and
strict-mode pruning consistently.
Also drop the context-less registerMethod and registerResource shims —
they became unreachable once the WithContext variants took over, and
were the source of new deadcode warnings. registerService is retained
because service_test.go still calls it directly.
Change-Id: I3fe5673aed663c7383bbbc5b0ae94d1f3491f22d
* refactor(cmd): hide --profile in single-app mode via build option
- GlobalOptions gains HideProfile; RegisterGlobalFlags stays pure and reads
the policy off the struct. No boolean-trap parameter, one call per site.
- buildConfig holds GlobalOptions inline so HideProfile(bool) BuildOption
mutates it directly. buildInternal stays a pure assembly function and
requires callers to supply WithIO — no implicit os.Std* fallback.
- Add WithIO BuildOption (wrapping raw io.Reader/Writer with automatic
*os.File TTY detection); Execute injects streams explicitly and decides
profile visibility via HideProfile(isSingleAppMode()).
- installTipsHelpFunc force-shows hidden root flags while rendering the
root command's own help, so single-app users still discover --profile
via lark-cli --help without it polluting subcommand helps.
Change-Id: I7755387e993992ca969e0a4a6f54441cc1993eef
* feat(transport): extension abort hook and shared base transport
Two transport-layer changes bundled because both reshape the base
round-tripper contract used by the HTTP client, the Lark SDK client,
and the in-process updater.
1. Extension abort hook (PreRoundTripE).
Extensions implementing exttransport.AbortableInterceptor can now
return an error from PreRoundTripE to skip the built-in chain. The
post hook still fires with (nil, reason) so extensions can unwind
resources. extensionMiddleware captures the provider name so the
returned *AbortError carries attribution.
2. Shared base transport to stop RPC leak.
util.NewBaseTransport cloned http.DefaultTransport on every call, so
each cmdutil.Factory produced a fresh *http.Transport whose
persistConn readLoop/writeLoop goroutines lingered until
IdleConnTimeout (~90s). Invisible in a single-process CLI, but the
fork is consumed by cli-server where each RPC request constructs a
new Factory, causing linear memory + goroutine growth under load.
Replace NewBaseTransport with SharedTransport — returns
http.DefaultTransport (the stdlib-wide singleton) by default, and
a cached proxy-disabled clone only when LARK_CLI_NO_PROXY is set.
Return type is http.RoundTripper to discourage in-place mutation of
the shared instance. FallbackTransport is kept as a thin
*http.Transport wrapper so existing callers in internal/auth and
internal/cmdutil transport decorators (which were already on the
singleton path) do not have to migrate.
Leak-site migrations: factory_default.go (HTTP + SDK base) and
update.go now call SharedTransport directly.
Change-Id: Ia82462134c5c5ee838be878b887860f41446a235
* fix: unblock Build() zero-opts path and sidecar demo build
Two regressions surfaced on refactor/build-execute-split:
1. cmd.Build(ctx, inv) without WithIO panicked at rootCmd.SetIn/Out/Err
because cfg.streams stayed nil — NewDefault normalized internally
but cmd/build.go never saw the normalized value. Default cfg.streams
to cmdutil.SystemIO() before the root command wires them, and add a
TestBuild_NoOptions regression guard.
2. sidecar/server-demo/main.go still called cmdutil.NewDefault(inv),
so `go build -tags authsidecar_demo ./sidecar/server-demo` failed
with "not enough arguments". Pass nil for the new streams parameter
to preserve the prior behavior (NewDefault substitutes SystemIO).
Change-Id: I20227b2355cde7d19e22eba3eb841c6d8611e8a7
* feat(doc): add pre-write semantic warnings to docs +update
Two static checks run before the MCP update-doc call:
1. replace_* + blank-line markdown: replace_range / replace_all only
swap text inside an existing block — a \n\n in the payload will
render as literal text, not a paragraph break. Hint to use
delete_range + insert_before instead.
2. Combined bold+italic emphases (***text***, **_text_**, _**text**_)
cannot round-trip through Lark and are silently downgraded to a
single emphasis. Hint to split into two separate emphases.
Both warnings go to stderr and never block the update — they inform,
not gate. Adds table-driven tests for each check plus an aggregation
test, and wires the checks into Execute right before CallMCPTool.
Closes the first batch of items from the docs +update pitfalls
review (Cases 1 and 5).
* fix(doc): exclude code regions and escaped markers from docs +update checks (#578)
* fix(doc): exclude code regions and escaped markers from docs +update checks
Addresses the three review comments on #569: the blank-line paragraph
check and the bold+italic emphasis check both operate on the raw
markdown string, so fenced code blocks / inline code spans / literal
escaped markers produce false-positive warnings on content users
expect to pass through verbatim.
Changes:
- Add proseHasBlankLine(): fence-aware detector that returns true only
when a blank line sits outside of ```...``` or ~~~...~~~ regions.
Replaces the raw strings.Contains("\n\n") check in
checkDocsUpdateReplaceMultilineMarkdown.
- Add stripMarkdownCodeRegions(): blanks out fenced code lines and
masks inline code spans (via scanInlineCodeSpans from markdown_fix.go)
with equal-length whitespace so byte offsets outside the stripped
regions are preserved.
- Add stripEscapedEmphasisMarkers(): removes "\*" and "\_" so literal
sequences like "\***text***" — which CommonMark renders as a literal
asterisk plus bold — don't match the combined bold+italic regex.
- Wire both helpers into checkDocsUpdateBoldItalic(): the regex now runs
on stripEscapedEmphasisMarkers(stripMarkdownCodeRegions(markdown)),
so code samples and escaped markers are sanitized away before
detection.
Shared fence-parsing helpers (codeFenceOpenMarker, isCodeFenceClose,
leadingRun) are kept local to this file to avoid touching files outside
the scope of the reviewed PR. If a future change wants to reuse them
across the doc package, they can be promoted then.
Tests:
- TestCheckDocsUpdateReplaceMultilineMarkdown: add 4 negative/positive
cases — blank line inside backtick and tilde fences (no flag), blank
line in prose while fence also has blanks (flag wins), fenced code
with no blank lines (no flag).
- TestCheckDocsUpdateBoldItalic: add 9 cases — ***text*** / **_text_** /
_**text**_ inside fenced code (backtick and tilde), inside inline
code spans, and escaped \***text*** / \*\*_text_\*\* (none flagged);
plus two positive cases to verify the strip doesn't over-sanitize
(real emphasis in prose still fires when inline/fenced code is nearby).
* fix(doc): close CommonMark gaps and add three more combined-emphasis shapes
Self-review of the first commit turned up three issues:
- isCodeFenceClose was strict on exact marker length. Per CommonMark
§4.5, a closing fence must be at least as long as the opener, not
exactly the same length. A 3-backtick open legitimately closed by a
4-backtick closer (used to embed triple-backticks inside the code
sample) was left open-ended, causing the rest of the document to be
treated as code and both checks to silently skip it.
- Both fence helpers accepted any amount of leading whitespace because
they ran on strings.TrimSpace(line). CommonMark allows 0..3 leading
spaces before a fence marker; 4+ spaces (or any tab in leading
position, which expands to 4 columns) makes the line indented code
block content, not a fence open/close. Indented fence-like lines now
correctly remain prose and blank lines around them are detected.
- The bold/italic check only covered three of the six documented
combined-emphasis shapes. Added ___text___, __*text*__, and
*__text__* so parity with the asterisk variants is complete. The
regex set is now table-driven (combinedEmphasisPatterns) to make
adding future shapes a one-line change.
Implementation changes:
- New fenceIndentOK(line) helper: returns (body, true) for 0..3 leading
spaces with no tabs, else (_, false). Used by both codeFenceOpenMarker
and isCodeFenceClose.
- isCodeFenceClose now counts the fence-char run and accepts any run
length >= len(marker), with trailing whitespace only.
- checkDocsUpdateBoldItalic replaced three named var regexes with a
table of six {shape, re} entries and a single early-exit loop.
- Updated docsUpdateWarnings top docstring to list all six shapes.
- Noted the known limitation of stripEscapedEmphasisMarkers around
doubled backslash escapes ("\\***text***"), which is a false negative
we accept in exchange for keeping this a simple string replace.
Test additions (docs_update_check_test.go):
- Fence close: longer-marker close correctly ends fence; real prose
blank after a longer-close fence is still detected.
- Indentation: 4-space indented fence-like line is not a fence open,
so a surrounding blank line still flags; tab-indented variant same;
3-space indented fence is still a real fence.
- New shapes: ___text___ positive + all three negative-guards (fenced
code, inline code, escaped); __*text*__ and *__text__* positive +
fenced/inline negative-guards; plus two composition tests to ensure
the strip does not over-sanitize across the six-regex alternative set.
All 53 sub-tests in this file pass; go vet and gofmt are clean.
---------
Co-authored-by: fangshuyu-768 <shuyufang768@outlook.com>
* fix(doc): address CodeRabbit review on docs +update warnings (#581)
Two CodeRabbit nits from #569:
1. Unit test hint assertion only checked for `delete_range` in the
remediation message; the companion `insert_before` half of the
guidance could regress undetected. Broaden the assertion to require
both tokens so a future edit that drops half the remediation
produces an immediate test failure.
2. No E2E coverage proved the dry-run contract in the PR description
("Not emitted in dry-run mode — kept quiet during planning"). The
helper itself is unit-tested, but nothing caught a regression where
a later refactor wired docsUpdateWarnings into the DryRun path.
Add tests/cli_e2e/docs/docs_update_dryrun_test.go:
TestDocs_UpdateDryRunSuppressesSemanticWarnings invokes
`docs +update --dry-run --mode=replace_range --markdown "***x***\n\nb"`
— an input crafted to trip BOTH pre-write warnings — and asserts
neither the "warning:" prefix, the blank-line message, nor the
combined-emphasis message appears on stdout or stderr.
Note: the file needs -f to add because .gitignore has a bare
`docs/` rule that accidentally matches tests/cli_e2e/docs/. The
existing tracked files under that directory predate the rule; new
additions have to be force-added until the ignore pattern is
narrowed. Not worth rewriting .gitignore for one file.
Verified manually that the new E2E fails cleanly when warnings are
injected into DryRun and passes again after reverting — the test has
real regression-detection power, not just a sticker.
Co-authored-by: fangshuyu-768 <shuyufang768@outlook.com>
- Register DocMediaUpload in doc/shortcuts.go (was defined but never
registered, so lark-cli docs +media-upload was unavailable)
- Rename MediaUpload to DocMediaUpload for consistency with
DocMediaInsert/DocMediaPreview/DocMediaDownload
- Add whiteboard to --parent-type flag description
- Update --parent-node description to mention board_token for whiteboard
Drive +upload (parent_type=explorer) produces file tokens that the
whiteboard API does not recognize (500 error). The correct approach
is docs +media-upload with parent_type=whiteboard.
* feat(doc): add --after-keyword/--before-keyword flags to +media-insert
Allows inserting images/files at a position relative to the first block
whose plain text matches a keyword (case-insensitive substring match).
- Add --after-keyword: insert after the matched root-level block
- Add --before-keyword: insert before the matched root-level block
- Flags are mutually exclusive; default behavior (append to end) unchanged
- fetchAllBlocks: paginated block listing (up to 50 pages × 200 blocks)
- extractBlockPlainText: covers text, heading1-9, bullet, ordered, todo, code, quote
- findInsertIndexByKeyword: walks parent_id chain to resolve nested blocks to their root-level ancestor
- DryRun updated to show block-listing step when keyword flag is set
* test(doc): add fetchAllBlocks pagination and keyword dry-run coverage
- TestFetchAllBlocksPaginationViaExecute: exercises fetchAllBlocks via a
full Execute flow with --after-keyword, covering multi-page block listing
(fetchAllBlocks was previously at 0% coverage)
- TestDocMediaInsertDryRunWithAfterKeyword: verifies that the dry-run output
includes a block-listing step and mentions "search blocks" in the
description when --after-keyword is provided
fetchAllBlocks coverage: 0% → 76.2%
* refactor(doc): use MCP locate-doc for keyword-based block positioning
Replace fetchAllBlocks + keyword scan with MCP locate-doc tool,
consistent with DriveAddComment. Flags changed from --after-keyword /
--before-keyword to --selection-with-ellipsis + --before.
* fix(doc): show <locate_index> in dry-run create-block when selection is set
When --selection-with-ellipsis is provided, the create-block step in dry-run
now shows index: "<locate_index>" instead of "<children_len>" to accurately
reflect that the insertion position is computed from MCP locate-doc, not
appended to end.
* fix(doc): address CodeRabbit review on +media-insert selection feature
- Validate: reject blank/whitespace --selection-with-ellipsis unconditionally
so a mis-typed empty value cannot silently fall back to append-mode.
- Redact the raw selection string when logging to stderr and when emitting
error messages. --selection-with-ellipsis is copied verbatim from document
content and may contain confidential text; the new redactSelection helper
keeps a short prefix and rune count so operators can still identify the
failing selection.
- Harden the after/before mode tests: root children now have three entries
so the two modes land on different indices, and the tests decode the
create-block request body to assert the computed `index` actually reaches
the /children API. A regression that ignored --before would now fail.
- Harden the nested-block test so it exercises the fallback parent-walk:
the anchor is now two levels deep (blk_grandchild under blk_section_child
under blk_section), which forces the walk to fetch the intermediate block
via GET /blocks/{id} to discover the root-level ancestor.
* fix(doc): harden +media-insert selection UX on top of #335 (#577)
Follow-up to #335 review: closes a handful of UX and robustness gaps in
the new --selection-with-ellipsis flow.
- Flag description rewritten to make the "insert at the top-level
ancestor" semantics explicit — when the selection is inside a callout,
table cell, or nested list, media lands outside that container, not
inside. Also calls out the 'start...end' disambiguator.
- locate-doc is now called with limit=2 so an ambiguous selection
(same phrase in more than one block) surfaces a stderr warning
pointing at 'start...end', instead of silently picking the first
match. The first-match return behaviour is unchanged.
- When the anchor is nested below the root, locateInsertIndex now
logs a note to stderr naming the walk depth and the root-level
ancestor's insert index. Users don't have to guess why the image
landed outside the callout they were editing.
- maxDepth bumped 8 → 32 with a comment explaining the invariants:
`visited` is the real cycle guard, `maxDepth` is belt-and-suspenders.
32 comfortably exceeds real docx nesting depth so a deeply-nested
but well-formed anchor is no longer silently rejected.
- Comment added before the parent-walk loop noting why the API calls
are serial (each level's parent_id is only known after the previous
GET returns; can't be batched or parallelised).
Tests:
- TestLocateInsertIndexWarnsOnMultipleMatches: stubs two matches,
asserts the stderr warning names the ambiguity and mentions
'start...end', and that the first-match insert index is unchanged.
- TestLocateInsertIndexLogsNestedAnchor: anchor two levels below root,
asserts stderr carries the "nested … top-level ancestor" note.
- TestLocateInsertIndexCycleDetection: malformed parent chain with
blk_x.parent = blk_y and blk_y.parent = blk_x, neither reachable
from root. Registering a single GET /blocks/blk_y stub also bounds
the call count — a regression that broke `visited` tracking would
either hang or fail via httpmock's extra-call guard.
Co-authored-by: fangshuyu-768 <shuyufang768@outlook.com>
Adds 5 invariant-level tests on top of #469's transforms:
- TestFixExportedMarkdownIdempotent — f(f(x)) == f(x) across rich
fixtures (kitchen sink, CJK, nested containers). Protects the core
round-trip promise from future transform interactions that rewrite
their own output.
- TestFixExportedMarkdownPreservesFencedCodeByteForByte — packs every
pipeline-touching shape into a fence and asserts byte-identical output.
Code samples must never be silently rewritten by a formatting pass.
- TestFixExportedMarkdownPreservesCRLF — CRLF input preserves line
endings AND still triggers transforms. Windows-authored markdown
should not be silently LF-normalized.
- TestFixExportedMarkdownTransformInteractions — composition regressions:
nested-list + trailing-space bold, text→list transition, callout
containing list with emphasis, heading vs paragraph bold.
- TestNormalizeNestedListIndentationDocumentedSkips — locks in the
deliberate no-op branches (odd-space indent, blank-line loose-list
sibling, 4-space indented code block, parentless two-space) as an
explicit spec so future heuristic tweaks surface in the test diff.
All transforms, fixtures, and expectations are derived from the head of
PR #469. No production code changes.
Co-authored-by: fangshuyu-768 <shuyufang768@outlook.com>
* fix(doc): preserve round-trip formatting in fetch output
- trim leading spaces inside bold and italic emphasis exported by docs +fetch
- normalize nested list indentation to avoid flattening and literal text on re-import
- add regression tests for emphasis spacing and nested list indentation
* fix(doc): avoid false positives in markdown spacing fixes
- keep literal * x * and ** x ** text unchanged
- only normalize indented nested list markers when a parent list item exists
- add regression coverage for both CodeRabbit findings
* fix(doc): 修正嵌套列表缩进的空行误判
- 遇到空行时停止向上查找父级列表项,避免把 loose list sibling 误改成嵌套列表
- 避免把列表项中的四空格缩进代码块误改成 tab 缩进列表项
- 补充两个回归测试,并更新 fixBoldSpacing 注释使其与当前实现一致
* fix(doc): 修复 Markdown emphasis 空格回写
- 将 fixBoldSpacingLine 改为按星号 run 扫描,修复 ** hello **、* hello * 和同一行多个 italic span 的空格清理
- 保留 inline code、heading 和 *** hello** 这类近邻字面量,避免误改 emphasis nesting
fix: address coderabbit review comments on role-config docs
- Update `allow_edit` field description to reflect conditional default:
`true` when table perm is `edit`, `false` for `read_only` or explicit restriction
- Move `record_operations.delete` out of "默认关闭项" into new "默认开启项(条件性)"
section to accurately reflect it is default-included when `perm = edit`
- Add `view_rule.allow_edit` to "默认开启项(条件性)" section with same logic
Co-authored-by: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
* feat(sidecar): add sidecar proxy for sandbox credential isolation
Keep real secrets (app_secret, access_token) out of sandbox environments.
CLI instances inside sandboxes connect to a trusted sidecar process via
HTTP; the sidecar verifies HMAC-signed requests and injects real tokens
before forwarding to the Lark API.
Key components:
- `auth proxy` subcommand to start the sidecar server (build tag: authsidecar)
- Noop credential provider returns sentinel tokens in sidecar mode
- Transport interceptor rewrites requests to sidecar with HMAC signature
- Env provider yields to sidecar provider when AUTH_PROXY is set
- Supports both feishu and lark brand endpoints
* feat(sidecar): implement priority ordering for credential providers
* feat(sidecar): strip client-supplied auth headers and improve shutdown logging
* feat(sidecar): buffer request body to prevent HMAC mismatches on read errors
* feat(sidecar): fix CI
* refactor(sidecar): publish protocol package and move server to reference demo
The sidecar server is no longer shipped as a `lark-cli auth proxy`
subcommand. Instead, the CLI provides only the standard sidecar *client*
(via `-tags authsidecar`), while the wire-protocol utilities are exposed
as a public package for integrators to implement their own server.
Changes:
- Move `internal/sidecar/` → `sidecar/` so external integrators can
import HMAC signing, headers, sentinels and address validators.
- Remove `cmd/auth/proxy.go`, `proxy_stub.go`, `proxy_test.go` and the
conditional registration in `cmd/auth/auth.go`.
- Add `sidecar/server-demo/` — a reference server implementation behind
the `authsidecar_demo` build tag. It reuses the lark-cli credential
pipeline for local development; production integrators are expected
to replace the credential layer with their own secrets source.
- Update all internal imports from `internal/sidecar` to `sidecar`.
Rationale:
- Each integrator has different secrets management / HA / multi-tenant
requirements, so a one-size-fits-all server doesn't belong in the
shipped CLI.
- Keeping the client in-tree guarantees all sandbox-side code stays
protocol-compatible without a second repo to sync.
- The public `sidecar/` package pins the wire protocol as a stable
contract third-party servers must conform to.
Build matrix after this change:
- `go build` → standard CLI, no sidecar code
- `go build -tags authsidecar` → CLI + sidecar client
- `go build -tags authsidecar_demo \
./sidecar/server-demo/` → reference server binary
No production users are affected today because the server was not yet
released; existing sidecar-client users are unchanged.
* feat(sidecar): close 5 pre-release security gaps
- Server: enforce https-only target (no path/query/userinfo), pin
forwardURL to https:// — blocks cleartext token leak
- Protocol v1: canonical now covers version/identity/auth-header,
blocks identity-flip replay within drift window
- Client: ValidateProxyAddr requires loopback or same-host alias,
rejects userinfo and https (interceptor is http-only); cross-machine
is out of scope
- Build: non-authsidecar builds exit(2) when AUTH_PROXY is set,
preventing silent fallback to env credentials
- Demo: whitelist auth-header to Authorization / X-Lark-MCP-{UAT,TAT},
blocks token injection into Cookie / UA / X-Forwarded-For exfil paths
/style and /styles_batch_update require full "A1:A1" form and reject
single-cell shorthand "A1". +set-style was using normalizeSheetRange
(prefix-only) and +batch-set-style passed --data through unchanged,
so both failed with `wrong range` when callers supplied a single cell.
Switch +set-style to normalizePointRange, and walk each ranges[]
entry in +batch-set-style through normalizePointRange before sending.
Multi-cell spans pass through unchanged.
Implement +create-float-image, +update-float-image, +get-float-image,
+list-float-images, and +delete-float-image shortcuts wrapping the v3
spreadsheet float_image API. The create reference doc includes the
prerequisite media upload step with the correct parent_type
(sheet_image) to avoid common token mismatch errors.
The POST /contact/v3/users/basic_batch endpoint caps user_ids at 1~10
per request, but batchResolveByBasicContact was chunking by 50. When
user identity needed to resolve >10 unresolved sender names, the
single oversized request was rejected, causing the batch resolver to
bail out and leave sender names empty for the rest.
Lower batchSize to 10 and add a unit test that exercises 25 missing
IDs and asserts they are sent as 10 / 10 / 5.
Both notices recommend the same fix command: `lark-cli update`. The skills notice's `current` field is `""` when skills have never been synced (cold start) and a version string when synced for an older binary (drift).
## Pre-PR Checks (match CI gates)
1.`make unit-test`
@@ -59,7 +75,31 @@ The one rule to internalize: **every error message you write will be parsed by a
### Structured errors in commands
`RunE` functions must return `output.Errorf` / `output.ErrWithHint` — never bare `fmt.Errorf`. AI agents parse stderr as JSON; bare errors break this contract.
Command-facing failures must be typed `errs.*` errors — never the legacy `output.Err*` helpers and never a final bare `fmt.Errorf`. AI agents parse the stderr envelope's `type` / `subtype` / `param` / `hint` fields to decide their next action; the full taxonomy lives in `errs/ERROR_CONTRACT.md`.
Picking a constructor:
| Failure | Constructor |
|---------|-------------|
| User flag/arg fails validation | `errs.NewValidationError(errs.SubtypeInvalidArgument, ...).WithParam("--flag")` |
| Valid request, wrong system state | `errs.NewValidationError(errs.SubtypeFailedPrecondition, ...).WithHint(...)` |
| Lark API returned `code != 0` | `runtime.CallAPITyped` (shortcuts) / `errclass.BuildAPIError` (raw responses) — never hand-build |
| Network / transport failure | `errs.NewNetworkError(errs.SubtypeNetworkTransport, ...)` |
| Local file I/O failure | `errs.NewInternalError(errs.SubtypeFileIO, ...)` — validate the path first (`validate.SafeInputPath` / `SafeOutputPath`) and use `vfs.*` |
| Unclassified lower-layer error as final | `errs.NewInternalError(errs.SubtypeUnknown, ...).WithCause(err)` |
| Lower layer already returned a typed error | pass it through unchanged — re-wrapping downgrades its classification |
Signatures that are easy to guess wrong:
-`runtime.CallAPITyped(method, url string, params map[string]interface{}, data interface{}) (map[string]interface{}, error)` — it performs the HTTP request itself and classifies `code != 0` into a typed error; just return the error it gives you.
- Typed pass-through check: `if _, ok := errs.ProblemOf(err); ok { return err }` — `ProblemOf` returns `(*errs.Problem, bool)`, not a nilable pointer.
-`.WithParam` exists only on `*errs.ValidationError`. `InternalError` / `NetworkError` have no param field — file or endpoint context goes in the message or `.WithHint(...)`.
`forbidigo` + `lint/errscontract` reject the legacy `output.Err*` helpers, bare final `fmt.Errorf` / `errors.New`, and legacy envelope literals on migrated paths. Beyond what lint catches, three authoring conventions apply:
- Preserve the underlying error with `.WithCause(err)` so `errors.Is` / `errors.Unwrap` keep working.
-`param` names only the user input that actually failed. Recovery guidance goes in `.WithHint(...)`; machine-readable recovery fields (`missing_scopes`, `log_id`) carry server/system ground truth only — never caller-side guesses.
- Error-path tests assert typed metadata via `errs.ProblemOf` (`category` / `subtype` / `param`) and cause preservation, not message substrings alone.
- **im**: Clarify audio message opus requirement (#1271)
### Build
- **ci**: Add public content safeguards and reduce false positives
## [v1.0.57] - 2026-06-23
### Features
- **slides**: Add `+screenshot` to capture slide page images (or render a single `<slide>` XML snippet), returning the local file path instead of Base64 (#1358)
- **base**: Support record comments (#1043)
- **search**: Surface search API notices (#1413)
### Bug Fixes
- **mail**: Resolve folder/label filter once per `+triage list` call (#1512)
- **meta**: Backfill enum value descriptions from options (#1541)
- **slides**: Add `+export` shortcut to export slides (#988)
- **sidecar**: Support multi-client identity isolation in `server-demo` via per-client HMAC keys, preventing UAT cross-contamination when multiple CLI sandboxes share one sidecar (#934)
- **im**: Support Markdown image rendering in post content (#893)
### Bug Fixes
- **scope**: Add 22 new scope entries to scope priorities (#1050)
- **apps**: Refine `lark-apps` skill description and surface, document `index.html` / `--path` hard constraints (#1040)
## [v1.0.38] - 2026-05-22
### Features
- **apps**: Gate the Miaoda apps domain off on the Lark brand — the `apps` shortcut subtree returns a structured brand-restriction error, `auth login --domain apps` is rejected, `--domain all` skips it, and `spark:*` scopes are no longer requested (#1025)
The official [Lark/Feishu](https://www.larksuite.com/) CLI tool, maintained by the [larksuite](https://github.com/larksuite) team — built for humans and AI Agents. Covers core business domains including Messenger, Docs, Base, Sheets, Slides, Calendar, Mail, Tasks, Meetings, and more, with 200+ commands and 22 AI Agent [Skills](./skills/).
The official [Lark/Feishu](https://www.larksuite.com/) CLI tool, maintained by the [larksuite](https://github.com/larksuite) team — built for humans and AI Agents. Covers core business domains including Messenger, Docs, Base, Sheets, Slides, Calendar, Mail, Tasks, Meetings, Markdown, and more, with 200+ commands and 26 AI Agent [Skills](./skills/).
- **Agent-Native Design** — 22 structured [Skills](./skills/) out of the box, compatible with popular AI tools — Agents can operate Lark with zero extra setup
- **Wide Coverage** — 14 business domains, 200+ curated commands, 22 AI Agent [Skills](./skills/)
- **Agent-Native Design** — 24 structured [Skills](./skills/) out of the box, compatible with popular AI tools — Agents can operate Lark with zero extra setup
- **Wide Coverage** — 18 business domains, 200+ curated commands, 26 AI Agent [Skills](./skills/)
- **AI-Friendly & Optimized** — Every command is tested with real Agents, featuring concise parameters, smart defaults, and structured output to maximize Agent call success rates
- **Open Source, Zero Barriers** — MIT license, ready to use, just `npm install`
- **Up and Running in 3 Minutes** — One-click app creation, interactive login, from install to first API call in just 3 steps
@@ -24,10 +24,11 @@ The official [Lark/Feishu](https://www.larksuite.com/) CLI tool, maintained by t
| 📋 Project | Meegle — manage work items, schedules, and data via the standalone [meegle-cli](https://github.com/larksuite/meegle-cli) (install separately) |
| 🔗 Apps | Create Spark/Miaoda apps, publish HTML/static sites, run cloud generation, and manage access scope |
## Installation & Quick Start
@@ -60,11 +63,7 @@ Choose **one** of the following methods:
"strict mode is %q, user login is disabled in this profile",mode).
WithHint("if the user explicitly wants to switch to user identity, see `lark-cli config strict-mode --help` (confirm with the user before switching; switching does NOT require re-bind)")
}
opts.Ctx=cmd.Context()
ifrunF!=nil{
@@ -62,17 +68,26 @@ browser. Run it in the background and retrieve the verification URL from its out
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"--exclude requires --scope, --domain, or --recommend to be specified").WithParam("--exclude")
log("Note: this command blocks until authorization is complete. Run it in the background and retrieve the verification URL from its output.")
returnoutput.ErrValidation("please specify the scopes to authorize")
log("Note: this command blocks until authorization is complete. For non-streaming agent harnesses, use --no-wait --json, send the verification URL as the final message of the turn, then run --device-code in a later step after the user confirms authorization.")
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"please specify the scopes to authorize").WithParam("--scope")
}
}
finalScope:=opts.Scope
// Normalize --scope so users can pass either OAuth-standard space-separated
// values or the more natural comma-separated list. RFC 6749 §3.3 mandates
// space-delimited scopes in the wire request, so the device authorization
// endpoint rejects raw "a,b" strings as a single malformed scope.
finalScope:=normalizeScopeInput(opts.Scope)
// Resolve scopes from domain/permission filters
// Resolve scopes from domain/permission filters and merge with --scope.
// --scope, --domain, and --recommend combine additively so callers can,
// for example, request all `docs` scopes plus a few specific `drive`
// scopes in a single command.
iflen(selectedDomains)>0||opts.Recommend{
ifopts.Scope!=""{
returnoutput.ErrValidation("cannot use --scope together with --domain/--recommend")
"hint":fmt.Sprintf("Show verification_url to user, then immediately execute: lark-cli auth login --device-code %s (blocks until authorized or timeout). Do not instruct the user to run this command themselves.",authResp.DeviceCode),
"hint":"**MUST generate QR code AND display it:** You MUST call lark-cli auth qrcode to convert verification_url into a QR code. This is a required step, do NOT skip it. Prefer PNG QR code (--output); use ASCII (--ascii) only when the user explicitly requests it."+
"**CRITICAL: You MUST include the QR image in your response.** Generating the file alone is NOT enough—use image tags, inline images, or file attachments to display it."+
"**Display order:** Output the URL first, then place the QR code image below the URL."+
"**URL Output Rules:** Treat verification_url as an opaque string that cannot be modified. Do NOT URL-encode/decode or add spaces/punctuation."+
"For agent harnesses that only deliver final turn messages, make the QR code image (or URL) the final message of the turn and return control to the user; do not block on --device-code in the same turn. **Before ending the turn, tell the user to come back and notify you after completing authorization.**"+
"**After the user confirms authorization:** YOU must execute `lark-cli auth login --device-code <device_code>` yourself."+
"**Do NOT cache verification_url or device_code for future use.** Always run `lark-cli auth login --no-wait --json` fresh when authorization is needed.",
}
encoder:=json.NewEncoder(f.IOStreams.Out)
encoder.SetEscapeHTML(false)
iferr:=encoder.Encode(data);err!=nil{
returnoutput.Errorf(output.ExitInternal,"internal","failed to write JSON output: %v",err)
returnerrs.NewInternalError(errs.SubtypeSDKError,"failed to write JSON output: %v",err).WithCause(err)
}
returnnil
}
// Step 2: Show user code and verification URL
// Step 2: Show user code and verification URL.
// JSON mode embeds AgentTimeoutHint as a structured field so agents that
// capture stdout into a JSON parser see it without stream-mixing surprises.
// Text mode prints the hint to stderr only when running under a non-TTY
// (i.e. piped / agent harness), since humans reading a terminal don't need
OpenURL:"Open this URL in your browser to authenticate:\n\n",
WaitingAuth:"Waiting for user authorization...",
AuthSuccess:"Authorization completed, fetching user info and validating granted scopes...",
AgentTimeoutHint:"[AI agent] This command blocks for up to ~10 minutes while waiting for the user to authorize in their browser. Make sure your runner's timeout is >= 600s. If your harness or agent tool only delivers final turn messages, use \"lark-cli auth login --no-wait --json\" to get device_code and verification_url, present verification_url to the user exactly as the final message of this turn, then end the turn; after the user replies that they authorized, run \"lark-cli auth login --device-code <code>\" in a later step to resume polling. **Do NOT show the URL and then immediately block on --device-code in the same turn**, and do not retry with a short timeout; each restart invalidates the previous device code and makes the earlier authorization URL useless.**MUST generate QR code AND display it:** You MUST call lark-cli auth qrcode to convert verification_url into a QR code. This is a required step, do NOT skip it. Prefer PNG QR code (--output); use ASCII (--ascii) only when the user explicitly requests it.**CRITICAL: You MUST include the QR image in your response.** Generating the file alone is NOT enough—use image tags, inline images, or file attachments to display it.**Display order:** Output the URL first, then place the QR code image below the URL.**URL Output Rules:** Treat verification_url as an opaque string that cannot be modified. Do NOT URL-encode/decode or add spaces/punctuation.",
AuthSuccess:"Authorization confirmed, fetching user info and validating granted scopes...",
ScopeMismatch:"authorization result is abnormal: these requested scopes were not granted: %s",
ScopeHint:"The result above is the user's final confirmation for this authorization request. Do not retry continuously. Scopes may be not granted for various reasons, such as a scope being disabled. The specific reason has already been shown to the user on the authorization page. Run `lark-cli auth status` to inspect all scopes currently granted to the account.",
RequestedScopes:" Requested scopes: %s\n",
NewlyGrantedScopes:" Newly granted scopes: %s\n",
MissingScopes:" Not granted scopes: %s\n",
NoScopes:"(none)",
StatusHint:"Run `lark-cli auth status` to inspect all scopes currently granted to the account.",
@@ -114,8 +116,9 @@ var loginMsgEn = &loginMsg{
HintFooter:" lark-cli auth login --help",
}
funcgetLoginMsg(langstring)*loginMsg{
iflang=="en"{
// getLoginMsg returns the login message bundle for the given language.
Long:`Generate a QR code image or ASCII representation for a verification URL.
This command is designed for AI agents to generate QR codes for OAuth authorization URLs.
For PNG output, the --output flag is required to specify the output file path (must be a relative path within the current directory).
For ASCII output, the result is printed to stdout with fixed size.`,
Args:cobra.ExactArgs(1),
RunE:func(cmd*cobra.Command,args[]string)error{
opts.URL=args[0]
opts.Ctx=cmd.Context()
ifrunF!=nil{
returnrunF(opts)
}
returnrunQRCode(opts)
},
}
cmd.Flags().IntVar(&opts.Size,"size",256,"Size of the QR code image in pixels (default: 256, for PNG mode only)")
cmd.Flags().BoolVar(&opts.ASCII,"ascii",false,"Output ASCII QR code to stdout")
cmd.Flags().StringVarP(&opts.Output,"output","o","","Output file path for PNG image (relative path within current directory, required for non-ASCII mode)")
returncmd
}
// runQRCode executes the auth qrcode command.
funcrunQRCode(opts*QRCodeOptions)error{
ifopts.URL==""{
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"url is required").WithParam("--url")
}
ifopts.ASCII{
varoutio.Writer=os.Stdout
ifopts.Factory!=nil{
out=opts.Factory.IOStreams.Out
}
returngenerateASCIIQRCode(opts.URL,out)
}
ifopts.Output==""{
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"output file path is required for PNG mode. Use --output or -o flag to specify the output file path.").WithParam("--output")
}
ifopts.Size<32{
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"size must be at least 32, got %d",opts.Size).WithParam("--size")
}
ifopts.Size>1024{
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"size must be at most 1024, got %d",opts.Size).WithParam("--size")
"hint":"You MUST include the QR image in your response. Generating the file alone is NOT enough—use image tags, inline images, or file attachments to display it.",
}
varoutio.Writer=os.Stdout
ifopts.Factory!=nil{
out=opts.Factory.IOStreams.Out
}
encoder:=json.NewEncoder(out)
encoder.SetEscapeHTML(false)
iferr:=encoder.Encode(result);err!=nil{
returnerrs.NewInternalError(errs.SubtypeSDKError,"failed to write output: %v",err).WithCause(err)
}
returnnil
}
// generateImageQRCode encodes the URL as a PNG QR code and writes it to outputPath.
result["note"]="User identity is "+identitydiag.StatusMessage(d.User.Status)+"; bot identity is ready for bot/tenant API calls. Run `lark-cli auth login` to enable user identity."
# Interactive (terminal user) — TUI prompts for everything:
lark-cli config bind`,
RunE:func(cmd*cobra.Command,args[]string)error{
opts.langExplicit=cmd.Flags().Changed("lang")
ifrunF!=nil{
returnrunF(opts)
}
returnconfigBindRun(opts)
},
}
cmd.Flags().StringVar(&opts.Source,"source","","Agent source to bind from (openclaw|hermes|lark-channel); auto-detected from env signals when omitted")
cmd.Flags().StringVar(&opts.AppID,"app-id","","App ID to bind (required for OpenClaw multi-account)")
cmd.Flags().StringVar(&opts.Identity,"identity","","identity preset (bot-only|user-default); defaults to bot-only in flag mode (safer: no impersonation)")
cmd.Flags().BoolVar(&opts.Force,"force",false,"confirm a risky transition (currently: bot-only → user-default identity change in flag mode)")
cmd.Flags().StringVar(&opts.Lang,"lang","","language preference (e.g. zh or zh_cn)")
cmdutil.SetRisk(cmd,"write")
returncmd
}
// configBindRun is the top-level orchestrator. Each step delegates to a named
// helper whose signature declares its contract; the body reads as the shape of
// the bind flow itself, not its mechanics.
funcconfigBindRun(opts*BindOptions)error{
iferr:=validateBindFlags(opts);err!=nil{
returnerr
}
// Decide TUI-vs-flag mode exactly once; every downstream branch reads
// opts.IsTUI instead of re-checking IOStreams.IsTerminal.
SelectSourceDesc:"lark-cli will read your %s app credentials from the selected Agent and apply them automatically.",
SourceOpenClaw:"OpenClaw — config: %s",
SourceHermes:"Hermes — config: %s",
SourceLarkChannel:"Lark Channel — config: %s",
// Args order (source, brand) matches the Chinese template; %[N]s lets the
// English reading order differ while the caller passes args in one order.
SelectAccount:"Multiple %[2]s apps configured in %[1]s — select one to continue.",
ConflictTitle:"Existing configuration found",
ConflictDesc:"lark-cli is already set up for %q:\n App ID: %s\n Brand: %s\n Config: %s",
ConflictForce:"Update config",
ConflictCancel:"Keep current config",
ConflictCancelled:"Current config kept. No changes made.",
MessageBotOnly:"Bound app %s to %s. The %s app (bot) identity is ready — you can now continue with the user's request.",
MessageUserDefault:"Bound app %s to %s. Next, in this %s chat, run `lark-cli auth login --recommend`. The command prints the verification URL to stderr and then blocks until the user authorizes it; relay the URL to the user so they can approve it in their own browser (do not call browser_navigate or any tool that opens a browser yourself — your browser is sandboxed and cannot complete the authorization). The command returns automatically once authorization completes.",
SelectIdentity:"How should the AI work with you?",
IdentityBotOnly:"As bot",
IdentityUserDefault:"As you",
IdentityBotOnlyDesc:"Works under its own identity in %s. Best for group chats, team notifications, and shared documents.",
IdentityUserDefaultDesc:"Works under your identity in %s, managing docs, messages, calendar, and more on your behalf. Personal use only.\n"+
"⚠️ Don't share this bot with others or add it to group chats. It has access to your personal %s data.",
BindSuccessHeader:"All set! lark-cli is now ready to use in %s.",
BindSuccessNotice:"Note: This is a one-time sync. To re-sync future changes, run `lark-cli config bind`",
IdentityEscalationMessage:"you are switching from bot-only to user-default — the AI will then act under your Feishu identity for all operations (docs, messages, calendar, etc.). ⚠️ Don't share this bot with others or add it to group chats. It has access to your personal Feishu data.",
IdentityEscalationHint:"if the user confirms the switch, re-run with --force: `lark-cli config bind --identity user-default --force`",
LangPreferenceSet:"Language preference set to: %s",
}
// getBindMsg picks the zh/en TUI bundle; non-English falls back to zh.
funcgetBindMsg(langi18n.Lang)*bindMsg{
iflang.IsEnglish(){
returnbindMsgEn
}
returnbindMsgZh
}
// brandDisplay returns the UI-friendly product name for the given brand
// identifier and display language. "lark" maps to "Lark" in both zh and en.
// "feishu" (or empty / unknown) maps to "飞书" in zh and "Feishu" in en —
// this is the safe default when the brand hasn't been resolved yet (for
// example, on the pre-binding source-selection screen).
For AI agents: use --new to create a new app. The command blocks until the user
completes setup in the browser. Run it in the background and retrieve the
verification URL from its output.`,
verification URL from its output.
Inside an Agent context (OPENCLAW_HOME / HERMES_HOME set) this command
refuses by default — use 'lark-cli config bind' to bind to the Agent's
existing app instead of creating a parallel one. Pass --force-init only
if the user explicitly wants a separate app inside the Agent workspace.`,
RunE:func(cmd*cobra.Command,args[]string)error{
opts.Ctx=cmd.Context()
opts.langExplicit=cmd.Flags().Changed("lang")
iferr:=validateInitLang(opts);err!=nil{
returnerr
}
iferr:=guardAgentWorkspace(opts);err!=nil{
returnerr
}
ifrunF!=nil{
returnrunF(opts)
}
@@ -61,12 +84,52 @@ verification URL from its output.`,
cmd.Flags().StringVar(&opts.AppID,"app-id","","App ID (non-interactive)")
cmd.Flags().BoolVar(&opts.AppSecretStdin,"app-secret-stdin",false,"Read App Secret from stdin to avoid process list exposure")
cmd.Flags().StringVar(&opts.Brand,"brand","feishu","feishu or lark (non-interactive, default feishu)")
cmd.Flags().StringVar(&opts.Lang,"lang","zh","language for interactive prompts (zh or en)")
cmd.Flags().StringVar(&opts.Lang,"lang","","language preference (e.g. zh or zh_cn)")
cmd.Flags().StringVar(&opts.ProfileName,"name","","create or update a named profile (append instead of replace)")
cmd.Flags().BoolVar(&opts.ForceInit,"force-init",false,"allow init inside an Agent workspace (OPENCLAW_HOME / HERMES_HOME); use config bind instead unless you really want a separate app")
cmdutil.SetRisk(cmd,"write")
returncmd
}
// printLangPreferenceConfirmation echoes the set preference to stderr, only
"config init is refused inside %s context (would create a parallel app and shadow the existing %s binding)",ws.Display(),ws.Display()).
WithHint("see `lark-cli config bind --help` to bind lark-cli to the Agent's existing app instead. Pass --force-init only if the user explicitly wants a separate app in this workspace.")
}
// hasAnyNonInteractiveFlag returns true if any non-interactive flag is set.
// Non-terminal: cannot run interactive mode, guide user to --new
if!f.IOStreams.IsTerminal{
returnoutput.ErrValidation("config init requires a terminal for interactive mode. Run with --new to create a new app:\n lark-cli config init --new\nThis command blocks until setup is complete and outputs a verification URL. Run it in the background, then retrieve the URL from its output.")
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"config init requires a terminal for interactive mode. Run with --new to create a new app:\n lark-cli config init --new\nThis command blocks until setup is complete and outputs a verification URL. Run it in the background, then retrieve the URL from its output.")
WithHint("This command must be run from an interactive macOS session (e.g. Terminal.app or iTerm) where the system Keychain is reachable. Running it from inside a sandbox / automation context that blocks Keychain access cannot succeed by design.").
WithCause(err)
}
switchresult{
casekeychain.DowngradeAlreadyDone:
output.PrintSuccess(f.IOStreams.ErrOut,fmt.Sprintf("keychain already downgraded; subsequent operations read from %s",keyPath))
casekeychain.DowngradeUsedKeychainKey:
output.PrintSuccess(f.IOStreams.ErrOut,fmt.Sprintf("downgraded: copied master key from system Keychain to %s. Subsequent operations will read from file, bypassing the OS Keychain (useful inside sandboxes like Codex).",keyPath))
casekeychain.DowngradeCreatedNewKey:
output.PrintSuccess(f.IOStreams.ErrOut,fmt.Sprintf("system Keychain was empty; generated a new master key and wrote it to %s. The OS Keychain was not modified.",keyPath))
Short:"Downgrade keychain storage to a local file (macOS only)",
Long:`Downgrade keychain storage to a local file. This subcommand is only supported on macOS; on this platform the keychain layer already uses local files.`,
RunE:func(cmd*cobra.Command,args[]string)error{
returnerrs.NewValidationError(errs.SubtypeInvalidArgument,"keychain-downgrade is only supported on macOS")
constrealisticPermError=`API GET /open-apis/application/v6/applications/cli_XXXXXXXXXXXXXXXX/app_versions?lang=zh_cn&page_size=2 returned 400: {"code":99991672,"msg":"Access denied. One of the following scopes is required: [application:application:self_manage, application:application.app_version:readonly].应用尚未开通所需的应用身份权限:[application:application:self_manage, application:application.app_version:readonly],点击链接申请并开通任一权限即可:https://open.feishu.cn/app/cli_XXXXXXXXXXXXXXXX/auth?q=application:application:self_manage,application:application.app_version:readonly&op_from=openapi&token_type=tenant","error":{"message":"Refer to the documentation...","log_id":"20260421101203E2A5F141245B6F43B3A6"}}`
cmd.Flags().StringVar(&o.jqExpr,"jq","","JQ expression to filter output")
cmd.Flags().BoolVar(&o.quiet,"quiet",false,"Suppress informational messages on stderr")
cmd.Flags().StringVar(&o.outputDir,"output-dir","","Write each event as a file in this directory (relative paths only; absolute paths and ~ are rejected to prevent path traversal)")
cmd.Flags().IntVar(&o.maxEvents,"max-events",0,"Exit after N successful emits (0 = unlimited). Multi-worker EventKeys may emit up to workers-1 past N before all workers stop. Bounded runs ignore stdin EOF.")
cmd.Flags().DurationVar(&o.timeout,"timeout",0,"Exit after DURATION (e.g. 30s, 2m). 0 = no timeout. Timeout is a normal exit (code 0; stderr 'reason: timeout'). Bounded runs ignore stdin EOF.")
cmd.Flags().String("as","auto","identity type: user | bot | auto (must match EventKey's declared AuthTypes)")
"run `lark-cli auth login --scope \"%s\"` in the background. It blocks and outputs a verification URL — retrieve the URL and open it in a browser to complete login.",
strings.Join(missing," "))
}
// preflightEventTypes verifies every RequiredConsoleEvents entry is subscribed
// in the app's console 底账 — published app_versions for event subscriptions,
// application/get subscribed_callbacks for callback subscriptions.
// resolveSchemaJSON returns the final JSON Schema for an EventKey (reflected base, V2-wrapped for Native, overlay applied); orphans lists unresolved FieldOverrides pointers.
Long:"Display detailed information about an EventKey including type, events, parameters, and response schema. Use --json for machine-readable output.",
Args:cobra.ExactArgs(1),
RunE:func(cmd*cobra.Command,args[]string)error{
returnrunSchema(f,args[0],asJSON)
},
}
cmd.Flags().BoolVar(&asJSON,"json",false,"Emit the EventKey definition + resolved schema as JSON (for AI / scripts)")
// ignoreBrokenPipe is a no-op on Windows (no SIGPIPE; closed-pipe writes return ERROR_BROKEN_PIPE directly).
funcignoreBrokenPipe(){}
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.