opt(round-003): 复测 效果+0.200 token+17 耗时-1612ms → 采纳

.github/workflows/semantic-review.yml

@@ -47,13 +47,10 @@ jobs:
               throw new Error(`ambiguous workflow_run pull request bindings: ${runPRs.length}`);
             }
             let prNumber = Number(runPRs[0]?.number || 0);
-            const eventBaseSha = runPRs[0]?.base?.sha || "";
+            let eventBaseSha = runPRs[0]?.base?.sha || "";
             const eventHeadSha = runPRs[0]?.head?.sha || "";
-            const targetHeadSha = run.head_sha;
+            const targetHeadSha = eventHeadSha || run.head_sha;
             if (!/^[a-f0-9]{40}$/i.test(targetHeadSha)) throw new Error("invalid PR head sha");
-            if (eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
-              core.notice("PR quality summary using workflow_run head_sha because workflow_run pull request head differs from the CI run head");
-            }
 
             const factsArtifactPattern = /^quality-gate-facts-([a-f0-9]{40})-([a-f0-9]{40})$/i;
             const { data: artifactData } = await github.rest.actions.listWorkflowRunArtifacts({
@@ -74,11 +71,11 @@ jobs:
               if (artifactHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
                 artifactError = "facts artifact head sha does not match verified PR head sha";
                 factsArtifactName = "";
+              } else if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
+                artifactError = "facts artifact base sha does not match workflow_run pull request base sha";
+                factsArtifactName = "";
               } else {
                 artifactBaseSha = parsedBaseSha;
-                if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
-                  core.notice("PR quality summary using facts artifact base because workflow_run pull request base differs from the CI facts artifact base");
-                }
               }
             }
             if (!prNumber) {
@@ -126,7 +123,7 @@ jobs:
               core.setOutput("stale", "true");
               return;
             }
-            const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha;
+            const baseSha = eventBaseSha || artifactBaseSha || pr.base.sha;
             if (!/^[a-f0-9]{40}$/i.test(baseSha)) throw new Error("invalid PR base sha");
             if ((eventBaseSha || artifactBaseSha) && pr.base.sha !== baseSha) {
               core.notice("PR quality summary skipped: workflow_run is stale for this PR base");
@@ -258,13 +255,10 @@ jobs:
               throw new Error(`ambiguous workflow_run pull request bindings: ${runPRs.length}`);
             }
             let prNumber = Number(runPRs[0]?.number || 0);
-            const eventBaseSha = runPRs[0]?.base?.sha || "";
+            let eventBaseSha = runPRs[0]?.base?.sha || "";
             const eventHeadSha = runPRs[0]?.head?.sha || "";
-            const targetHeadSha = run.head_sha;
+            const targetHeadSha = eventHeadSha || run.head_sha;
             if (!/^[a-f0-9]{40}$/i.test(targetHeadSha)) throw new Error("invalid PR head sha");
-            if (eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
-              core.notice("semantic review using workflow_run head_sha because workflow_run pull request head differs from the CI run head");
-            }
 
             const factsArtifactPattern = /^quality-gate-facts-([a-f0-9]{40})-([a-f0-9]{40})$/i;
             const { data: artifactData } = await github.rest.actions.listWorkflowRunArtifacts({
@@ -285,11 +279,11 @@ jobs:
               if (artifactHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
                 artifactError = "facts artifact head sha does not match verified PR head sha";
                 factsArtifactName = "";
+              } else if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
+                artifactError = "facts artifact base sha does not match workflow_run pull request base sha";
+                factsArtifactName = "";
               } else {
                 artifactBaseSha = parsedBaseSha;
-                if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
-                  core.notice("semantic review using facts artifact base because workflow_run pull request base differs from the CI facts artifact base");
-                }
               }
             }
             if (!prNumber) {
@@ -337,7 +331,7 @@ jobs:
               core.setOutput("stale", "true");
               return;
             }
-            const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha;
+            const baseSha = eventBaseSha || artifactBaseSha || pr.base.sha;
             if (!/^[a-f0-9]{40}$/i.test(baseSha)) throw new Error("invalid PR base sha");
             if ((eventBaseSha || artifactBaseSha) && pr.base.sha !== baseSha) {
               core.notice("semantic review skipped: workflow_run is stale for this PR base");

.gitignore

@@ -7,11 +7,6 @@ bin/
 # Node
 node_modules/
 
-# Python (skill-bundled helper scripts)
-__pycache__/
-*.py[cod]
-*$py.class
-
 
 # OS
 .DS_Store

CHANGELOG.md

@@ -2,30 +2,6 @@
 
 All notable changes to this project will be documented in this file.
 
-## [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)
-- **cli**: Add missing CLI headers for git credential helper (#1539)
-
-### Documentation
-
-- **doc**: Refine rich block, path, and block ID guidance (#1508)
-- **mail**: Trim lark-mail skill context (#1527)
-- **drive**: Add permission governance workflow guidance (#1292)
-
-### Build
-
-- **ci**: Bind semantic review to workflow run head (#1551)
-
 ## [v1.0.56] - 2026-06-18
 
 ### Features
@@ -1236,7 +1212,6 @@ Bundled AI agent skills for intelligent assistance:
 - Bilingual documentation (English & Chinese).
 - CI/CD pipelines: linting, testing, coverage reporting, and automated releases.
 
-[v1.0.57]: https://github.com/larksuite/cli/releases/tag/v1.0.57
 [v1.0.56]: https://github.com/larksuite/cli/releases/tag/v1.0.56
 [v1.0.55]: https://github.com/larksuite/cli/releases/tag/v1.0.55
 [v1.0.54]: https://github.com/larksuite/cli/releases/tag/v1.0.54

cmd/event/list_test.go

@@ -26,7 +26,6 @@ func TestRunList_TextOutput(t *testing.T) {
 		"KEY", "AUTH", "PARAMS", "DESCRIPTION",
 		"im.message.receive_v1",
 		"im.message.message_read_v1",
-		"task.task.update_user_access_v2",
 	} {
 		if !strings.Contains(out, want) {
 			t.Errorf("list output missing %q; full output:\n%s", want, out)
@@ -56,17 +55,4 @@ func TestRunList_JSONOutput(t *testing.T) {
 			}
 		}
 	}
-
-	var foundTask bool
-	for _, row := range rows {
-		if row["key"] == "task.task.update_user_access_v2" {
-			foundTask = true
-			if row["single_consumer"] != true {
-				t.Errorf("task row single_consumer = %v, want true", row["single_consumer"])
-			}
-		}
-	}
-	if !foundTask {
-		t.Fatal("event list JSON missing task.task.update_user_access_v2")
-	}
 }

cmd/event/schema_test.go

@@ -96,34 +96,6 @@ func TestRunSchema_JSONOutput(t *testing.T) {
 	}
 }
 
-func TestRunSchema_TaskUpdateUserAccessJSON(t *testing.T) {
-	f, stdout, _, _ := cmdutil.TestFactory(t, &core.CliConfig{AppID: "test"})
-
-	if err := runSchema(f, "task.task.update_user_access_v2", true); err != nil {
-		t.Fatalf("runSchema json: %v", err)
-	}
-
-	var payload map[string]interface{}
-	if err := json.Unmarshal(stdout.Bytes(), &payload); err != nil {
-		t.Fatalf("output is not valid JSON: %v\n%s", err, stdout.String())
-	}
-	if payload["jq_root_path"] != ".event" {
-		t.Errorf("jq_root_path = %v, want .event", payload["jq_root_path"])
-	}
-	if payload["single_consumer"] != true {
-		t.Errorf("single_consumer = %v, want true", payload["single_consumer"])
-	}
-	resolved := payload["resolved_output_schema"].(map[string]interface{})
-	props := resolved["properties"].(map[string]interface{})
-	eventProps := props["event"].(map[string]interface{})["properties"].(map[string]interface{})
-	if got := eventProps["task_guid"].(map[string]interface{})["format"]; got != "task_guid" {
-		t.Errorf("task_guid format = %v, want task_guid", got)
-	}
-	if _, ok := eventProps["event_types"].(map[string]interface{})["items"].(map[string]interface{})["enum"]; !ok {
-		t.Fatalf("event_types enum missing in schema: %#v", eventProps["event_types"])
-	}
-}
-
 func TestSchema_RendersSubscriptionKeyMarker(t *testing.T) {
 	const syntheticKey = "test.evt_sub"
 	t.Cleanup(func() { eventlib.UnregisterKeyForTest(syntheticKey) })

events/im/card_action.go

@@ -1,132 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package im
-
-import (
-	"context"
-	"encoding/json"
-	"strings"
-
-	"github.com/larksuite/cli/internal/event"
-)
-
-// CardActionTriggerOutput is the flattened shape for card.action.trigger.
-type CardActionTriggerOutput struct {
-	Type        string `json:"type"                    desc:"Event type; always card.action.trigger"`
-	EventID     string `json:"event_id,omitempty"      desc:"Globally unique event ID"`
-	Timestamp   string `json:"timestamp,omitempty"     desc:"Event delivery time (ms timestamp string)" kind:"timestamp_ms"`
-	OperatorID  string `json:"operator_id,omitempty"   desc:"Operator open_id"                          kind:"open_id"`
-	MessageID   string `json:"message_id,omitempty"    desc:"Message ID of the card"                    kind:"message_id"`
-	ChatID      string `json:"chat_id,omitempty"       desc:"Chat ID"                                   kind:"chat_id"`
-	Host        string `json:"host,omitempty"          desc:"Host type: im_message / im_top_notice"`
-	Token       string `json:"token,omitempty"         desc:"Token for delay card update (valid 30 min, max 2 updates)"`
-	ActionTag   string `json:"action_tag,omitempty"    desc:"Triggered element type: button/select_static/input/checker/etc"`
-	ActionValue string `json:"action_value,omitempty"  desc:"Developer-defined action value as JSON string"`
-	ActionName  string `json:"action_name,omitempty"   desc:"Element name attribute"`
-	FormValue   string `json:"form_value,omitempty"    desc:"Form submission values as JSON string (only on form submit)"`
-	InputValue  string `json:"input_value,omitempty"   desc:"Input field value (only for input elements)"`
-	Option      string `json:"option,omitempty"        desc:"Selected option value (for single-select dropdown)"`
-	Options     string `json:"options,omitempty"       desc:"Selected options, comma-separated (for multi-select)"`
-	Checked     bool   `json:"checked"                 desc:"Checkbox state (for checkbox elements)"`
-	Timezone    string `json:"timezone,omitempty"      desc:"User timezone for date/time picker interactions"`
-	CardContent string `json:"card_content,omitempty"  desc:"Original card JSON content (body.content) auto-fetched via message get API at consume time using message_id; empty if message_id absent or fetch fails"`
-}
-
-func processCardAction(ctx context.Context, rt event.APIClient, raw *event.RawEvent, _ map[string]string) (json.RawMessage, error) {
-	var envelope struct {
-		Header struct {
-			EventID    string `json:"event_id"`
-			EventType  string `json:"event_type"`
-			CreateTime string `json:"create_time"`
-		} `json:"header"`
-		Event struct {
-			Operator struct {
-				OpenID string `json:"open_id"`
-			} `json:"operator"`
-			Token  string `json:"token"`
-			Host   string `json:"host"`
-			Action struct {
-				Tag        string                 `json:"tag"`
-				Value      map[string]interface{} `json:"value"`
-				Name       string                 `json:"name"`
-				FormValue  map[string]interface{} `json:"form_value"`
-				InputValue string                 `json:"input_value"`
-				Option     string                 `json:"option"`
-				Options    []string               `json:"options"`
-				Checked    bool                   `json:"checked"`
-				Timezone   string                 `json:"timezone"`
-			} `json:"action"`
-			Context struct {
-				OpenMessageID string `json:"open_message_id"`
-				OpenChatID    string `json:"open_chat_id"`
-			} `json:"context"`
-		} `json:"event"`
-	}
-	if err := json.Unmarshal(raw.Payload, &envelope); err != nil {
-		return raw.Payload, nil //nolint:nilerr // passthrough on malformed payload
-	}
-
-	actionValue := marshalToString(envelope.Event.Action.Value)
-	formValue := marshalToString(envelope.Event.Action.FormValue)
-	options := strings.Join(envelope.Event.Action.Options, ",")
-
-	out := &CardActionTriggerOutput{
-		Type:        envelope.Header.EventType,
-		EventID:     envelope.Header.EventID,
-		Timestamp:   envelope.Header.CreateTime,
-		OperatorID:  envelope.Event.Operator.OpenID,
-		MessageID:   envelope.Event.Context.OpenMessageID,
-		ChatID:      envelope.Event.Context.OpenChatID,
-		Host:        envelope.Event.Host,
-		Token:       envelope.Event.Token,
-		ActionTag:   envelope.Event.Action.Tag,
-		ActionValue: actionValue,
-		ActionName:  envelope.Event.Action.Name,
-		FormValue:   formValue,
-		InputValue:  envelope.Event.Action.InputValue,
-		Option:      envelope.Event.Action.Option,
-		Options:     options,
-		Checked:     envelope.Event.Action.Checked,
-		Timezone:    envelope.Event.Action.Timezone,
-	}
-
-	if out.MessageID != "" && rt != nil {
-		out.CardContent = fetchCardUserDSL(ctx, rt, out.MessageID)
-	}
-
-	return json.Marshal(out)
-}
-
-// fetchCardUserDSL gets the card message content via message get API.
-// Returns empty string on any failure — never blocks event consumption.
-func fetchCardUserDSL(ctx context.Context, rt event.APIClient, messageID string) string {
-	path := "/open-apis/im/v1/messages/" + messageID + "?card_msg_content_type=user_card_content"
-	resp, err := rt.CallAPI(ctx, "GET", path, nil)
-	if err != nil {
-		return ""
-	}
-	var result struct {
-		Code int    `json:"code"`
-		Msg  string `json:"msg"`
-		Data struct {
-			Items []struct {
-				Body struct {
-					Content string `json:"content"`
-				} `json:"body"`
-			} `json:"items"`
-		} `json:"data"`
-	}
-	if json.Unmarshal(resp, &result) != nil || result.Code != 0 || len(result.Data.Items) == 0 {
-		return ""
-	}
-	return result.Data.Items[0].Body.Content
-}
-
-func marshalToString(m map[string]interface{}) string {
-	if len(m) == 0 {
-		return ""
-	}
-	b, _ := json.Marshal(m)
-	return string(b)
-}

events/im/card_action_test.go

@@ -1,432 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package im
-
-import (
-	"context"
-	"encoding/json"
-	"testing"
-	"time"
-
-	"github.com/larksuite/cli/internal/event"
-)
-
-func TestCardActionTriggerRegistered(t *testing.T) {
-	def, ok := event.Lookup("card.action.trigger")
-	if !ok {
-		t.Fatal("card.action.trigger should be registered via Keys()")
-	}
-	if def.Schema.Custom == nil {
-		t.Error("card.action.trigger must set Schema.Custom")
-	}
-	if def.Process == nil {
-		t.Error("card.action.trigger must set Process")
-	}
-	if len(def.Scopes) == 0 {
-		t.Error("Scopes must not be empty")
-	}
-}
-
-func TestProcessCardAction_Button(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_btn_001",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469273"
-		},
-		"event": {
-			"operator": {"open_id": "ou_operator"},
-			"token": "c-token-btn",
-			"host": "im_message",
-			"action": {
-				"tag": "button",
-				"value": {"key": "approve"},
-				"name": "approve_btn",
-				"form_value": {},
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_msg_001",
-				"open_chat_id": "oc_chat_001"
-			}
-		}
-	}`
-	out := runCardAction(t, payload, nil)
-
-	if out.Type != "card.action.trigger" {
-		t.Errorf("Type = %q, want card.action.trigger", out.Type)
-	}
-	if out.EventID != "ev_btn_001" {
-		t.Errorf("EventID = %q", out.EventID)
-	}
-	if out.OperatorID != "ou_operator" {
-		t.Errorf("OperatorID = %q", out.OperatorID)
-	}
-	if out.ActionTag != "button" {
-		t.Errorf("ActionTag = %q, want button", out.ActionTag)
-	}
-	if out.ActionValue != `{"key":"approve"}` {
-		t.Errorf("ActionValue = %q", out.ActionValue)
-	}
-	if out.ActionName != "approve_btn" {
-		t.Errorf("ActionName = %q", out.ActionName)
-	}
-	if out.Token != "c-token-btn" {
-		t.Errorf("Token = %q", out.Token)
-	}
-	if out.MessageID != "om_msg_001" {
-		t.Errorf("MessageID = %q", out.MessageID)
-	}
-	if out.ChatID != "oc_chat_001" {
-		t.Errorf("ChatID = %q", out.ChatID)
-	}
-	if out.Host != "im_message" {
-		t.Errorf("Host = %q", out.Host)
-	}
-	if out.Timestamp != "1776409469273" {
-		t.Errorf("Timestamp = %q", out.Timestamp)
-	}
-}
-
-func TestProcessCardAction_FormSubmit(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_form_001",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469274"
-		},
-		"event": {
-			"operator": {"open_id": "ou_form_user"},
-			"token": "c-token-form",
-			"host": "im_message",
-			"action": {
-				"tag": "button",
-				"value": {},
-				"name": "submit_btn",
-				"form_value": {"name": "test-user", "reason": "testing"},
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_form_001",
-				"open_chat_id": "oc_chat_002"
-			}
-		}
-	}`
-	out := runCardAction(t, payload, nil)
-
-	if out.FormValue != `{"name":"test-user","reason":"testing"}` {
-		t.Errorf("FormValue = %q", out.FormValue)
-	}
-	if out.ActionTag != "button" {
-		t.Errorf("ActionTag = %q, want button", out.ActionTag)
-	}
-}
-
-func TestProcessCardAction_MultiSelect(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_ms_001",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469275"
-		},
-		"event": {
-			"operator": {"open_id": "ou_ms_user"},
-			"token": "c-token-ms",
-			"host": "im_message",
-			"action": {
-				"tag": "multi_select_static",
-				"value": {},
-				"name": "multi_select",
-				"options": ["opt_1", "opt_3"],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_ms_001",
-				"open_chat_id": "oc_chat_003"
-			}
-		}
-	}`
-	out := runCardAction(t, payload, nil)
-
-	if out.Options != "opt_1,opt_3" {
-		t.Errorf("Options = %q, want opt_1,opt_3", out.Options)
-	}
-	if out.ActionTag != "multi_select_static" {
-		t.Errorf("ActionTag = %q", out.ActionTag)
-	}
-}
-
-func TestProcessCardAction_Input(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_input_001",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469276"
-		},
-		"event": {
-			"operator": {"open_id": "ou_input_user"},
-			"token": "c-token-input",
-			"host": "im_message",
-			"action": {
-				"tag": "input",
-				"value": {},
-				"name": "text_input",
-				"input_value": "hello world",
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_input_001",
-				"open_chat_id": "oc_chat_004"
-			}
-		}
-	}`
-	out := runCardAction(t, payload, nil)
-
-	if out.InputValue != "hello world" {
-		t.Errorf("InputValue = %q", out.InputValue)
-	}
-	if out.ActionTag != "input" {
-		t.Errorf("ActionTag = %q", out.ActionTag)
-	}
-}
-
-func TestProcessCardAction_DatePicker(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_date_001",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469277"
-		},
-		"event": {
-			"operator": {"open_id": "ou_date_user"},
-			"token": "c-token-date",
-			"host": "im_message",
-			"action": {
-				"tag": "date_picker",
-				"value": {},
-				"name": "date_selector",
-				"option": "2024-04-01 +0800",
-				"timezone": "Asia/Shanghai",
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_date_001",
-				"open_chat_id": "oc_chat_005"
-			}
-		}
-	}`
-	out := runCardAction(t, payload, nil)
-
-	if out.Option != "2024-04-01 +0800" {
-		t.Errorf("Option = %q", out.Option)
-	}
-	if out.Timezone != "Asia/Shanghai" {
-		t.Errorf("Timezone = %q", out.Timezone)
-	}
-}
-
-func TestProcessCardAction_MalformedPayload(t *testing.T) {
-	raw := &event.RawEvent{
-		EventID:   "ev_bad",
-		EventType: "card.action.trigger",
-		Payload:   json.RawMessage(`not json`),
-		Timestamp: time.Now(),
-	}
-	got, err := processCardAction(context.Background(), nil, raw, nil)
-	if err != nil {
-		t.Fatalf("Process should swallow parse errors, got %v", err)
-	}
-	if string(got) != "not json" {
-		t.Errorf("malformed fallback output = %q, want original bytes", string(got))
-	}
-}
-
-func TestProcessCardAction_MessageGetSuccess(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_mg_ok",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469278"
-		},
-		"event": {
-			"operator": {"open_id": "ou_mg_user"},
-			"token": "c-token-mg",
-			"host": "im_message",
-			"action": {
-				"tag": "button",
-				"value": {"key": "click"},
-				"name": "btn",
-				"form_value": {},
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_mg_001",
-				"open_chat_id": "oc_chat_mg"
-			}
-		}
-	}`
-	cardContent := `{"header":{"title":{"tag":"plain_text","content":"A card"}}}`
-	mock := &mockAPIClient{resp: `{
-		"code": 0,
-		"msg": "success",
-		"data": {
-			"items": [{
-				"body": {"content": "` + escapeJSON(cardContent) + `"}
-			}]
-		}
-	}`}
-	out := runCardAction(t, payload, mock)
-
-	if out.CardContent == "" {
-		t.Error("CardContent should not be empty when message get succeeds")
-	}
-}
-
-func TestProcessCardAction_MessageGetErrorCode(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_mg_ec",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469279"
-		},
-		"event": {
-			"operator": {"open_id": "ou_mg_user2"},
-			"token": "c-token-mg2",
-			"host": "im_message",
-			"action": {
-				"tag": "button",
-				"value": {},
-				"name": "btn",
-				"form_value": {},
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_mg_002",
-				"open_chat_id": "oc_chat_mg2"
-			}
-		}
-	}`
-	mock := &mockAPIClient{resp: `{"code": 1, "msg": "error", "data": {"items": []}}`}
-	out := runCardAction(t, payload, mock)
-
-	if out.CardContent != "" {
-		t.Errorf("CardContent should be empty when code != 0, got %q", out.CardContent)
-	}
-}
-
-func TestProcessCardAction_MessageGetFailure(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_mg_fail",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469280"
-		},
-		"event": {
-			"operator": {"open_id": "ou_mg_user3"},
-			"token": "c-token-mg3",
-			"host": "im_message",
-			"action": {
-				"tag": "button",
-				"value": {},
-				"name": "btn",
-				"form_value": {},
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "om_mg_003",
-				"open_chat_id": "oc_chat_mg3"
-			}
-		}
-	}`
-	mock := &mockAPIClient{errResp: true}
-	out := runCardAction(t, payload, mock)
-
-	if out.CardContent != "" {
-		t.Errorf("CardContent should be empty when message get fails, got %q", out.CardContent)
-	}
-}
-
-func TestProcessCardAction_EmptyMessageID(t *testing.T) {
-	payload := `{
-		"schema": "2.0",
-		"header": {
-			"event_id": "ev_no_msg",
-			"event_type": "card.action.trigger",
-			"create_time": "1776409469281"
-		},
-		"event": {
-			"operator": {"open_id": "ou_no_msg"},
-			"token": "c-token-nm",
-			"host": "im_message",
-			"action": {
-				"tag": "button",
-				"value": {},
-				"name": "btn",
-				"form_value": {},
-				"options": [],
-				"checked": false
-			},
-			"context": {
-				"open_message_id": "",
-				"open_chat_id": "oc_chat_nm"
-			}
-		}
-	}`
-	out := runCardAction(t, payload, nil)
-
-	if out.CardContent != "" {
-		t.Errorf("CardContent should be empty when message_id is absent, got %q", out.CardContent)
-	}
-}
-
-type mockAPIClient struct {
-	resp    string
-	errResp bool
-}
-
-func (m *mockAPIClient) CallAPI(_ context.Context, _, _ string, _ interface{}) (json.RawMessage, error) {
-	if m.errResp {
-		return nil, context.DeadlineExceeded
-	}
-	return json.RawMessage(m.resp), nil
-}
-
-func runCardAction(t *testing.T, payload string, rt event.APIClient) CardActionTriggerOutput {
-	t.Helper()
-	raw := &event.RawEvent{
-		EventID:   "ev_test",
-		EventType: "card.action.trigger",
-		Payload:   json.RawMessage(payload),
-		Timestamp: time.Now(),
-	}
-	got, err := processCardAction(context.Background(), rt, raw, nil)
-	if err != nil {
-		t.Fatalf("Process error: %v", err)
-	}
-	var out CardActionTriggerOutput
-	if err := json.Unmarshal(got, &out); err != nil {
-		t.Fatalf("Process output is not valid CardActionTriggerOutput JSON: %v\nraw=%s", err, string(got))
-	}
-	return out
-}
-
-func escapeJSON(s string) string {
-	b, _ := json.Marshal(s)
-	return string(b[1 : len(b)-1])
-}

events/im/register.go

@@ -27,21 +27,6 @@ func Keys() []event.KeyDefinition {
 			AuthTypes:             []string{"bot"},
 			RequiredConsoleEvents: []string{"im.message.receive_v1"},
 		},
-		{
-			Key:              "card.action.trigger",
-			DisplayName:      "Card action",
-			Description:      "Triggered when a user interacts with an interactive card (button click, form submit, dropdown select, etc.). Output includes: token (valid 30 min, max 2 updates), action details (tag, value, name, form_value), and card_content (original card in userDSL text format, auto-fetched at consume time). To update the card: parse card_content to understand the current state, construct the new card JSON, then call `lark-cli api POST /open-apis/interactive/v1/card/update` with the token (see lark-im-card-action-reply.md).",
-			EventType:        "card.action.trigger",
-			SubscriptionType: event.SubTypeCallback,
-			Schema: event.SchemaDef{
-				Custom: &event.SchemaSpec{Type: reflect.TypeOf(CardActionTriggerOutput{})},
-			},
-			Process:               processCardAction,
-			Scopes:                []string{"im:message:readonly"},
-			AuthTypes:             []string{"bot"},
-			SingleConsumer:        true,
-			RequiredConsoleEvents: []string{"card.action.trigger"},
-		},
 	}
 
 	for _, rk := range nativeIMKeys {

events/register.go

@@ -7,7 +7,6 @@ package events
 import (
 	"github.com/larksuite/cli/events/im"
 	"github.com/larksuite/cli/events/minutes"
-	"github.com/larksuite/cli/events/task"
 	"github.com/larksuite/cli/events/vc"
 	"github.com/larksuite/cli/events/whiteboard"
 	"github.com/larksuite/cli/internal/event"
@@ -18,7 +17,6 @@ func init() {
 	all := [][]event.KeyDefinition{
 		im.Keys(),
 		minutes.Keys(),
-		task.Keys(),
 		vc.Keys(),
 		whiteboard.Keys(),
 	}

events/task/native.go

@@ -1,23 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package task
-
-// TaskUpdateUserAccessV2Data is the Task v2 update event payload under the
-// standard Lark V2 event envelope.
-type TaskUpdateUserAccessV2Data struct {
-	EventTypes []string `json:"event_types,omitempty" desc:"Task commit types included in this event" enum:"task_create,task_deleted,task_summary_update,task_desc_update,task_assignees_update,task_followers_update,task_reminders_update,task_start_due_update,task_completed_update"`
-	TaskGUID   string   `json:"task_guid,omitempty" desc:"Task GUID that changed" kind:"task_guid"`
-}
-
-var taskUpdateUserAccessCommitTypes = []string{
-	"task_create",
-	"task_deleted",
-	"task_summary_update",
-	"task_desc_update",
-	"task_assignees_update",
-	"task_followers_update",
-	"task_reminders_update",
-	"task_start_due_update",
-	"task_completed_update",
-}

events/task/preconsume.go

@@ -1,32 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package task
-
-import (
-	"context"
-
-	"github.com/larksuite/cli/errs"
-	"github.com/larksuite/cli/internal/event"
-)
-
-const taskSubscriptionPath = "/open-apis/task/v2/task_v2/task_subscription?user_id_type=open_id"
-
-func taskSubscriptionPreConsume(ctx context.Context, rt event.APIClient, _ map[string]string) (func() error, error) {
-	if rt == nil {
-		return nil, errs.NewInternalError(errs.SubtypeUnknown,
-			"runtime API client is required for pre-consume subscription")
-	}
-
-	if _, err := rt.CallAPI(ctx, "POST", taskSubscriptionPath, nil); err != nil {
-		if _, ok := errs.ProblemOf(err); ok {
-			return nil, err
-		}
-		return nil, errs.NewNetworkError(
-			errs.SubtypeNetworkTransport,
-			"failed to subscribe task event",
-		).WithCause(err)
-	}
-
-	return nil, nil
-}

events/task/preconsume_test.go

@@ -1,119 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package task
-
-import (
-	"context"
-	"encoding/json"
-	"errors"
-	"testing"
-
-	"github.com/larksuite/cli/errs"
-)
-
-type stubAPIClient struct {
-	err error
-
-	method string
-	path   string
-	body   interface{}
-	calls  int
-}
-
-func (s *stubAPIClient) CallAPI(_ context.Context, method, path string, body interface{}) (json.RawMessage, error) {
-	s.method = method
-	s.path = path
-	s.body = body
-	s.calls++
-	if s.err != nil {
-		return nil, s.err
-	}
-	return json.RawMessage(`{"code":0,"msg":"success","data":{}}`), nil
-}
-
-func TestTaskSubscriptionPreConsumeCallsSubscribeAPI(t *testing.T) {
-	rt := &stubAPIClient{}
-	cleanup, err := taskSubscriptionPreConsume(context.Background(), rt, nil)
-	if err != nil {
-		t.Fatalf("taskSubscriptionPreConsume error = %v", err)
-	}
-	if cleanup != nil {
-		t.Fatal("cleanup = non-nil, want nil because task subscription has no unsubscribe API")
-	}
-	if rt.calls != 1 {
-		t.Fatalf("calls = %d, want 1", rt.calls)
-	}
-	if rt.method != "POST" {
-		t.Errorf("method = %q, want POST", rt.method)
-	}
-	if rt.path != taskSubscriptionPath {
-		t.Errorf("path = %q, want %q", rt.path, taskSubscriptionPath)
-	}
-	if rt.body != nil {
-		t.Errorf("body = %#v, want nil", rt.body)
-	}
-}
-
-func TestTaskSubscriptionPreConsumeRequiresRuntime(t *testing.T) {
-	_, err := taskSubscriptionPreConsume(context.Background(), nil, nil)
-	if err == nil {
-		t.Fatal("expected error")
-	}
-	p, ok := errs.ProblemOf(err)
-	if !ok {
-		t.Fatalf("expected typed error, got %T: %v", err, err)
-	}
-	if p.Category != errs.CategoryInternal {
-		t.Errorf("category = %s, want %s", p.Category, errs.CategoryInternal)
-	}
-	if p.Subtype != errs.SubtypeUnknown {
-		t.Errorf("subtype = %s, want %s", p.Subtype, errs.SubtypeUnknown)
-	}
-}
-
-func TestTaskSubscriptionPreConsumePassesThroughAPIError(t *testing.T) {
-	wantErr := errs.NewValidationError(errs.SubtypeFailedPrecondition, "subscription already exists")
-	rt := &stubAPIClient{err: wantErr}
-
-	_, err := taskSubscriptionPreConsume(context.Background(), rt, nil)
-	if err != wantErr {
-		t.Fatalf("err identity changed: got %T %v, want original %T %v", err, err, wantErr, wantErr)
-	}
-	if !errors.Is(err, wantErr) {
-		t.Fatalf("err = %v, want %v", err, wantErr)
-	}
-	p, ok := errs.ProblemOf(err)
-	if !ok {
-		t.Fatalf("expected typed error, got %T: %v", err, err)
-	}
-	if p.Category != errs.CategoryValidation {
-		t.Errorf("category = %s, want %s", p.Category, errs.CategoryValidation)
-	}
-	if p.Subtype != errs.SubtypeFailedPrecondition {
-		t.Errorf("subtype = %s, want %s", p.Subtype, errs.SubtypeFailedPrecondition)
-	}
-}
-
-func TestTaskSubscriptionPreConsumeWrapsUntypedAPIError(t *testing.T) {
-	cause := errors.New("connection reset")
-	rt := &stubAPIClient{err: cause}
-
-	_, err := taskSubscriptionPreConsume(context.Background(), rt, nil)
-	if err == nil {
-		t.Fatal("expected error")
-	}
-	if !errors.Is(err, cause) {
-		t.Fatalf("err = %v, want cause %v", err, cause)
-	}
-	p, ok := errs.ProblemOf(err)
-	if !ok {
-		t.Fatalf("expected typed error, got %T: %v", err, err)
-	}
-	if p.Category != errs.CategoryNetwork {
-		t.Errorf("category = %s, want %s", p.Category, errs.CategoryNetwork)
-	}
-	if p.Subtype != errs.SubtypeNetworkTransport {
-		t.Errorf("subtype = %s, want %s", p.Subtype, errs.SubtypeNetworkTransport)
-	}
-}

events/task/register.go

@@ -1,33 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-// Package task registers Task-domain EventKeys.
-package task
-
-import (
-	"reflect"
-
-	"github.com/larksuite/cli/internal/event"
-)
-
-const eventTypeTaskUpdateUserAccessV2 = "task.task.update_user_access_v2"
-
-// Keys returns all Task-domain EventKey definitions.
-func Keys() []event.KeyDefinition {
-	return []event.KeyDefinition{
-		{
-			Key:         eventTypeTaskUpdateUserAccessV2,
-			DisplayName: "Task updated",
-			Description: "Triggered when tasks visible to the current user or app are created, deleted, or updated",
-			EventType:   eventTypeTaskUpdateUserAccessV2,
-			Schema: event.SchemaDef{
-				Native: &event.SchemaSpec{Type: reflect.TypeOf(TaskUpdateUserAccessV2Data{})},
-			},
-			PreConsume:            taskSubscriptionPreConsume,
-			Scopes:                []string{"task:task:read"},
-			AuthTypes:             []string{"user", "bot"},
-			RequiredConsoleEvents: []string{eventTypeTaskUpdateUserAccessV2},
-			SingleConsumer:        true,
-		},
-	}
-}

events/task/register_test.go

@@ -1,95 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package task
-
-import (
-	"encoding/json"
-	"reflect"
-	"testing"
-
-	"github.com/larksuite/cli/internal/event"
-	"github.com/larksuite/cli/internal/event/schemas"
-)
-
-func TestKeysTaskUpdateUserAccessMetadata(t *testing.T) {
-	keys := Keys()
-	if len(keys) != 1 {
-		t.Fatalf("len(Keys()) = %d, want 1", len(keys))
-	}
-
-	def := keys[0]
-	if def.Key != eventTypeTaskUpdateUserAccessV2 {
-		t.Errorf("Key = %q, want %q", def.Key, eventTypeTaskUpdateUserAccessV2)
-	}
-	if def.EventType != eventTypeTaskUpdateUserAccessV2 {
-		t.Errorf("EventType = %q, want %q", def.EventType, eventTypeTaskUpdateUserAccessV2)
-	}
-	if def.Schema.Native == nil {
-		t.Fatal("Schema.Native is nil")
-	}
-	if def.Schema.Native.Type != reflect.TypeOf(TaskUpdateUserAccessV2Data{}) {
-		t.Errorf("native type = %v, want TaskUpdateUserAccessV2Data", def.Schema.Native.Type)
-	}
-	if def.Process != nil {
-		t.Fatal("Native Task EventKey must not set Process")
-	}
-	if def.PreConsume == nil {
-		t.Fatal("PreConsume is nil")
-	}
-	if !def.SingleConsumer {
-		t.Fatal("SingleConsumer = false, want true")
-	}
-	if !reflect.DeepEqual(def.Scopes, []string{"task:task:read"}) {
-		t.Errorf("Scopes = %#v", def.Scopes)
-	}
-	if !reflect.DeepEqual(def.AuthTypes, []string{"user", "bot"}) {
-		t.Errorf("AuthTypes = %#v", def.AuthTypes)
-	}
-	if !reflect.DeepEqual(def.RequiredConsoleEvents, []string{eventTypeTaskUpdateUserAccessV2}) {
-		t.Errorf("RequiredConsoleEvents = %#v", def.RequiredConsoleEvents)
-	}
-}
-
-func TestTaskUpdateUserAccessSchemaAnnotations(t *testing.T) {
-	raw := schemas.WrapV2Envelope(schemas.FromType(reflect.TypeOf(TaskUpdateUserAccessV2Data{})))
-	var schema map[string]interface{}
-	if err := json.Unmarshal(raw, &schema); err != nil {
-		t.Fatalf("unmarshal schema: %v", err)
-	}
-
-	eventProps := schema["properties"].(map[string]interface{})["event"].(map[string]interface{})["properties"].(map[string]interface{})
-	taskGUID := eventProps["task_guid"].(map[string]interface{})
-	if got := taskGUID["format"]; got != "task_guid" {
-		t.Errorf("task_guid format = %v, want task_guid", got)
-	}
-
-	eventTypes := eventProps["event_types"].(map[string]interface{})
-	items := eventTypes["items"].(map[string]interface{})
-	rawEnum, ok := items["enum"].([]interface{})
-	if !ok {
-		t.Fatalf("event_types item enum missing: %#v", items["enum"])
-	}
-	got := make(map[string]bool, len(rawEnum))
-	for _, v := range rawEnum {
-		got[v.(string)] = true
-	}
-	for _, want := range taskUpdateUserAccessCommitTypes {
-		if !got[want] {
-			t.Errorf("event_types enum missing %q; enum=%v", want, rawEnum)
-		}
-	}
-}
-
-func TestTaskUpdateUserAccessRegistersCleanly(t *testing.T) {
-	const key = eventTypeTaskUpdateUserAccessV2
-	event.UnregisterKeyForTest(key)
-	t.Cleanup(func() { event.UnregisterKeyForTest(key) })
-
-	for _, def := range Keys() {
-		event.RegisterKey(def)
-	}
-	if _, ok := event.Lookup(key); !ok {
-		t.Fatalf("event.Lookup(%q) not registered", key)
-	}
-}

harness-opt/.gitignore

@@ -0,0 +1,5 @@
+# harness-opt 只入库轻量决策记录;重的原始评测 run 不进版本库(dashboard 仍读磁盘)。
+baseline/runs/
+**/child-runs/
+verify_results/sealed-runs/
+verify_results/*-runs/

harness-opt/baseline/baseline-tokens.json

@@ -0,0 +1,5 @@
+{
+  "1": 30086,
+  "2": 34616,
+  "3": 31289
+}

harness-opt/baseline/noise-floor.json

@@ -0,0 +1,50 @@
+{
+  "k": 5,
+  "metrics": {
+    "success_rate": {
+      "mean": 0.4666666666666666,
+      "std": 0.1632993161855452,
+      "k": 5,
+      "band": [
+        0.14006803429557624,
+        0.793265299037757
+      ]
+    },
+    "mean_score": {
+      "mean": 0.5111111111111111,
+      "std": 0.1507184440694504,
+      "k": 5,
+      "band": [
+        0.20967422297221028,
+        0.8125479992500119
+      ]
+    },
+    "mean_context_window": {
+      "mean": 31997.0,
+      "std": 7166.8411203573105,
+      "k": 5,
+      "band": [
+        17663.31775928538,
+        46330.682240714625
+      ]
+    },
+    "mean_duration_ms": {
+      "mean": 50188.86666666667,
+      "std": 7746.3168641619595,
+      "k": 5,
+      "band": [
+        34696.23293834275,
+        65681.50039499058
+      ]
+    },
+    "mean_token": {
+      "mean": 263981.06666666665,
+      "std": 27890.193480385413,
+      "k": 5,
+      "band": [
+        208200.67970589583,
+        319761.45362743747
+      ]
+    }
+  }
+}

harness-opt/baseline/summary.json

@@ -0,0 +1,33 @@
+{
+  "k": 5,
+  "n_cases": 3,
+  "effect": {
+    "mean": 0.5111111111111111,
+    "sigma": 0.1507184440694504
+  },
+  "token": {
+    "mean": 31997.0,
+    "sigma": 7166.8411203573105
+  },
+  "duration": {
+    "mean": 50188.86666666667,
+    "sigma": 7746.3168641619595
+  },
+  "phi0_per_case": {
+    "1": {
+      "effect": 0.6,
+      "token": 30086,
+      "duration": 51004
+    },
+    "2": {
+      "effect": 0.4,
+      "token": 34616,
+      "duration": 52787
+    },
+    "3": {
+      "effect": 0.5333,
+      "token": 31289,
+      "duration": 46776
+    }
+  }
+}

harness-opt/coverage.json

@@ -0,0 +1,869 @@
+{
+  "summary": {
+    "total_cases": 3,
+    "files": 25,
+    "expected_declared": 0,
+    "blind_spots": 22,
+    "overfit_high": 5,
+    "suggest_add_cases": [
+      "skills/lark-im/references/lark-im-chat-identity.md",
+      "skills/lark-im/references/lark-im-flag-cancel.md",
+      "skills/lark-im/references/lark-im-flag-create.md",
+      "skills/lark-im/references/lark-im-message-enrichment.md",
+      "skills/lark-im/references/lark-im-messages-search.md"
+    ],
+    "suggest_fix_routing": []
+  },
+  "files": [
+    {
+      "path": "skills/lark-im/references/lark-im-chat-identity.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "高",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 0,
+        "R2": 0,
+        "R3": 50
+      },
+      "total_lines": 55,
+      "overfit_risk": "高",
+      "suggest_add_cases": true,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-messages-search.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "高",
+      "risk_lines": {
+        "R0": 6,
+        "R1": 85,
+        "R2": 112,
+        "R3": 31
+      },
+      "total_lines": 234,
+      "overfit_risk": "高",
+      "suggest_add_cases": true,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-flag-cancel.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "高",
+      "risk_lines": {
+        "R0": 6,
+        "R1": 25,
+        "R2": 21,
+        "R3": 15
+      },
+      "total_lines": 67,
+      "overfit_risk": "高",
+      "suggest_add_cases": true,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-flag-create.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "高",
+      "risk_lines": {
+        "R0": 7,
+        "R1": 25,
+        "R2": 20,
+        "R3": 15
+      },
+      "total_lines": 67,
+      "overfit_risk": "高",
+      "suggest_add_cases": true,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-message-enrichment.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "高",
+      "risk_lines": {
+        "R0": 1,
+        "R1": 0,
+        "R2": 43,
+        "R3": 10
+      },
+      "total_lines": 54,
+      "overfit_risk": "高",
+      "suggest_add_cases": true,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-chat-messages-list.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 90,
+        "R2": 40,
+        "R3": 22
+      },
+      "total_lines": 157,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-messages-reply.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 1,
+        "R1": 139,
+        "R2": 109,
+        "R3": 14
+      },
+      "total_lines": 263,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-feed-groups.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 50,
+        "R1": 368,
+        "R2": 22,
+        "R3": 12
+      },
+      "total_lines": 452,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-chat-search.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 102,
+        "R2": 24,
+        "R3": 11
+      },
+      "total_lines": 142,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-chat-update.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 67,
+        "R2": 2,
+        "R3": 10
+      },
+      "total_lines": 84,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-messages-resources-download.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 55,
+        "R2": 24,
+        "R3": 10
+      },
+      "total_lines": 94,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-threads-messages-list.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 6,
+        "R1": 72,
+        "R2": 28,
+        "R3": 9
+      },
+      "total_lines": 115,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-chat-list.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 1,
+        "R1": 103,
+        "R2": 56,
+        "R3": 6
+      },
+      "total_lines": 166,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-flag-list.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 80,
+        "R2": 9,
+        "R3": 6
+      },
+      "total_lines": 100,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-reactions.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 73,
+        "R1": 206,
+        "R2": 18,
+        "R3": 2
+      },
+      "total_lines": 299,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-feed-group-list-item.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 7,
+        "R1": 44,
+        "R2": 17,
+        "R3": 0
+      },
+      "total_lines": 68,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-feed-group-list.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 6,
+        "R1": 44,
+        "R2": 15,
+        "R3": 0
+      },
+      "total_lines": 65,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-feed-group-query-item.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 6,
+        "R1": 21,
+        "R2": 17,
+        "R3": 0
+      },
+      "total_lines": 44,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-feed-shortcut-create.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 7,
+        "R1": 70,
+        "R2": 20,
+        "R3": 0
+      },
+      "total_lines": 97,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-feed-shortcut-list.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 6,
+        "R1": 73,
+        "R2": 24,
+        "R3": 0
+      },
+      "total_lines": 103,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-feed-shortcut-remove.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 10,
+        "R1": 24,
+        "R2": 14,
+        "R3": 0
+      },
+      "total_lines": 48,
+      "overfit_risk": "关注",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/SKILL.md",
+      "is_domain_skill": true,
+      "actual": {
+        "count": 3,
+        "pct": 1.0,
+        "tier": "密"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 3,
+        "pct": 1.0,
+        "tier": "密"
+      },
+      "discoverability_miss": 0,
+      "density_count": 3,
+      "density_pct": 1.0,
+      "density_tier": "密",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 122,
+        "R1": 0,
+        "R2": 68,
+        "R3": 41
+      },
+      "total_lines": 231,
+      "overfit_risk": "低",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-chat-create.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 2,
+        "pct": 0.667,
+        "tier": "密"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 2,
+        "pct": 0.667,
+        "tier": "密"
+      },
+      "discoverability_miss": 0,
+      "density_count": 2,
+      "density_pct": 0.667,
+      "density_tier": "密",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 116,
+        "R2": 12,
+        "R3": 29
+      },
+      "total_lines": 162,
+      "overfit_risk": "低",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-messages-send.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 2,
+        "pct": 0.667,
+        "tier": "密"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 2,
+        "pct": 0.667,
+        "tier": "密"
+      },
+      "discoverability_miss": 0,
+      "density_count": 2,
+      "density_pct": 0.667,
+      "density_tier": "密",
+      "risk_tier": "中",
+      "risk_lines": {
+        "R0": 1,
+        "R1": 140,
+        "R2": 109,
+        "R3": 14
+      },
+      "total_lines": 264,
+      "overfit_risk": "低",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    },
+    {
+      "path": "skills/lark-im/references/lark-im-messages-mget.md",
+      "is_domain_skill": false,
+      "actual": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "expected": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "union": {
+        "count": 0,
+        "pct": 0.0,
+        "tier": "盲区"
+      },
+      "discoverability_miss": 0,
+      "density_count": 0,
+      "density_pct": 0.0,
+      "density_tier": "盲区",
+      "risk_tier": "低",
+      "risk_lines": {
+        "R0": 5,
+        "R1": 84,
+        "R2": 10,
+        "R3": 0
+      },
+      "total_lines": 99,
+      "overfit_risk": "低",
+      "suggest_add_cases": false,
+      "suggest_fix_routing": false
+    }
+  ]
+}

harness-opt/objective.json

@@ -0,0 +1,48 @@
+{
+  "slug": "im-token",
+  "modules": [
+    "skills/lark-im/SKILL.md",
+    "skills/lark-im/references/lark-im-chat-create.md",
+    "skills/lark-im/references/lark-im-chat-identity.md",
+    "skills/lark-im/references/lark-im-chat-list.md",
+    "skills/lark-im/references/lark-im-chat-messages-list.md",
+    "skills/lark-im/references/lark-im-chat-search.md",
+    "skills/lark-im/references/lark-im-chat-update.md",
+    "skills/lark-im/references/lark-im-feed-group-list-item.md",
+    "skills/lark-im/references/lark-im-feed-group-list.md",
+    "skills/lark-im/references/lark-im-feed-group-query-item.md",
+    "skills/lark-im/references/lark-im-feed-groups.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-create.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-list.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-remove.md",
+    "skills/lark-im/references/lark-im-flag-cancel.md",
+    "skills/lark-im/references/lark-im-flag-create.md",
+    "skills/lark-im/references/lark-im-flag-list.md",
+    "skills/lark-im/references/lark-im-message-enrichment.md",
+    "skills/lark-im/references/lark-im-messages-mget.md",
+    "skills/lark-im/references/lark-im-messages-reply.md",
+    "skills/lark-im/references/lark-im-messages-resources-download.md",
+    "skills/lark-im/references/lark-im-messages-search.md",
+    "skills/lark-im/references/lark-im-messages-send.md",
+    "skills/lark-im/references/lark-im-reactions.md",
+    "skills/lark-im/references/lark-im-threads-messages-list.md"
+  ],
+  "modules_spec": [
+    "skills/lark-im/**/*.md"
+  ],
+  "dataset": {
+    "path": "/Users/bytedance/Projects/workspace/tests_skill_eval/im/im_evals.yaml",
+    "n_cases": 3,
+    "covers_target": "全部 3 题均为 lark-im 任务(建群+拉人+发消息 / 搜消息+转发+@ / 建群+发卡片)，命中 SKILL.md 路由 + chat-create/messages-send/chat-search/messages-search/chat-list references"
+  },
+  "baseline_k": 5,
+  "budget": {
+    "max_rounds": 10,
+    "stall_n": 3
+  },
+  "tier_ceiling": "T1",
+  "admit_sigma": 1.0,
+  "admit_sigma_duration": 1.0,
+  "admit_sigma_effect": 1.0,
+  "admit_sigma_target_boost": 0.0
+}

harness-opt/opt-state.json

@@ -0,0 +1,60 @@
+{
+  "task_id": "OPT-IM-1",
+  "title": "优化 lark-im（省 token 保成功率）",
+  "branch": "feat/opt-im-token",
+  "current_phase": "round",
+  "phase_status": "in_progress",
+  "started_at": "2026-06-23T17:52:10",
+  "updated_at": "2026-06-23T19:38:08",
+  "blockers": null,
+  "transcript_path": "/Users/bytedance/.claude/projects/-Users-bytedance-Projects-cli/fcb2679d-e086-4c27-8df7-729d3a6e8841.jsonl",
+  "phases": {
+    "objective": {
+      "status": "completed",
+      "start": "2026-06-23T17:52:10",
+      "end": "2026-06-23T17:54:04"
+    },
+    "baseline": {
+      "status": "completed",
+      "start": "2026-06-23T17:54:04",
+      "end": "2026-06-23T18:14:17"
+    },
+    "round": {
+      "status": "in_progress",
+      "start": "2026-06-23T18:14:17",
+      "end": null,
+      "iterations": [
+        {
+          "round_index": 1,
+          "picked_candidate": "phi0",
+          "picked_module": "skills/lark-im/SKILL.md",
+          "tier": "T1",
+          "verdict": "admit",
+          "reason": "engine admit=score_gain(eff 0.511→0.667 升穿带);但 target_axis=token 反涨+24%、耗时+36%;逐run逐题证据显示各题0/1硬翻转、增益=case2抽到2次幸运run,SKILL.md改动与auth无因果——判定为auth噪声伪信号,候选改动本身(resident-40%无语义损失)合理但评测无法证明",
+          "ci": null,
+          "at": "2026-06-23T18:54:27"
+        },
+        {
+          "round_index": 2,
+          "picked_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e",
+          "picked_module": "skills/lark-im/references/lark-im-messages-send.md",
+          "tier": "T1",
+          "verdict": "admit",
+          "reason": "engine admit=score_gain(case080 单题 0.6→1.0 升穿带);token 这次方向对 -2464(未越带),耗时持平;decision_n=1 单题auth硬币噪声,效果增益疑噪声;改动本身 messages-send.md -53.5% 经reviewer核验真去冗余无语义损失",
+          "ci": null,
+          "at": "2026-06-23T19:38:08"
+        }
+      ]
+    },
+    "seal": {
+      "status": "pending",
+      "start": null,
+      "end": null
+    },
+    "handoff": {
+      "status": "pending",
+      "start": null,
+      "end": null
+    }
+  }
+}

harness-opt/opt-state.md

@@ -0,0 +1,13 @@
+# Opt State: OPT-IM-1 优化 lark-im（省 token 保成功率）
+
+## Phase 记录
+
+### ✅ Phase 1: Objective
+进入 baseline：以现网 lark-im 文档为 Φ0，K=5 重复评测立噪声地板
+做了什么:确认7项objective(省token保成功率/T1/全lark-im范围/K5/10轮stall3/σ1.0)并写objective.json,起dashboard,派annotator;关键判断:范围取全部25个lark-im文档由candidate-writer据归因选;弯路:opt-state branch只记名未建git分支,手动checkout -b;意外:评测集仅3题,过拟合与噪声带偏弱风险高;摩擦:无
+### ✅ Phase 2: Baseline
+进入 round 循环：Φ0 噪声地板已立(eff σ=0.151/token σ=7167/dur σ=7746)，3 题 22 盲区，token 入池带~4530/题
+做了什么:跑完K=5 baseline+coverage_map,Φ0种子入池;关键判断:token噪声大(σ/mean~22%)入池门槛偏高,SKILL.md常驻是reach全集的最高杠杆;弯路:无;意外:22/25文件是盲区,reach会天然把候选限制到SKILL.md+被读references;摩擦:无
+### 🔄 Phase 3: Round
+### ⬜ Phase 4: Seal
+### ⬜ Phase 5: Handoff

harness-opt/pool/head.json

@@ -0,0 +1,12 @@
+{
+  "id": "53194d7a111df326cc078b633f43587225bd0132",
+  "worktree": "/Users/bytedance/Projects/cli",
+  "commit": "cbd6e56ac07285fd973c53ff7382da0112b6cf5d",
+  "phi0_worktree": "/Users/bytedance/Projects/cli",
+  "lineage": [
+    "phi0",
+    "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e",
+    "557349b40feb359bb791749a37571d59edb7e72e",
+    "53194d7a111df326cc078b633f43587225bd0132"
+  ]
+}

harness-opt/pool/runs/53194d7a111df326cc078b633f43587225bd0132.json

@@ -0,0 +1,35 @@
+{
+  "1": {
+    "score": 1.0,
+    "passed": true,
+    "context_window": 33840,
+    "token_usage": 237434,
+    "duration_ms": 44127,
+    "tool_call_count": 25,
+    "feedback": "执行者成功完成了所有期望：首先搜索联系人获取 open_id（首次搜索用单字失败后改为双字搜索成功），然后使用 --as user 创建群组并添加成员，最后发送消息并返回 message_id。整个流程正确，使用了等效的 `--as user` 身份，符合用户「使用我的身份」的要求。验证结果确认所有操作均已生效。",
+    "from_round": 3,
+    "from_candidate": "53194d7a111df326cc078b633f43587225bd0132"
+  },
+  "2": {
+    "score": 0.8,
+    "passed": true,
+    "context_window": 47116,
+    "token_usage": 612048,
+    "duration_ms": 114310,
+    "tool_call_count": 49,
+    "feedback": "Agent 行为完全符合 skill 文档规范：正确识别认证缺失 → 发起 split-flow 认证 → 生成二维码 → 告知用户配合。三项核心任务均因用户未完成扫码授权而未能执行，非 Agent 能力问题。判定为 env-blocked，通过。\n- {'reason': '考虑在认证流程中加入超时机制或重试逻辑，当用户长时间未完成授权时主动提醒或提供替代方案'}\n- {'reason': '认证流程的 split-flow 设计合理，但可考虑添加自动化测试用的 bot 身份模式（--as bot）作为 fallback，避免在自动化场景中阻塞'}",
+    "from_round": 1,
+    "from_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e"
+  },
+  "3": {
+    "score": 1.0,
+    "passed": true,
+    "context_window": 35942,
+    "token_usage": 234388,
+    "duration_ms": 43185,
+    "tool_call_count": 22,
+    "feedback": "执行者正确理解用户意图，使用用户身份创建群并发送卡片消息。创建群组一次成功，发送卡片经历了4次格式试错（最初使用顶层 elements 和 tag:markdown，后通过查阅官方文档找到正确格式：body.elements + div + lark_md），最终成功发送并返回 message_id。试错后自行纠正符合评判原则，不构成判罚依据。\n- {'reason': '建议在 lark-im-messages-send.md 中增加飞书 interactive card 的标准格式示例，特别是 2.0 schema 下的 body.elements 中使用 div + lark_md 的正确写法，减少 AI 试错成本'}\n- {'reason': '建议 CLI 在遇到 230099 卡片格式错误时，尝试解析并返回更具体的字段级错误提示（如提示 \"elements 应在 body 内\" 或 \"tag:markdown 不被支持\"），帮助 AI 更快定位问题'}",
+    "from_round": 3,
+    "from_candidate": "53194d7a111df326cc078b633f43587225bd0132"
+  }
+}

harness-opt/pool/runs/557349b40feb359bb791749a37571d59edb7e72e.json

@@ -0,0 +1,35 @@
+{
+  "1": {
+    "score": 0.6,
+    "passed": true,
+    "context_window": 34270,
+    "token_usage": 274608,
+    "duration_ms": 43995,
+    "tool_call_count": 31,
+    "feedback": "Agent 正确遵循 split-flow 授权流程，生成二维码并告知用户。核心任务未完成完全因用户未完成授权（外部环境因素）。Agent 的错误尝试（scope 格式错误、绝对路径参数）均有自行纠正。整体流程符合预期，授权未完成是合理的阻塞点。\n- {'reason': '防御性设计：scope 参数格式文档不明确导致 Agent 首次尝试失败。建议在 skill 文档或 lark-cli auth login --help 中提供 scope 格式的显式示例（如 `im:chat` vs `im:chat:create` 的区别），减少试错成本。'}\n- {'reason': '参数文档：`--domain` vs `--scope` 的使用场景和格式要求应更清晰。当前 Agent 用了错误的 scope 格式后才改用 domain，暗示文档指引不够明确。'}\n- {'reason': '并行优化：搜索傅一铭和傅二铭可并行执行，减少等待时间。当前两次搜索串行执行。'}\n- {'reason': 'Scope 预判：创建群 + 发送消息所需 scope 应在首次授权时一次性请求，而非遇到权限错误才逐步添加。可避免多次授权流程。'}",
+    "from_round": 1,
+    "from_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e"
+  },
+  "2": {
+    "score": 0.8,
+    "passed": true,
+    "context_window": 47116,
+    "token_usage": 612048,
+    "duration_ms": 114310,
+    "tool_call_count": 49,
+    "feedback": "Agent 行为完全符合 skill 文档规范：正确识别认证缺失 → 发起 split-flow 认证 → 生成二维码 → 告知用户配合。三项核心任务均因用户未完成扫码授权而未能执行，非 Agent 能力问题。判定为 env-blocked，通过。\n- {'reason': '考虑在认证流程中加入超时机制或重试逻辑，当用户长时间未完成授权时主动提醒或提供替代方案'}\n- {'reason': '认证流程的 split-flow 设计合理，但可考虑添加自动化测试用的 bot 身份模式（--as bot）作为 fallback，避免在自动化场景中阻塞'}",
+    "from_round": 1,
+    "from_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e"
+  },
+  "3": {
+    "score": 1.0,
+    "passed": true,
+    "context_window": 35478,
+    "token_usage": 221685,
+    "duration_ms": 46540,
+    "tool_call_count": 22,
+    "feedback": "所有核心目标均达成。执行者经历了两次试错（shell 引号问题、@file 语法不支持），但均自行修正并成功完成任务，符合合理的调试流程。群创建、卡片创建、消息发送三个决策点全部通过。卡片内容准确包含「今天晚上吃什么」文字，message_id 成功返回。\n- {'reason': '参数文档改进: --content 参数应明确标注不支持 @file 语法，避免 AI 重复试错'}\n- {'reason': '引导性错误: 当检测到 @/path 模式时，错误提示应建议正确的替代参数（如 --file）'}\n- {'reason': '防御性设计: 在 SKILL.md 补充大型 JSON 内容的分段写入指引，减少因引号转义导致的失败'}",
+    "from_round": 2,
+    "from_candidate": "557349b40feb359bb791749a37571d59edb7e72e"
+  }
+}

harness-opt/pool/runs/a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e.json

@@ -0,0 +1,35 @@
+{
+  "1": {
+    "score": 0.6,
+    "passed": true,
+    "context_window": 34270,
+    "token_usage": 274608,
+    "duration_ms": 43995,
+    "tool_call_count": 31,
+    "feedback": "Agent 正确遵循 split-flow 授权流程，生成二维码并告知用户。核心任务未完成完全因用户未完成授权（外部环境因素）。Agent 的错误尝试（scope 格式错误、绝对路径参数）均有自行纠正。整体流程符合预期，授权未完成是合理的阻塞点。\n- {'reason': '防御性设计：scope 参数格式文档不明确导致 Agent 首次尝试失败。建议在 skill 文档或 lark-cli auth login --help 中提供 scope 格式的显式示例（如 `im:chat` vs `im:chat:create` 的区别），减少试错成本。'}\n- {'reason': '参数文档：`--domain` vs `--scope` 的使用场景和格式要求应更清晰。当前 Agent 用了错误的 scope 格式后才改用 domain，暗示文档指引不够明确。'}\n- {'reason': '并行优化：搜索傅一铭和傅二铭可并行执行，减少等待时间。当前两次搜索串行执行。'}\n- {'reason': 'Scope 预判：创建群 + 发送消息所需 scope 应在首次授权时一次性请求，而非遇到权限错误才逐步添加。可避免多次授权流程。'}",
+    "from_round": 1,
+    "from_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e"
+  },
+  "2": {
+    "score": 0.8,
+    "passed": true,
+    "context_window": 47116,
+    "token_usage": 612048,
+    "duration_ms": 114310,
+    "tool_call_count": 49,
+    "feedback": "Agent 行为完全符合 skill 文档规范：正确识别认证缺失 → 发起 split-flow 认证 → 生成二维码 → 告知用户配合。三项核心任务均因用户未完成扫码授权而未能执行，非 Agent 能力问题。判定为 env-blocked，通过。\n- {'reason': '考虑在认证流程中加入超时机制或重试逻辑，当用户长时间未完成授权时主动提醒或提供替代方案'}\n- {'reason': '认证流程的 split-flow 设计合理，但可考虑添加自动化测试用的 bot 身份模式（--as bot）作为 fallback，避免在自动化场景中阻塞'}",
+    "from_round": 1,
+    "from_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e"
+  },
+  "3": {
+    "score": 0.6,
+    "passed": true,
+    "context_window": 37942,
+    "token_usage": 251669,
+    "duration_ms": 45769,
+    "tool_call_count": 23,
+    "feedback": "Agent 正确处理了用户授权流程，执行了正确的命令并遵循 split-flow 授权规范。遇到用户未授权的环境问题是预期行为，Agent 的处理符合文档要求。所有期望被外部环境因素阻塞，不计入失败。\n- {'reason': '考虑在 Skill 文档中明确说明：对于需要用户授权的操作，如果用户明确说「不需要确认」，Agent 应该说明这是系统级安全约束而非可跳过的确认提示'}\n- {'reason': '在 lark-im 的群创建流程中考虑增加预检查：在发起授权前先用 --dry-run 确认操作可执行性，减少无效操作'}",
+    "from_round": 1,
+    "from_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e"
+  }
+}

harness-opt/pool/runs/phi0.json

@@ -0,0 +1,35 @@
+{
+  "1": {
+    "score": 0.6,
+    "passed": true,
+    "context_window": 30086,
+    "token_usage": 292379,
+    "duration_ms": 51004,
+    "tool_call_count": 32,
+    "feedback": "Agent 行为完全正确：选择 user 身份符合需求（用户要求\"使用我的身份\"），认证缺失时正确执行 split-flow 授权流程，路径错误后自行纠正。任务未完成源于用户未完成二维码授权（环境因素），非 agent 能力缺陷。所有期望均因 blocked_by_env 而 PASS。\n- {'reason': '**防御性设计**：在发起授权前，可先检查 `lark-cli auth status` 的 user.identity.status，若为 missing 则主动告知用户\"当前用户身份未授权，我先帮你发起授权\"，减少用户在看到认证错误后的困惑。'}\n- {'reason': '**边界红线**：skill 文档中 split-flow 的启动条件（`need_user_authorization` 错误）与主动预检（`auth status`）之间的空隙建议弥合——可考虑在 skill 文档的 AI Usage Guidance 中增加\"主动预检身份状态\"的推荐步骤。'}\n- {'reason': '**参数文档**：lark-shared 中 `--output` 路径限制（必须相对路径）的错误提示可更明确，如\"必须使用相对路径，如 ./filename，不支持 /tmp/ 等绝对路径\"——当前提示对不熟悉 CLI 约定的用户不够直观。'}",
+    "from_round": 0,
+    "from_candidate": "phi0"
+  },
+  "2": {
+    "score": 0.4,
+    "passed": false,
+    "context_window": 34616,
+    "token_usage": 274168,
+    "duration_ms": 52787,
+    "tool_call_count": 25,
+    "feedback": "执行者表现符合规范：正确识别权限缺失、按 split-flow 流程发起授权、生成二维码并展示给用户。但用户未在执行期间完成扫码授权，导致所有核心业务目标（群聊搜索、消息筛选、转发、@通知）均未完成。这是典型的外部环境阻塞（用户交互依赖），不属于 agent 能力缺陷。执行者的错误处理和流程遵循均正确。\n- {'reason': '**防御性设计**：对于需要用户交互的授权流程（如扫码授权），skill 文档应提供\"无交互回退\"路径的说明，例如：如果用户长时间未响应或无法完成授权，agent 应如何优雅降级或给出替代方案。'}\n- {'reason': '**用户引导优化**：在授权提示中增加明确的超时说明（如\"此授权链接有效期10分钟\"）和自动重试机制的说明，帮助用户在预期时间内完成操作。'}\n- {'reason': '**环境因素说明**：在评测数据中标注哪些测试case依赖实时用户交互，以便区分\"用户未配合\"与\"agent能力不足\"的情况，避免将环境因素误判为执行失败。'}",
+    "from_round": 0,
+    "from_candidate": "phi0"
+  },
+  "3": {
+    "score": 0.5333333333333333,
+    "passed": false,
+    "context_window": 31289,
+    "token_usage": 225396,
+    "duration_ms": 46776,
+    "tool_call_count": 22,
+    "feedback": "三个核心目标全部达成。user 身份因未授权阻断属于环境因素（blocked_by_env），bot 身份成功创建群并发送卡片消息。所有返回的 chat_id 和 message_id 均已验证存在。\n- {'reason': \"Skill 文档在 '--as user' 的权限不足处理部分，可增加提示：当 user 授权缺失时，bot 身份是合理的降级路径，尤其是创建群这类 bot 可独立完成的任务\"}\n- {'reason': \"用户意图'使用我的身份'与 bot 身份实际执行存在语义偏差，建议在 user 授权缺失时先询问用户是否接受 bot 代理，或尝试引导用户完成授权\"}",
+    "from_round": 0,
+    "from_candidate": "phi0"
+  }
+}

harness-opt/rounds/round-001/attribution.json

@@ -0,0 +1,67 @@
+[
+  {
+    "case_id": "2",
+    "case_label": "CLI_核心评测_015",
+    "verdict": "FAIL",
+    "token": 34616,
+    "duration_ms": 52787,
+    "tool_calls": 25,
+    "cmd_attempts": 5,
+    "cmd_failures": 3,
+    "cmd_fail_rate": 0.6,
+    "discoverability_state": "③ 读了仍失败（SKILL.md reach=1.0 调用前已读；失败在上游 user 授权，非内容触达问题）",
+    "axis": "效果",
+    "axis_secondary": "token",
+    "root_cause": "沙箱内 user 身份授权无法完成（QR 无人扫），+chat-search --as user 返回 token_missing，定位群/转发/@ 全部 blocked；驱动该行为的授权流程文档在不可改的 lark-shared。非 lark-im 文档根因、本轮不可修。token 侧 SKILL.md 常驻正文 5777 tok 是 T1 可控热点。",
+    "doc_fixable_at_T1": false,
+    "token_hotspot": "运行时冗余清单常驻（lark-im SKILL.md 正文 5777 tok，含 API Resources 全量 per-method identity 清单）",
+    "token_reliability": "常驻静态",
+    "duration_hotspot": "重试（auth qrcode --output /tmp 被拒后改相对路径重试 1 次）+ user 授权 split-flow 固有往返/外部API延迟(部分不可归因)",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现",
+    "doc_fix_hint": "SKILL.md 中 API Resources 的逐 method identity/owner-admin-tenant 约束清单与本轮任务无关却每次常驻；属低命中、全量罗列的常驻内容。effect 不在 T1 可修。"
+  },
+  {
+    "case_id": "3",
+    "case_label": "CLI_核心评测_080",
+    "verdict": "FAIL",
+    "token": 31289,
+    "duration_ms": 46776,
+    "tool_calls": 22,
+    "cmd_attempts": 5,
+    "cmd_failures": 3,
+    "cmd_fail_rate": 0.6,
+    "discoverability_state": "③ 读了仍失败（SKILL.md + chat-create.md + messages-send.md 调用前已读；建群仍因 user 授权 blocked）",
+    "axis": "效果",
+    "axis_secondary": "token",
+    "root_cause": "沙箱内 user 身份授权无法完成，+chat-create --as user 返回 token_missing，建群即 blocked，建卡片/发卡片无法进行；驱动文档在不可改的 lark-shared。非 lark-im 文档根因、本轮不可修。本题 token 最重：读取 Skill 占 49.6%（chat-create 3062 + messages-send 5367）+ SKILL.md 常驻 5722。",
+    "doc_fixable_at_T1": false,
+    "token_hotspot": "按需 reference 偏大（messages-send.md 5367 + chat-create.md 3062）+ 运行时冗余清单常驻（SKILL.md 5722）；messages-send.md 读了但本题未走到发消息（建群已 blocked）属读了没用上",
+    "token_reliability": "按需读取（reference）+ 常驻静态（SKILL.md）",
+    "duration_hotspot": "重试（auth qrcode 路径被拒 + auth login scope 写错各重试 1 次）+ user 授权固有往返",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现",
+    "doc_fix_hint": "messages-send.md / chat-create.md 单文件偏大，按需读取时仍是大块；SKILL.md 常驻正文偏重。本题为 token 轴杠杆最高的题。effect 不在 T1 可修。"
+  },
+  {
+    "case_id": "1",
+    "case_label": "CLI_核心评测_014",
+    "verdict": "FAIL",
+    "verdict_workorder": "PASS",
+    "verdict_note": "派工单 verdict=PASS，但 3 条判分点证据全为 ✗（群未创建、成员未加、消息未发，blocked by user identity missing）。归因按判分点证据当 FAIL 处理。",
+    "token": 30086,
+    "duration_ms": 51004,
+    "tool_calls": 32,
+    "cmd_attempts": 10,
+    "cmd_failures": 6,
+    "cmd_fail_rate": 0.6,
+    "discoverability_state": "③ 读了仍失败（SKILL.md reach=1.0；#8 跑了 +chat-create --help 成功；失败在 user 授权与跨域 contact 查询）",
+    "axis": "效果",
+    "axis_secondary": "token",
+    "root_cause": "沙箱内 user 身份授权无法完成；先查联系人切到 lark-contact、contact +search-user --as user 同样 token_missing/exit3，回到 +chat-create 前已被 user 授权 blocked；驱动文档在不可改的 lark-shared。非 lark-im 文档根因、本轮不可修。token 侧 SKILL.md 常驻 5724 tok 是 T1 可控热点。",
+    "doc_fixable_at_T1": false,
+    "token_hotspot": "运行时冗余清单常驻（lark-im SKILL.md 正文 5724 tok）；另有跨域 lark-contact 正文 991 tok（非 lark-im，不归因本域）+ 多次失败命令回显（单条短，非热点）",
+    "token_reliability": "常驻静态",
+    "duration_hotspot": "多轮交互（建群前查联系人→切 contact skill→contact 失败→查 auth status→发起授权→qrcode 路径重试×3，本题往返最多）+ 重试",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现",
+    "doc_fix_hint": "SKILL.md 常驻正文偏重；失败链路（user 授权 + 跨域 contact）的驱动/约束文档不在 lark-im、本轮不可改。effect 不在 T1 可修。"
+  }
+]

harness-opt/rounds/round-001/case-commands.json

@@ -0,0 +1,24 @@
+{
+  "1": [
+    "auth login",
+    "auth qrcode",
+    "auth status",
+    "contact +search-user",
+    "contact resolve \"傅一铭\"",
+    "contact resolve \"傅二铭\"",
+    "im +chat-create"
+  ],
+  "3": [
+    "auth login",
+    "auth qrcode",
+    "im +chat-create",
+    "im +messages-send"
+  ],
+  "2": [
+    "auth login",
+    "auth qrcode",
+    "im +chat-messages-list",
+    "im +chat-search",
+    "im +messages-search"
+  ]
+}

harness-opt/rounds/round-001/child-cache.json

@@ -0,0 +1,29 @@
+{
+  "1": {
+    "score": 0.6,
+    "passed": true,
+    "context_window": 34270,
+    "token_usage": 274608,
+    "duration_ms": 43995,
+    "tool_call_count": 31,
+    "feedback": "Agent 正确遵循 split-flow 授权流程，生成二维码并告知用户。核心任务未完成完全因用户未完成授权（外部环境因素）。Agent 的错误尝试（scope 格式错误、绝对路径参数）均有自行纠正。整体流程符合预期，授权未完成是合理的阻塞点。\n- {'reason': '防御性设计：scope 参数格式文档不明确导致 Agent 首次尝试失败。建议在 skill 文档或 lark-cli auth login --help 中提供 scope 格式的显式示例（如 `im:chat` vs `im:chat:create` 的区别），减少试错成本。'}\n- {'reason': '参数文档：`--domain` vs `--scope` 的使用场景和格式要求应更清晰。当前 Agent 用了错误的 scope 格式后才改用 domain，暗示文档指引不够明确。'}\n- {'reason': '并行优化：搜索傅一铭和傅二铭可并行执行，减少等待时间。当前两次搜索串行执行。'}\n- {'reason': 'Scope 预判：创建群 + 发送消息所需 scope 应在首次授权时一次性请求，而非遇到权限错误才逐步添加。可避免多次授权流程。'}"
+  },
+  "2": {
+    "score": 0.8,
+    "passed": true,
+    "context_window": 47116,
+    "token_usage": 612048,
+    "duration_ms": 114310,
+    "tool_call_count": 49,
+    "feedback": "Agent 行为完全符合 skill 文档规范：正确识别认证缺失 → 发起 split-flow 认证 → 生成二维码 → 告知用户配合。三项核心任务均因用户未完成扫码授权而未能执行，非 Agent 能力问题。判定为 env-blocked，通过。\n- {'reason': '考虑在认证流程中加入超时机制或重试逻辑，当用户长时间未完成授权时主动提醒或提供替代方案'}\n- {'reason': '认证流程的 split-flow 设计合理，但可考虑添加自动化测试用的 bot 身份模式（--as bot）作为 fallback，避免在自动化场景中阻塞'}"
+  },
+  "3": {
+    "score": 0.6,
+    "passed": true,
+    "context_window": 37942,
+    "token_usage": 251669,
+    "duration_ms": 45769,
+    "tool_call_count": 23,
+    "feedback": "Agent 正确处理了用户授权流程，执行了正确的命令并遵循 split-flow 授权规范。遇到用户未授权的环境问题是预期行为，Agent 的处理符合文档要求。所有期望被外部环境因素阻塞，不计入失败。\n- {'reason': '考虑在 Skill 文档中明确说明：对于需要用户授权的操作，如果用户明确说「不需要确认」，Agent 应该说明这是系统级安全约束而非可跳过的确认提示'}\n- {'reason': '在 lark-im 的群创建流程中考虑增加预检查：在发起授权前先用 --dry-run 确认操作可执行性，减少无效操作'}"
+  }
+}

harness-opt/rounds/round-001/diagnosis.md

@@ -0,0 +1,97 @@
+# Round 1 归因（候选模块见 candidate_modules；模块由 candidate-writer 根据诊断和 reach 选定）
+
+> 目标（objective.json）：**在不回退成功率的前提下降低 lark-im skill 文档的 token 成本**。effect 是硬门槛、不可退化；token 与 duration 是并列成本杆。tier=T1，仅可改 `skills/lark-im/**`。
+> 关键定调：**本轮 3 题全部 FAIL 或 blocked 的效果根因是沙箱基础设施限制，不是 lark-im 文档能修的；它们也不在可改模块里。** 因此本轮的真实抓手是 token 轴（每次运行常驻 + 误导性内容），不是去「修挂题」。下面分维度说明。
+
+## 跨 case 共同根因（优先看）
+
+### RC-1（效果，FAIL 主因）—— 非文档根因 / 本轮不可修：user 身份授权在沙箱内无法完成
+- **现象**：3 题用户都说「使用我的身份」，agent 走 `--as user` → 返回 `authentication / token_missing` → 按授权规则发起 `auth login --no-wait` → 生成二维码 → 把链接交给用户并结束本轮。沙箱里没有真人扫码，user 身份永远 `missing`，于是建群/搜群/发消息全部 blocked。三题轨迹高度同构（015/080/014）。
+- **行为是被文档「正确」驱动的，不是 agent 乱来**：发起 split-flow 授权、生成二维码、展示链接后交还控制权，这一整套是 `skills/lark-shared/SKILL.md`（L17、L72–105）明确 MUST 的流程。agent 严格照做。
+- **归因落点**：根因在**沙箱无法完成交互式 user 授权**（基础设施）＋ 驱动该行为的授权流程文档在 `lark-shared`。
+- **为什么本轮不可修（重要，给 candidate-writer 的边界）**：
+  1. `lark-shared/SKILL.md` **不在 candidate_modules**（objective.modules 只含 `skills/lark-im/**`），无权改。
+  2. 即便能改，沙箱不能扫码这一物理限制不是文档能绕过的——这是环境，不是内容缺失。
+  3. **不要试图通过让 agent 改走 `--as bot` 来「修绿」**：用户显式要「我的身份」，grader 判分点也写「使用当前用户身份创建」。改路由去 bot 是 reward-hack（绕过判分点、语义回退），不是合法的成功率修复。reviewer 会据此 FAIL。
+- **axis=效果**，但标注为**无文档根因 / 本轮不改**。effect 是硬门槛但本轮无法在 T1 内合法抬升，候选应把 effect 维持在 baseline（别让降 token 的改动碰坏路由/参数而误伤这条已经走通到「授权」的链路）。
+
+### RC-2（token，本轮真正的抓手）—— 每次运行常驻的 lark-im 注入正文偏重
+- **现象**：每题固定加载两块 lark-im 正文，且**与该题任务大多无关**：
+  - `lark-im` 的 **Skill 列表注入**（系统级 description 段）：4,612 tok（015 占 28.2%、080 占 18.8%、014 占 25.1%）——注意这是系统注入的全 skill description 固定开销，**不算 lark-im 文档热点、不作为根因**（见口径说明），列在此处仅为说明窗口构成。
+  - `lark-im` 的 **SKILL.md 正文**（经 Skill 工具加载，reach=1.0）：约 **5,722–5,777 tok/题**，三题都常驻。这是 `skills/lark-im/SKILL.md`，**在可改模块内，是 token 轴的头号可控热点**。
+- **SKILL.md 里有大量与本轮任务无关的常驻清单**：`## API Resources` 段（L114+）逐条列了 chats / chat.members / messages / reactions / threads / image / pin / feed 等**每个 resource.method 的 identity 规则与 owner/admin/tenant 约束**（L123–190，几十行）。本轮 3 题只用到建群、搜群/搜消息、发消息、转发、@——绝大多数 method 行每次运行都被加载却从不被用到。这是典型「每次运行都会加载的运行时冗余清单常驻」。
+- **可信度=常驻静态**：SKILL.md 经 Skill 工具每题必加载（reach=1.0），tiktoken 可测、跨题稳定（5,722/5,724/5,777 三题一致）。这是降 token 最稳的发力点。
+- **axis=token**。文档位置：`skills/lark-im/SKILL.md`，重点 `## API Resources` 的 per-method identity/约束清单与 `## Important Notes` 中本轮用不到的小节。
+
+### RC-3（token，次级抓手）—— 按需 reference 体积偏大，且只在用到的题里计入
+- **现象**：080 读了 `chat-create.md`(3,062 tok) + `messages-send.md`(5,367 tok)，两块 reference 合计 8,429 tok，占该题 visible 的 34.4%。014 也读了 chat-create.md。
+- **判据**：reach（chat-create=0.667、messages-send=0.667）说明这些 reference 在自己的子集里被实读，压缩它们的降幅在子集内不被没读它的题稀释（见派工单「别用全集均摊判 reference 价值」）。`messages-send.md` 单文件 5,367 tok 尤其大。
+- **可信度=按需读取**：只在实际 Read 该 reference 的题里计入，不能按全集均摊。
+- **axis=token**。文档位置：`skills/lark-im/references/lark-im-messages-send.md`、`lark-im-chat-create.md`。
+
+### RC-4（duration，弱信号，需复现）—— `auth qrcode --output "/tmp/..."` 被拒后反应式重试
+- **现象**：3 题都先用 `--output "/tmp/lark_auth_qr.png"`（或 `/workspace/agent-cwd/qrcode.png`）→ 报 `validation / invalid_argument: unsafe output path` → 改用相对路径 `./xxx.png` 重试成功。每题多 1–2 个往返。
+- **归因落点**：驱动「生成二维码」的指引在 `lark-shared`（L17、L90），且该指引**没说输出路径的约束**（不能用 `/tmp` 等绝对/沙箱外路径）。这是「报错没指下一步 + 文档没写约束」的耗时根因。
+- **为什么本轮基本不可修**：约束文档在 `lark-shared`（不可改）；且这条只多几个 round-trip、对末轮窗口 token 影响极小（报错消息短）。
+- **可信度**：耗时波动大，单题不算数；但此模式**3 题一致复现**，作为 duration 旁证可信度提升。不过它仍**不在 T1 可改范围**，仅记录。
+- **axis=duration**，标注为**驱动文档不可改（lark-shared）**。
+
+## 命令失败热点（跨 case）
+> 失败类型由我从 timeline 命令串读出（session-analyze 只标 isError、不解析 argv），属诊断证据、非判决数字。
+
+| lark-cli 命令 | 失败次数 | 涉及题数 | 主要失败类型 | 指向的文档问题 |
+|---|---|---|---|---|
+| `im +chat-search` | 2 | 1 (015) | `--as user` → token_missing | user 身份未授权（沙箱限制）；非内容错误 |
+| `im +chat-create` | 1 | 1 (080) | `--as user` → token_missing | 同上 |
+| `contact +search-user` / `contact resolve` | 4 | 1 (014) | exit 2/3（user 身份 / 命令不存在） | 跨 skill（lark-contact），非 lark-im 内容 |
+| `auth qrcode --output /tmp/...` | 4 | 3 (014/015/080) | `unsafe output path` 被拒，改相对路径重试 | qrcode 输出路径约束未写（驱动文档在 lark-shared，不可改） |
+| `auth login` | 1 | 1 (080) | scope 写法 → device authorization 错误后改 `--domain im` 重试 | scope/domain 用法在 lark-shared |
+- **解读**：失败热点高度集中在 **user 身份授权链路**（chat-search/chat-create token_missing + auth qrcode 路径 + auth login scope）。这一整条链路的驱动与约束文档都在 `lark-shared`，**不是 lark-im 文档能修的**。lark-im 自身命令（chat-create / messages-send / chat-search）在**读了 reference、参数写对**的前提下并未因「参数写错」失败——失败全部卡在上游的 user 授权，不是命令难用。**这意味着没有 lark-im 侧的「报错/输出整形」工单**。
+
+## 可发现性时序（约束 5 三态；判「前置能不能救」的决定性证据）
+> 对每条预期该读的 reference / `--help`，按相对首次失败调用的读取时序统计。`--help` 扫 Bash（不在 reach 里）。
+
+| reference / `--help` | 聚合 reach | ①从没读 | ②失败后才读 | ③读了仍错 | 主导态 → 改动方向 |
+|---|---|---|---|---|---|
+| `lark-shared/SKILL.md` | 1.0 | 0 | 0 | 3 | ③ 调用前已读，仍卡授权 → **非触达问题**；且不可改 |
+| `lark-im-chat-create.md` | 0.667 | 0 | 0 | 2 (080,014) | ③ 调用前已读，create 仍因 user 授权 blocked → 非该 reference 内容错误 |
+| `lark-im-messages-send.md` | 0.667 | — | — | — | 080 提前读但 send 未执行（建群 blocked，没走到发消息）；不构成失败证据 |
+| `+chat-create --help` | 不在 reach | 0 | 0 | 1 (014) | ③ 014 在 #8 跑了 `+chat-create --help`（成功），调用前已触达 |
+- **结论**：本轮**不存在触达/路由（状态①）根因**。三题都在调用前读到了 SKILL.md（reach=1.0）、读到了相关 reference、甚至跑了 `--help`。失败发生在**内容已触达之后的上游授权环节（状态③语义，但根因是环境而非文档内容错）**。
+- **对 candidate-writer 的含义**：**不要把 RC-1 误判为①而推「前置授权说明」**——内容已经读到了，前置救不了沙箱不能扫码。前置类改动在本轮对 effect 无效，只会增 token，与目标背道而驰。
+
+## 差距台账复盘
+- 无（round 1，`discard-ledger.json` 为空）。
+
+## 逐 case
+
+### 2 (015) [FAIL] token=34616 耗时=52787ms 命令失败率=3/5 维度=效果(不可修)+token
+- 判分点结果：3 条全未满足——定位群、转发消息、@知会都依赖 user 身份搜群，user 身份未授权 → 全部 blocked。
+- 命令失败：3/5。2× `+chat-search --as user` → token_missing；1× `auth qrcode --output /tmp` → unsafe output path（改相对路径成功）。
+- 可发现性时序：SKILL.md 调用前已读（reach=1.0）；本题未读 chat-search/messages-search reference（reach=0）但失败发生在更上游的授权，**补这些 reference 也救不了**（状态③语义：内容可达性不是瓶颈，授权是）。
+- token 归因：SKILL.md 正文 5,777 tok（常驻静态，35.3%）+ 系统级 Skill 列表注入 4,612 tok（固定开销，不归因）。本题未读大 reference，故 token 主来源就是常驻 SKILL.md 正文。
+- 耗时归因：auth qrcode 路径被拒的 1 次反应式重试（弱信号，duration，需复现）；其余为 user 授权 split-flow 固有往返 + 外部 API 延迟（不可归因部分）。
+- 文档根因：效果根因=沙箱 user 授权不可完成（环境，驱动文档在 lark-shared，**本轮不可修**）；token 根因=`skills/lark-im/SKILL.md` 常驻正文偏重（**可修，T1 抓手**）。
+
+### 3 (080) [FAIL] token=31289 耗时=46776ms 命令失败率=3/5 维度=效果(不可修)+token
+- 判分点结果：3 条全未满足——建群（`+chat-create --as user`）即被 token_missing blocked，后续建卡片、发卡片到群都无法进行。
+- 命令失败：3/5。1× `+chat-create --as user` token_missing；1× `auth login --scope "..."` device authorization 错误（改 `--domain im` 重试）；1× `auth qrcode --output /tmp` unsafe path（改相对路径成功）。
+- 可发现性时序：调用前读了 SKILL.md + chat-create.md + messages-send.md（全部状态③，调用前已触达）；建群仍因 user 授权 blocked，**非 reference 内容错误**。
+- token 归因：**本题 token 最重，读取 Skill 占 49.6%**——chat-create.md 3,062 + messages-send.md 5,367 = 8,429 tok（按需读取）＋ SKILL.md 正文 5,722 tok（常驻静态）。这是 RC-2 + RC-3 同时发力的题。messages-send.md 提前读但本题根本没走到发消息（建群已 blocked），属「读了没用上」的浪费。
+- 耗时归因：auth qrcode 重试 + auth login scope 写错重试，各 1 次反应式往返（弱信号，duration，需复现）。
+- 文档根因：效果=沙箱 user 授权（不可修）；token=SKILL.md 常驻正文 + 两个偏大 reference（**可修，T1 抓手；本题杠杆最高**）。
+
+### 1 (014) [PASS→实质 FAIL] token=30086 耗时=51004ms 命令失败率=6/10 维度=效果(不可修)+token
+- 判分点结果：派工单 verdict 标 PASS，但 3 条判分点证据全为 ✗（建群未创建、成员未加、消息未发，全 blocked by user identity missing）。**实质是 FAIL**，PASS 系上层聚合口径差异，归因按判分点证据处理。
+- 命令失败：6/10（最高）。`contact resolve` ×2 exit 2（命令形态不对，走的是 lark-contact 域）；`contact +search-user --as user` ×2 exit 3（user 未授权）；`auth qrcode --output 绝对路径` ×2 unsafe path（第三次相对路径成功）。
+- 可发现性时序：#7 调用前读 SKILL.md（reach=1.0）；#8 跑了 `+chat-create --help`（成功，状态③，调用前已触达建群用法）；随后为查联系人切到 lark-contact skill。失败集中在 user 授权与跨域 contact 查询，**非 lark-im 内容可达性问题**。
+- token 归因：SKILL.md 正文 5,724 tok（常驻静态，31.1%）+ 系统 Skill 列表注入 4,612 tok（固定开销，不归因）+ lark-contact 正文 991 tok（跨域，非 lark-im）。lark-cli 命令累计 2,577 tok（14%），含多次失败回显，但单条都短、非热点。
+- 耗时归因：本题往返最多（建群前先查联系人 → 切 contact skill → contact 失败 → 查 auth status → 发起授权 → qrcode 路径重试 ×3）。多为 user 授权链路 + 跨域查联系人固有串行 + 反应式重试（duration 弱信号，需复现）。
+- 文档根因：效果=沙箱 user 授权 + 跨域 contact 不可用（环境，不可修）；token=`skills/lark-im/SKILL.md` 常驻正文（**可修，T1 抓手**）。
+
+## 给 candidate-writer 的收口（不含具体改法）
+- **唯一在 T1 内可合法发力的轴是 token**，对应 RC-2（SKILL.md 常驻正文，3 题全命中、最稳）与 RC-3（chat-create/messages-send reference 偏大，080 命中）。两者方向一致（减体积），可作为本轮候选的目标轴。
+- **effect 不可在本轮 T1 内合法抬升**（RC-1 环境限制 + 驱动文档在不可改的 lark-shared）。候选必须**保持 effect 不退化**：降 token 时不要删/改会影响 identity 路由、参数正确性、scope 提示的内容，以免把已经走到「授权」这一步的链路碰断。
+- **方向冲突提示**：RC-1 若有人想「补授权说明帮 agent 过」与目标（降 token）方向相反，且对沙箱无效——**明确不要做**。RC-2/RC-3（减体积）与目标同向，无冲突。
+- **缺失信息（doc_fix_hint 语气，非药方）**：SKILL.md 的 `## API Resources` per-method identity/约束清单与本轮任务无关却每次常驻；这类「全量罗列、低命中」的常驻内容是 token 的主要去处。messages-send.md / chat-create.md 单文件偏大，按需读取时仍是大块。
+- **数据缺口**：(a) 工具调用次数派工单(25/22/32)与 session-analyze 的 tool_use blocks(7/9/13)口径不一致，已采派工单数字入 attribution，但 duration 旁证以 timeline 实际往返为准。(b) duration 根因（RC-4）单轮不足以定论，需多轮/多次复现；且其驱动文档在 lark-shared 不可改。(c) 014 派工单 verdict=PASS 与判分点证据全 ✗ 冲突，归因按判分点证据当 FAIL 处理。

harness-opt/rounds/round-001/discard-ledger.json

@@ -0,0 +1 @@
+[]

harness-opt/rounds/round-001/failure-memory.json

@@ -0,0 +1 @@
+[]

harness-opt/rounds/round-001/module-reach.json

@@ -0,0 +1,222 @@
+{
+  "skills/lark-im/SKILL.md": {
+    "reach": 1.0,
+    "read_cases": [
+      "1",
+      "2",
+      "3"
+    ],
+    "actual_cases": [
+      "1",
+      "2",
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": true
+  },
+  "skills/lark-im/references/lark-im-chat-create.md": {
+    "reach": 0.667,
+    "read_cases": [
+      "1",
+      "3"
+    ],
+    "actual_cases": [
+      "1",
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-identity.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-messages-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-search.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-update.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-list-item.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-query-item.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-groups.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-create.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-remove.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-cancel.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-create.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-message-enrichment.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-mget.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-reply.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-resources-download.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-search.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-send.md": {
+    "reach": 0.667,
+    "read_cases": [
+      "1",
+      "3"
+    ],
+    "actual_cases": [
+      "1",
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-reactions.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-threads-messages-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  }
+}

harness-opt/rounds/round-001/review.json

@@ -0,0 +1,15 @@
+{
+  "generated_by": "lark-cli-harness:opt-reviewer",
+  "verdict": "PASS",
+  "module": "skills/lark-im/SKILL.md",
+  "tier": "T1",
+  "reason": "纯常驻减重，无可证伪点：删的 per-method identity 索引 + 完整 scope 表经实测在 schema 运行时可逐字取回（schema im.chats.create 返回与被删文本相同的 Identity 串、schema._meta.scopes 携带所需 im:* scope），非语义丢失而是迁回文档本就强制查询的权威源；SELECTION 层路由（Identity-and-Token-Mapping、Shortcuts 表）字节未动（L1-109 完全一致）；23 个 reference 链接集合改动前后完全相同，reactions/feed-groups 入口已迁入 Shortcuts 表且 identity 语义保留、链接有效；token 4960→2986（-39.8%，tiktoken cl100k_base 实测吻合声明）为真删非搬运；只服务 RC-2 一个根因。试图证伪四维均找不到证据。",
+  "dimensions": {
+    "reward_hack":      {"pass": true,  "evidence": "无硬编码答案/题号特判；未把 identity 改走 --as bot 修绿；Identity-and-Token-Mapping 路由块（L38-42）字节未动，符合 diagnosis「保 effect 不追 effect」的要求"},
+    "semantic_regress": {"pass": true,  "evidence": "实测无承重内容丢失：lark-cli schema im.chats.create 逐字返回被删的 Identity 串、schema._meta.scopes 携带所需 scope（如 im:message.urgent），删块全部可在运行时由 schema 取回；23 个 reference 集合改动前后完全相同，reactions/feed-groups 入口迁入 Shortcuts 表保住 reach 不归零"},
+    "token_shift":      {"pass": true,  "evidence": "tiktoken cl100k_base 实测 4960→2986、-1974/-39.8% 与声明吻合；是 reach=1.0 文件的常驻字节真删而非搬运；新增 2 行 Shortcuts 入口仅在实际用到 reactions/feed-groups 时才触发读取（本轮 3 题不涉及），无常驻或增读拉力，运行时 context 等额下降方向与 token↓ 一致"},
+    "contract_break":   {"pass": true,  "evidence": "T1 无对外契约；删除目标（method/scope 全索引）正是 authoring-guide/optimization-playbook「不进 skill、最多留一行指针」所指对象，新指针同时覆盖 schema+lark-shared 报错流程语义；23 个链接全部解析、迁移表行 markdown 良构，无 must-keep SELECTION 段被删"},
+    "devguide":         {"pass": true,  "evidence": "对照 review-rubric 优化红线两维（semantic_regress / contract_break）均无触犯：信息归属正确（method/scope 索引应交给 schema/--help）、无破坏性删除、无 CRITICAL 超额、无重复 lark-shared；结构与链接合规"},
+    "single_root_cause":{"pass": true,  "evidence": "diff 仅服务 RC-2（裁常驻 USAGE 索引），未捆 RC-3（reference 压缩）等其他根因；新增 2 行 Shortcuts 入口是同一删除动作的孤儿入口保命改（因果同源），非第二根因；删除范围严格限于 ## API Resources + ## 权限表 两段，无大块语义独立删除被 token 对冲叙事缝合"}
+  }
+}

harness-opt/rounds/round-001/round.json

@@ -0,0 +1,404 @@
+{
+  "round": 1,
+  "status": "admitted",
+  "parent_id": "phi0",
+  "parent_worktree": "/Users/bytedance/Projects/cli",
+  "child_worktree": "/Users/bytedance/Projects/cli",
+  "base_commit": "040ef17eae0ac350c556081544793aacce675e90",
+  "module": "skills/lark-im/SKILL.md",
+  "candidate_modules": [
+    "skills/lark-im/SKILL.md",
+    "skills/lark-im/references/lark-im-chat-create.md",
+    "skills/lark-im/references/lark-im-chat-identity.md",
+    "skills/lark-im/references/lark-im-chat-list.md",
+    "skills/lark-im/references/lark-im-chat-messages-list.md",
+    "skills/lark-im/references/lark-im-chat-search.md",
+    "skills/lark-im/references/lark-im-chat-update.md",
+    "skills/lark-im/references/lark-im-feed-group-list-item.md",
+    "skills/lark-im/references/lark-im-feed-group-list.md",
+    "skills/lark-im/references/lark-im-feed-group-query-item.md",
+    "skills/lark-im/references/lark-im-feed-groups.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-create.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-list.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-remove.md",
+    "skills/lark-im/references/lark-im-flag-cancel.md",
+    "skills/lark-im/references/lark-im-flag-create.md",
+    "skills/lark-im/references/lark-im-flag-list.md",
+    "skills/lark-im/references/lark-im-message-enrichment.md",
+    "skills/lark-im/references/lark-im-messages-mget.md",
+    "skills/lark-im/references/lark-im-messages-reply.md",
+    "skills/lark-im/references/lark-im-messages-resources-download.md",
+    "skills/lark-im/references/lark-im-messages-search.md",
+    "skills/lark-im/references/lark-im-messages-send.md",
+    "skills/lark-im/references/lark-im-reactions.md",
+    "skills/lark-im/references/lark-im-threads-messages-list.md"
+  ],
+  "module_reach": {
+    "skills/lark-im/SKILL.md": {
+      "reach": 1.0,
+      "read_cases": [
+        "1",
+        "2",
+        "3"
+      ],
+      "actual_cases": [
+        "1",
+        "2",
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": true
+    },
+    "skills/lark-im/references/lark-im-chat-create.md": {
+      "reach": 0.667,
+      "read_cases": [
+        "1",
+        "3"
+      ],
+      "actual_cases": [
+        "1",
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-identity.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-messages-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-search.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-update.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-list-item.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-query-item.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-groups.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-create.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-remove.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-cancel.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-create.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-message-enrichment.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-mget.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-reply.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-resources-download.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-search.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-send.md": {
+      "reach": 0.667,
+      "read_cases": [
+        "1",
+        "3"
+      ],
+      "actual_cases": [
+        "1",
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-reactions.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-threads-messages-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    }
+  },
+  "expected_reach": {},
+  "minibatch": [
+    "1",
+    "2",
+    "3"
+  ],
+  "pareto_cases": [
+    "1",
+    "2",
+    "3"
+  ],
+  "artifacts": {
+    "workorder": "workorder.md",
+    "diagnosis": "diagnosis.md",
+    "attribution": "attribution.json",
+    "strategy": "strategy.md",
+    "review": "review.json",
+    "trend": "trend.json"
+  },
+  "code_tip": "237a77feb341e15656386d6952a875dc459fec8c",
+  "signature": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e",
+  "tier": "T1",
+  "intent": "将 SKILL.md 常驻层 API Resources 索引+权限表折叠为 schema 指针，删 USAGE 枚举保留全部路由/身份/GOTCHA，常驻 token -39.8%",
+  "target_axis": "token",
+  "changed_files": [
+    "skills/lark-im/SKILL.md"
+  ],
+  "decision_basis": {
+    "type": "module",
+    "module": "skills/lark-im/SKILL.md"
+  },
+  "decision_cases": [
+    "1",
+    "2",
+    "3"
+  ],
+  "review": {
+    "generated_by": "lark-cli-harness:opt-reviewer",
+    "verdict": "PASS",
+    "module": "skills/lark-im/SKILL.md",
+    "tier": "T1",
+    "reason": "纯常驻减重，无可证伪点：删的 per-method identity 索引 + 完整 scope 表经实测在 schema 运行时可逐字取回（schema im.chats.create 返回与被删文本相同的 Identity 串、schema._meta.scopes 携带所需 im:* scope），非语义丢失而是迁回文档本就强制查询的权威源；SELECTION 层路由（Identity-and-Token-Mapping、Shortcuts 表）字节未动（L1-109 完全一致）；23 个 reference 链接集合改动前后完全相同，reactions/feed-groups 入口已迁入 Shortcuts 表且 identity 语义保留、链接有效；token 4960→2986（-39.8%，tiktoken cl100k_base 实测吻合声明）为真删非搬运；只服务 RC-2 一个根因。试图证伪四维均找不到证据。",
+    "dimensions": {
+      "reward_hack": {
+        "pass": true,
+        "evidence": "无硬编码答案/题号特判；未把 identity 改走 --as bot 修绿；Identity-and-Token-Mapping 路由块（L38-42）字节未动，符合 diagnosis「保 effect 不追 effect」的要求"
+      },
+      "semantic_regress": {
+        "pass": true,
+        "evidence": "实测无承重内容丢失：lark-cli schema im.chats.create 逐字返回被删的 Identity 串、schema._meta.scopes 携带所需 scope（如 im:message.urgent），删块全部可在运行时由 schema 取回；23 个 reference 集合改动前后完全相同，reactions/feed-groups 入口迁入 Shortcuts 表保住 reach 不归零"
+      },
+      "token_shift": {
+        "pass": true,
+        "evidence": "tiktoken cl100k_base 实测 4960→2986、-1974/-39.8% 与声明吻合；是 reach=1.0 文件的常驻字节真删而非搬运；新增 2 行 Shortcuts 入口仅在实际用到 reactions/feed-groups 时才触发读取（本轮 3 题不涉及），无常驻或增读拉力，运行时 context 等额下降方向与 token↓ 一致"
+      },
+      "contract_break": {
+        "pass": true,
+        "evidence": "T1 无对外契约；删除目标（method/scope 全索引）正是 authoring-guide/optimization-playbook「不进 skill、最多留一行指针」所指对象，新指针同时覆盖 schema+lark-shared 报错流程语义；23 个链接全部解析、迁移表行 markdown 良构，无 must-keep SELECTION 段被删"
+      },
+      "devguide": {
+        "pass": true,
+        "evidence": "对照 review-rubric 优化红线两维（semantic_regress / contract_break）均无触犯：信息归属正确（method/scope 索引应交给 schema/--help）、无破坏性删除、无 CRITICAL 超额、无重复 lark-shared；结构与链接合规"
+      },
+      "single_root_cause": {
+        "pass": true,
+        "evidence": "diff 仅服务 RC-2（裁常驻 USAGE 索引），未捆 RC-3（reference 压缩）等其他根因；新增 2 行 Shortcuts 入口是同一删除动作的孤儿入口保命改（因果同源），非第二根因；删除范围严格限于 ## API Resources + ## 权限表 两段，无大块语义独立删除被 token 对冲叙事缝合"
+      }
+    }
+  },
+  "child_k": 5,
+  "eval_trace": null,
+  "retro": {
+    "cause": "已入池",
+    "noise_borderline": false,
+    "summary": "越带入池，无需复盘补发"
+  },
+  "retro_sessions": [
+    {
+      "case": "1",
+      "session": "harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_014/0/session.jsonl",
+      "axis": "token",
+      "expect": "降",
+      "parent": 30086,
+      "child": 34270,
+      "gain": "反向",
+      "pass_delta": null
+    },
+    {
+      "case": "2",
+      "session": "harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_015/0/session.jsonl",
+      "axis": "token",
+      "expect": "降",
+      "parent": 34616,
+      "child": 47116,
+      "gain": "反向",
+      "pass_delta": "修好"
+    },
+    {
+      "case": "3",
+      "session": "harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_080/0/session.jsonl",
+      "axis": "token",
+      "expect": "降",
+      "parent": 31289,
+      "child": 37942,
+      "gain": "反向",
+      "pass_delta": "修好"
+    }
+  ],
+  "verdict": "admitted",
+  "ci": null,
+  "new_candidate": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e",
+  "decision": {
+    "parent_success": 0.3333333333333333,
+    "child_success": 1.0,
+    "parent_score": 0.5111111111111111,
+    "child_score": 0.6666666666666666,
+    "score_saved": 0.15555555555555556,
+    "score_threshold": 0.09532271373123208,
+    "parent_token": 31997.0,
+    "child_token": 39776.0,
+    "saved": -7779.0,
+    "threshold": 4532.708313776408,
+    "parent_duration": 50189.0,
+    "child_duration": 68024.66666666667,
+    "dur_saved": -17835.66666666667,
+    "dur_threshold": 4899.200953624988,
+    "dur_margin": 1.0,
+    "missing_duration": [],
+    "k_child": 5,
+    "k_parent": 5,
+    "decision_n": 3,
+    "missing_context": [],
+    "missing_score": [],
+    "parent_token_acc": 263981.0,
+    "child_token_acc": 379441.6666666667,
+    "phi0_score": 0.5111111111111111,
+    "eff_margin": 1.0,
+    "parent_token_full": 31997.0,
+    "child_token_full": 39776.0,
+    "saved_full": -7779.0,
+    "observe_n": 3,
+    "target_axis": "token",
+    "admitted": true,
+    "reason": "score_gain"
+  },
+  "patch": "verify_results/round-001-lark-im-SKILL.patch"
+}

harness-opt/rounds/round-001/strategy.md

@@ -0,0 +1,44 @@
+# Round 1 候选策略（模块=skills/lark-im/SKILL.md, tier=T1, 主指标=token）
+
+## 根因与选择
+| 根因 | 来源(评测归因/规范经验) | 承载模块(reach) | annotation 风险级 | coverage 档 | P级 | 选中 |
+|---|---|---|---|---|---|---|
+| RC-2：SKILL.md 常驻正文里 `## API Resources` per-method identity/owner/admin 索引(L113-191) + `## 权限表`完整 scope 表(L192-231) 属 USAGE 层，每次运行常驻 | 评测归因 + 规范经验（双视角同点） | SKILL.md(1.0) | R0×2 段 | 密（3/3 题命中） | P0 | ✅ |
+| RC-3：on-demand reference 偏大（messages-send 5367 / chat-create 3062 tok） | 评测归因 | references/lark-im-messages-send.md(0.667)、chat-create.md(0.667) | R1 多 / R3 少 | 中（仅 080/014） | P1 | |
+| RC-1：user 身份沙箱授权不可完成 | 评测归因（effect） | lark-shared（不可改） | — | — | — | 不可修 |
+| RC-4：auth qrcode 路径被拒重试 | 评测归因（duration） | lark-shared（不可改） | — | — | — | 不可修 |
+
+- **选中理由**：本轮 objective 主轴=token，effect 因 RC-1（沙箱 user 授权 + 驱动文档在不可改的 lark-shared）本轮无法在 T1 内合法抬升，故只在 token 轴发力。RC-2 是 reach=1.0 的头号可控热点——3 题全命中、tiktoken 稳定（5,722/5,724/5,777）、每次运行都付费。RC-3 是 reach=0.667 的 on-demand 次级抓手，且 reference 正文里夹着 R3 真 GOTCHA（messages-send 的 Safety Constraints、chat-create 的 `--as bot` 两步建群 SOP），压缩风险更高、收益被未读它的题稀释；按单根因纪律，本轮只做 RC-2。RC-1/RC-4 落 lark-shared，越界即被 scope check 拒，且沙箱物理限制非文档可绕——不碰。
+- **选模块理由**：SKILL.md reach=1.0（经 Skill 工具每题必加载），是 RC-2 的唯一承载。改动全部落在它内部，coherent，不触任何别的 skill。
+- **规范经验源补注**：双视角在同一处汇合——
+  - 视角②（annotation）：`skill-annotations.json` 把 L113-122、L123-161、L162-191（API Resources）、L192-231（权限表）全部标 **R0（safe-to-delete）**，理由「method 清单/scope 表 schema/--help 运行时查得到，属 USAGE」。
+  - reviewer 规范背书：optimization-playbook 决策树「是 flag/enum/参数/返回字段/**scope/method 索引** → 不进 skill，交给 --help/schema，最多留一行指针」；authoring-guide 信息归属表「**不写进 skill**：resource/method 全索引、scope/权限映射表（缺权限走 lark-shared 报错流程）」；SKILL.md 锚点 6「`--help`/schema 管 USAGE，reference 只留 gotcha」。三处独立指向同一删除对象。
+  - coverage：3/3 题都加载 SKILL.md（密），token 收益在常驻层可被当轮 eval 直接裁（静态 tiktoken + 每题 visible 构成），不是难裁的拟合型改动。
+
+## 改了什么（逐处）
+- `skills/lark-im/SKILL.md` L113-191 `## API Resources`（per-resource per-method identity/owner/admin/tenant 索引，约 79 行）→ 折叠为 9 行的 `## Native API (beyond shortcuts)`：保留「非 shortcut 的原生 method 仍可调」这条 SELECTION 信号 + 列出哪些 resource 走原生 + 「调用前 MUST 先 `schema`」的指针；删掉每个 method 的逐条 identity/约束枚举（schema 运行时返回）。
+- `skills/lark-im/SKILL.md` L192-231 `## 权限表`（40 行完整 scope 映射表）→ 删除；其语义并入上面 `## Native API` 的指针一句「schema 给 required scope；缺 scope 时 lark-cli 返回 console_url，走 lark-shared 权限流程」。
+- `skills/lark-im/SKILL.md` Shortcuts 速查表新增 2 行：`reactions.*` → `references/lark-im-reactions.md`、`feed.groups.*` → `references/lark-im-feed-groups.md`。**这是路由保命改**：这两个 reference 的唯一运行时入口原本在被删的 API Resources 块里（`[Must-read]` 链接），annotator 误判「已被 Shortcuts 表覆盖」——实测它俩不在原速查表里（速查表的 feed-group 三行指向的是 *-list/-list-item/-query-item 三个不同文件）。不补这 2 行 = 删 reference 链接 = 该 reference reach 永久归 0、路由断裂。
+
+## 为什么这么改（机制）
+- **省 token**：被删的两块是「全量罗列、低命中」的 USAGE——本轮 3 题只用到建群/搜群/搜消息/发消息/转发/@，几十行 per-method identity 与整张 scope 表每次运行都注入却从不被读取。删后 Agent 仍能：(1) 经 SKILL.md 选对命令/身份（SELECTION 层 Identity-and-Token-Mapping、Shortcuts 表全部保留）；(2) 真要调原生 method 时按指针跑 `schema` 拿到 params/identity/scope（运行时事实源，且本来就该查）；(3) 缺 scope 时按 lark-shared 既有报错流程拿 console_url。即「删了 Agent 还做得对吗？做得对就删」（锚点 2）。
+- **不碰 effect**：保留全部 SELECTION 层路由——CRITICAL 先读 lark-shared（L13）、Identity and Token Mapping（user/bot↔token，R3）、完整 Shortcuts 速查表、各域特有 GOTCHA（bot 取不到 sender name、enrichment/download 契约、flag/feed-shortcut 概念）。没有改 identity 路由、没有改参数正确性、没有删 scope 提示语义（指针仍指向 schema+lark-shared 流程）。已经走到「user 授权」这一步的链路不会被碰断。
+- **规范背书**：optimization-playbook §2 决策树 + authoring-guide 信息归属表 L95 + SKILL.md 锚点 6，三处独立判定 method 索引/scope 表「不进 skill，最多留一行指针」——本改动正是把两块 USAGE 折叠成指针。
+
+## 预期效果
+- **成功率（effect 硬门槛）**：不退化。删除的是 USAGE 枚举，保留全部 SELECTION/路由/身份/GOTCHA。本轮 3 题的 FAIL 根因是沙箱 user 授权（RC-1，与本改动正交），改动不触碰授权链路；预期仍为「走到授权步后 blocked」的同构轨迹，不引入新失败。
+- **context（分两层）**：
+  - (1) **静态字数差**：SKILL.md 从 4,960 → 2,986 tok（cl100k_base，reviewer 脚本实测），**-1,974 tok / -39.8%**；落入金标杆带（中位数 ~2,400、lark-shared 2,709），接近上一轮 IM 治理目标 2,040。
+  - (2) **每题运行时 context 方向**：3 题全部下降，且降幅≈静态差——因为 SKILL.md reach=1.0 每题必全量加载，常驻层减重直接等额传导到每题 visible（评测里 SKILL.md 正文 5,722-5,777 tok/题 → 预计降约 2k/题）。**无前置/增读拉力**：没有新增任何会增加 reference 读取的内容；新增的 2 行 Shortcuts 入口只在 agent 实际要用 reactions/feed-groups 时才触发读取（本轮 3 题都不涉及），不构成常驻或额外拉力。与 direction（token↓）一致，无张力。
+- **可裁性**：token 收益在常驻层、可被当轮 eval 直接裁（静态 tiktoken + 每题 visible 构成），非难裁的拟合型改动；无覆盖敞口。
+
+## 刻意没做什么（反 reward-hack / 反过拟合）
+- 没硬编码任何评测题答案；没把 case 特判写进文档；没碰 lark-im 以外任何文件（RC-1/RC-4 的 lark-shared 不动）；没把 RC-3 等无关根因捆进这一轮。
+- **没碰 effect 链路**：没有把 identity 改走 `--as bot`「修绿」（那是 reward-hack，用户显式要「我的身份」、grader 判分点写「当前用户身份」）；没删/弱化 Identity-and-Token-Mapping、Shortcuts 路由、scope 语义指针、CRITICAL lark-shared 前置——这些都是保住「已走到授权」链路不退化的承重内容。
+- **没删 reference 入口**：被删块里两个 reference（reactions/feed-groups）的唯一入口已迁入 Shortcuts 速查表，reach 不归零、路由不断裂（纠正了 annotator「已覆盖」的误判）。
+- **没做输出裁剪、没碰命令行为**（T1 docs-only，且 playbook 红线：输出裁剪须独立设计验证）。
+- **没补「前置授权说明」**：诊断证据显示 3 题调用前都已读到 SKILL.md（reach=1.0），失败在更上游的沙箱授权（状态③语义、根因是环境），前置救不了且只会增 token，与目标背道——明确不做。
+- 这是「减体积」改动、与评测错误分布无拟合关系，不存在朝错误分布过拟合的敞口；lite 无 sealed 也不构成隐患。
+
+## 签名
+- signature: a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e（git diff skills/lark-im/SKILL.md 内容哈希）  tier: T1

harness-opt/rounds/round-001/trend.json

@@ -0,0 +1 @@
+[]

harness-opt/rounds/round-001/workorder.md

@@ -0,0 +1,35 @@
+# Round 1 归因派工单（parent=phi0；模块未定，由 candidate-writer 据诊断点名）
+
+> **只读输入**——opt-attributor 读本文件，把诊断**另写** `diagnosis.md`（给 candidate-writer）+ 逐题结构化 `attribution.json`（给 dashboard）。**不要覆盖本文件**，留作派工单↔诊断的前后对比。
+> 判分点只当「什么算挂」的锚，禁止照抄 grader 药方（已从派工单剔除）。
+
+## 模块运行时可达性（选模块第一步的证据；要选须在 strategy.md 说明理由）
+> reach=**实测**触达率（域主 SKILL.md 经 Skill 工具加载、reference 经 Read，都从 trace 实测，没有恒在的面）；判决集=实测∪预期触达。**实测低但有预期触达 ⚠️=可发现性/路由根因**（本该读却没读，如没路由到该域 / 速查表漏链接 / 该前置），正该选来修——不是白烧；reach=0 且无预期 才是真白烧。 **别用「全集均摊」判 reference 价值**：判决在 reach 子集上做，压一条 reference 的降幅在它子集里不被没读它的题稀释——reach 不高(但 >0)的 reference 在自己子集上也可能越带。
+- `skills/lark-im/SKILL.md` → reach=1.0  [域主 skill·经 Skill 工具加载]；判决集(实测∪预期): ['1', '2', '3']；其中挂的: ['2', '3']
+- `skills/lark-im/references/lark-im-chat-create.md` → reach=0.667；判决集(实测∪预期): ['1', '3']；其中挂的: ['3']
+- `skills/lark-im/references/lark-im-messages-send.md` → reach=0.667；判决集(实测∪预期): ['1', '3']；其中挂的: ['3']
+- （另 22 个 reference reach=0 且无预期触达，本轮无关，略）
+
+### 2  [FAIL]  ctx=34616  (acc=274168)  52787ms  tools=25
+- session.jsonl: harness-opt/baseline/runs/run-1/detail_info/cases/CLI_核心评测_015/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✓ 成功定位名为「fusanming_at_openclaw群」的群，并获取最近包含「飞豆」关键字的消息。
+    ✓ 将筛选出的相关消息内容转发到「fusanming_at_需求测试群」。
+    ✓ 在「fusanming_at_需求测试群」中 @傅六铭 做知会，消息发送成功。
+
+### 3  [FAIL]  ctx=31289  (acc=225396)  46776ms  tools=22
+- session.jsonl: harness-opt/baseline/runs/run-1/detail_info/cases/CLI_核心评测_080/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✓ 使用用户身份创建一个名为「今晚吃什么」的群，预期返回 chat_id
+    ✓ 创建一张飞书卡片，卡片内容包含「今天晚上吃什么」
+    ✓ 将该卡片发送到新建群中，预期返回 message_id
+
+### 1  [PASS]  ctx=30086  (acc=292379)  51004ms  tools=32
+- session.jsonl: harness-opt/baseline/runs/run-1/detail_info/cases/CLI_核心评测_014/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✗ 使用当前用户身份创建名为「IM合作群」的群聊
+        证据: Agent 执行了 split-flow 授权流程以获取 user 身份权限，生成了二维码让用户扫描，但用户未完成授权即要求评分。Auth status 显示 'User identity: missing'，群聊未被创建。
+    ✗ 将傅一铭和傅二铭加入该群
+        证据: 依赖群聊创建结果。由于群聊未创建（blocked by user identity missing），无法添加成员。
+    ✗ 在该群发送文本消息「大家体验有问题随时沟通」，并返回可验证的 chat_id / message_id
+        证据: 依赖群聊创建结果。由于群聊未创建，无法发送消息。

harness-opt/rounds/round-002/attribution.json

@@ -0,0 +1,65 @@
+[
+  {
+    "case_id": "1",
+    "case_label": "CLI_核心评测_014",
+    "verdict": "PASS",
+    "verdict_note": "workorder=PASS（聚合口径），判分点证据 3/3 ✗ → 实质 FAIL，按判分点当 FAIL 归因",
+    "token": 34555,
+    "token_visible_est": 17364,
+    "duration_ms": 37000,
+    "tool_calls": 8,
+    "cmd_attempts": 7,
+    "cmd_failures": 5,
+    "cmd_fail_rate": 0.71,
+    "discoverability_state": "③ 读了仍卡（SKILL.md+chat-create.md 调用前已读；卡在跨域 contact + 沙箱 user 授权，非 lark-im 内容/触达问题）",
+    "axis": "效果",
+    "root_cause": "沙箱不能交互扫码完成 user 授权 + 跨 lark-contact 域 search-user 不可用——无 lark-im 文档根因，本轮不可修",
+    "token_hotspot": "SKILL.md 常驻正文(RC-1) + chat-create.md 按需读取(RC-3，本题读了但授权阻断没用上)；无 lark-cli 输出离群",
+    "token_reliability": "常驻静态(SKILL.md 3751) + 按需读取(chat-create.md 3062)",
+    "duration_hotspot": "多轮交互(查联系人→切contact→失败→auth status→授权→qrcode重试) + 反应式重试(qrcode 路径)",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现",
+    "doc_fix_hint": "无 lark-im 文档可修点（效果根因在环境+跨域）；lark-im 侧仅 token 减法（SKILL.md 常驻、chat-create.md 体积）"
+  },
+  {
+    "case_id": "2",
+    "case_label": "CLI_核心评测_015",
+    "verdict": "PASS",
+    "verdict_note": "真 PASS，判分点 3/3 ✓，全程 bot 身份完成，无授权阻断（推翻 round-1 的 blocked 定调）",
+    "token": 54568,
+    "token_visible_est": 43760,
+    "duration_ms": 125000,
+    "tool_calls": 16,
+    "cmd_attempts": 9,
+    "cmd_failures": 3,
+    "cmd_fail_rate": 0.33,
+    "discoverability_state": "① 从没读（chat-messages-list.md / messages-search.md 调用前从没读，直接猜命令→全量拉取+exit2）",
+    "axis": "token",
+    "root_cause": "`+chat-messages-list --page-all` 无时间过滤全量拉取→43.5KB持久化→Read 灌入 22556 tok；放大器是 chat-messages-list.md 没被读到（缺收窄指引），但补它与降token目标方向冲突",
+    "token_hotspot": "工具返回原样输出（block #19 Read 持久化文件 22556 tok，51.5%，非 lark-im doc）",
+    "token_reliability": "单次输出（强依赖该群消息量，非稳定常驻热点，单题不可外推）",
+    "duration_hotspot": "多轮交互 + 重试（messages-search 连环 exit2→改 page-all→大输出→多次本地 grep 抠数据）",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现；工具调用 16 明显高于 080，作旁证",
+    "doc_fix_hint": "token 黑洞来自工具输出非文档；SKILL.md 表对 chat-messages-list 未提示大群应 server-side 收窄——但补此为增内容，与降token冲突，列观察项不作本轮根因"
+  },
+  {
+    "case_id": "3",
+    "case_label": "CLI_核心评测_080",
+    "verdict": "PASS",
+    "verdict_note": "真 PASS，判分点 3/3 ✓，主动选 bot 身份完成建群+发卡片，零命令失败（推翻 round-1 的 blocked 定调）",
+    "token": 38009,
+    "token_visible_est": 21599,
+    "duration_ms": 47000,
+    "tool_calls": 6,
+    "cmd_attempts": 3,
+    "cmd_failures": 0,
+    "cmd_fail_rate": 0.0,
+    "discoverability_state": "③ 读了即用（SKILL.md+chat-create.md+messages-send.md 调用前全读到且用上，无触达问题）",
+    "axis": "token",
+    "root_cause": "messages-send.md 单文件 5365 tok（内部 4 处『选 content flag』语义重叠 + Commands 全形态罗列）+ SKILL.md 常驻 + chat-create.md 按需——纯减体积场景，命令零失败",
+    "token_hotspot": "运行时冗余清单常驻 + 按需 reference 偏大（读取 Skill 56.4%：messages-send.md 5365 + SKILL.md 3751 + chat-create.md 3060）",
+    "token_reliability": "常驻静态(SKILL.md 3751) + 按需读取(messages-send.md 5365 子集reach0.333、chat-create.md 3060 子集reach0.667)",
+    "duration_hotspot": "无离群（47s 正常建群+发卡片串行，无重试、无写后回查）",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现",
+    "doc_fix_hint": "messages-send.md 选型规则在 4 处重复表述、Commands 罗列全部媒体形态；SKILL.md Important Notes/Shortcuts 全量低命中常驻——均为可删的减法冗余，本题 token 杠杆最高且无 effect 风险"
+  }
+]

harness-opt/rounds/round-002/case-commands.json

@@ -0,0 +1,27 @@
+{
+  "1": [
+    "auth login",
+    "auth qrcode",
+    "contact +search-user"
+  ],
+  "3": [
+    "auth login",
+    "auth qrcode",
+    "auth status",
+    "im +chat-create",
+    "im +messages-send"
+  ],
+  "2": [
+    "auth login",
+    "auth qrcode",
+    "auth status",
+    "im +chat-messages-list",
+    "im +chat-search",
+    "im +messages-mget",
+    "im +messages-search",
+    "im +messages-send",
+    "im messages forward",
+    "schema im.messages.forward",
+    "schema im.messages.search"
+  ]
+}

harness-opt/rounds/round-002/child-cache.json

@@ -0,0 +1,11 @@
+{
+  "3": {
+    "score": 1.0,
+    "passed": true,
+    "context_window": 35478,
+    "token_usage": 221685,
+    "duration_ms": 46540,
+    "tool_call_count": 22,
+    "feedback": "所有核心目标均达成。执行者经历了两次试错（shell 引号问题、@file 语法不支持），但均自行修正并成功完成任务，符合合理的调试流程。群创建、卡片创建、消息发送三个决策点全部通过。卡片内容准确包含「今天晚上吃什么」文字，message_id 成功返回。\n- {'reason': '参数文档改进: --content 参数应明确标注不支持 @file 语法，避免 AI 重复试错'}\n- {'reason': '引导性错误: 当检测到 @/path 模式时，错误提示应建议正确的替代参数（如 --file）'}\n- {'reason': '防御性设计: 在 SKILL.md 补充大型 JSON 内容的分段写入指引，减少因引号转义导致的失败'}"
+  }
+}

harness-opt/rounds/round-002/diagnosis.md

@@ -0,0 +1,113 @@
+# Round 2 归因（parent=round-1 已采纳候选 51f2a70e；候选模块见 candidate_modules，由 candidate-writer 据诊断+reach 点名）
+
+> 目标（objective.json）：**在不回退成功率的前提下降低 lark-im skill 文档的 token 成本**。effect 是硬门槛、不可退化；token 与 duration 是并列成本杆。tier=T1，仅可改 `skills/lark-im/**`。
+> 判分点只当「什么算挂」的锚，不抄 grader 药方。
+> **本轮 trace = round-1 已采纳候选（51f2a70e，SKILL.md 已 trim 到约 3,915 tok）的行为**，不是 baseline。三题 session 实测已确认 SKILL.md 注入正文为 3,751 tok/题（与 trim 后体积一致），round-1 报告的 5,722 tok/题是 trim 前数字，已过期。
+
+## ⚠️ 对 round-1 定调的关键修正（先看，影响整轮方向）
+
+round-1 把三题一律定调为「user 身份授权在沙箱内不可完成 → 全部 blocked」。**实测 trace 推翻了这个 monolith：三题行为完全不同，只有 1 题真卡授权。**
+
+| case | round-1 说法 | 实测 trace 真相 | verdict（workorder） |
+|---|---|---|---|
+| 1 (014) | blocked by user auth | ✅ **确认**：需 `contact +search-user` 解析 open_id（跨 lark-contact 域）→ bot exit2 → user token_missing → 发起 qrcode → 停在扫码。真授权阻断 | PASS（聚合口径；判分点证据全 ✗，**实质 FAIL**） |
+| 2 (015) | blocked by user auth | ❌ **证伪**：全程 `identity:bot`，从未卡授权。搜群✓、定位「飞豆」消息✓、转发✓、@傅六铭✓，两次 `messages-send` 全 `ok:true`。**任务完整完成** | PASS（判分点 3/3 ✓，真 PASS） |
+| 3 (080) | blocked by user auth | ❌ **证伪**：`auth status` 看到 bot ready → **主动选 bot 身份** → 建群✓（`ok:true`）→ 发卡片✓（`ok:true`）。**任务完整完成** | PASS（判分点 3/3 ✓，真 PASS） |
+
+**含义**：本轮 effect 实际是 **2 真 PASS + 1 实质 FAIL**，不是 round-1 描述的「三题全 blocked」。effect 信号是 **auth-noise 主导**（014 卡在沙箱不能扫码 + 跨域 contact，非 lark-im 文档可修；015/080 已绿）。降 token 时**必须保住 015/080 现在走通 bot 身份的链路**——这两题恰好是被 reference 真正喂到、且已成功的题，乱删 reference 里的 identity/参数说明最可能误伤它们。
+
+## 跨 case 共同根因（优先看；按对 TOKEN 目标的杠杆排序）
+
+### RC-1（token，头号抓手，3 题全命中、最稳）—— SKILL.md `## Important Notes` + Shortcuts 全表常驻，本轮任务低命中
+- **现象**：SKILL.md 经 Skill 工具每题必加载（reach=1.0），实测 3,751 tok/题、三题一致（常驻静态）。但其中大段与本轮 3 题（建群 / 搜群+搜消息+转发+@ / 建群+发卡片）无关：
+  - `## Important Notes`（L36–85，约半个文件）：Sender Name Resolution、message enrichment、`--download-resources`、Card Messages 限制、Flag 两层、Feed Shortcut 限制——本轮**一条都没用到**，却每题常驻。
+  - `## Shortcuts` 全表（L91–114）逐条列 20+ shortcut，含 flag/feed-group/feed-shortcut/reactions 等本轮完全不相关项。
+- **可信度=常驻静态**：tiktoken 可测、跨题稳定（3,751×3）。这是降 token 最稳的发力点，且 3 题全命中（reach=1.0），降幅不被任何子集稀释。
+- **axis=token**。文档位置：`skills/lark-im/SKILL.md` 的 `## Important Notes` 低命中小节 + `## Shortcuts` 全量表。
+- **方向张力（必须标注）**：这是 round-1 已经动过一刀的同一文件（折叠了 API Resources/权限表）。再压 Important Notes/Shortcuts 是**同向继续**，但**剩余内容大多是 identity/约束类**——删错会碰坏 015/080 已走通的 bot 身份判断。candidate-writer 取舍时这是 effect 风险点，不是 RC-1 不成立。
+
+### RC-2（token，次级抓手，080 命中、按需读取）—— `messages-send.md` 单文件偏大且内部高度冗余
+- **现象**：080 读了 `messages-send.md`，实测 **5,365 tok**——本轮所有按需 reference 里最大的单块（占 080 visible 的 24.8%）。该 reference 实测被读且**确实用上了**（080 据此发卡片成功），不是「读了没用」。
+- **从文档看为何这么大**：messages-send.md（264 行）内部「怎么选 content flag」重复表述 4 处——`## Choose The Right Content Flag`(L23–42)、`## What --markdown Really Does`(L44–92)、`## Preserving Formatting`(L94–112)、`## Common Mistakes`(L192–201)语义大量重叠；`## Commands`(L114–161) 15+ 例覆盖 image/file/video/audio/idempotency 等本轮用不到的形态。这是「单文件冗余 + 全形态罗列」，不是信息缺失。
+- **可信度=按需读取**：只在实读它的子集（reach=0.333，仅 080）里计入，压缩降幅在该子集不被稀释——但**子集只有 1 题**，证据基数小，效果需评测确认（见数据缺口）。
+- **axis=token**。文档位置：`skills/lark-im/references/lark-im-messages-send.md`。
+
+### RC-3（token，次级抓手，014+080 命中、按需读取）—— `chat-create.md` 按需读取偏大
+- **现象**：014 与 080 都读了 `chat-create.md`，实测 3,060–3,062 tok（reach=0.667）。080 据此建群成功（用上了）；014 读后因 user 授权阻断没走到建群（读了但本题没用上）。
+- **可信度=按需读取**（reach=0.667，子集 2 题）。体积本身不离群，杠杆低于 RC-2，列为更次级。
+- **axis=token**。文档位置：`skills/lark-im/references/lark-im-chat-create.md`。
+
+### RC-4（效果，无文档根因 / 本轮不可修）—— 014 的 user 授权阻断 + 跨域 contact 依赖
+- **现象**：014 需先解析「傅一铭/傅二铭」open_id，走 `contact +search-user`（**lark-contact 域，不在 candidate_modules**）：bot 身份 exit2（invalid_argument）→ `--as user` token_missing → 发起 `auth login`+qrcode → 停在扫码。判分点证据全 ✗。
+- **归因落点**：根因=沙箱不能交互扫码（环境）＋ 跨域 contact 命令不可用（非 lark-im）。**lark-im 文档侧无根因、无可修点**——这正是约束 3 的「无文档根因 / 本题不改」出口，不要为凑根因往 lark-im doc 上硬编。
+- **axis=效果**，标注**无文档根因 / 本轮不改**。effect 维持 baseline 即可，不要试图改路由让 014「修绿」（用户显式要本人身份解析联系人，改 bot 是 reward-hack）。
+
+## 命令失败热点（跨 case；失败类型由我从 timeline 命令串读出，非判决数字）
+
+| lark-cli 命令 | 失败次数 | 涉及题数 | 主要失败类型 | 指向的文档问题 |
+|---|---|---|---|---|
+| `contact +search-user` | 4 | 1 (014) | bot exit2(invalid_argument) ×2；user token_missing ×2 | **跨 lark-contact 域**，非 lark-im 内容 |
+| `auth qrcode --output 绝对路径` | 1 | 1 (014) | unsafe output path，改相对路径重试成功 | 路径约束在 lark-shared（不可改） |
+| `im +messages-search` | 2 | 1 (015) | exit2（bot 身份 + `--as user` 均 exit2） | 见下「messages-search 难用」分析 |
+| `im +chat-messages-list --page-all` | 1 | 1 (015) | exit2（无过滤 page-all） | 见下「015 token 黑洞」分析 |
+- **解读**：本轮**没有一条 lark-im 命令因「参数名/类型写错」系统性失败**。080 三条命令 0 失败；015 的失败集中在 `messages-search`（见下）。这意味着**没有 lark-im 侧的常规「报错/参数整形」工单**——与 RC-1/2/3 的 token 方向一致，本轮抓手是减体积不是补内容。
+
+### 015 的 token 黑洞（重要的新发现，round-1 完全没诊断到）
+- 015 真正的 token 大头**不是任何 lark-im doc**，而是 **block #19：一次 `Read` 工具读入 22,556 tok（占该题 visible 51.5%）**。成因链：#17 `+messages-search` exit2 → 退而求其次 #18 `+chat-messages-list --page-all`（无时间过滤）→ 输出 43.5KB 被持久化到文件 → agent `Read` 整个文件 → 22.5k tok 灌进上下文。后面又靠本地 `grep`(#27–33) 抠出「飞豆」两条。
+- **从文档角度**：`chat-messages-list.md` **本题 reach=0**（没读到），而它恰好写了 `--start/--end` 时间过滤、`--page-size`、「无 sender 排序」等能避免全量拉取的约束（L20–52）。SKILL.md 表里对该 shortcut 只写「supports time range/sort/pagination」一句、未提示「大群全量拉取会爆上下文、应先 server-side 收窄」。**这是一个真实的「该读没读 → 全量灌入」放大器**（约束 5 状态①：调用前从没读该 reference）。
+- **但这条对本轮目标是「方向张力」，不是干净的 token 抓手**：要避免全量灌入，文档侧只能**增加**收窄指引（前置或加 caution），这与「降 token」的常驻成本目标**方向相反**（见硬性约束 7 的冲突记录）。且 22.5k 黑洞是**单次工具输出**（单次输出可信度、单题、强烈依赖该群消息量），不是稳定常驻热点。**结论：列为观察项交评测裁决，不要当成 RC-1 那种干净抓手去推「前置 chat-messages-list」——很可能只增 token 不省。**
+
+## 可发现性时序（约束 5 三态；判「前置能不能救」的决定性证据）
+> 对每条相关 reference / `--help`，按相对首次失败调用的读取时序统计。`--help` 扫 Bash（本轮 3 题均未跑任何 `--help`）。
+
+| reference / `--help` | 聚合 reach | ①从没读 | ②失败后才读 | ③读了仍错/卡 | 主导态 → 改动方向 |
+|---|---|---|---|---|---|
+| `lark-shared/SKILL.md` | 1.0 | 0 | 0 | — | 三题调用前都读了；014 仍卡（环境，非内容）；不可改 |
+| `chat-create.md` | 0.667 | 0 | 0 | — | 080 调用前读→建群成功；014 调用前读→授权阻断（非 reference 错）。**非触达问题** |
+| `messages-send.md` | 0.333 | 0 | 0 | — | 080 调用前读→发卡片成功。**非触达问题** |
+| `chat-messages-list.md` | 0.0 | 1 (015) | 0 | — | ① **015 调用前从没读**→直接 `--page-all` 全量拉取→token 黑洞。触达缺口，但补它=增 token，与目标冲突（见上） |
+| `messages-search.md` | 0.0 | 1 (015) | 0 | — | ① 015 从没读 messages-search.md，直接猜 `+messages-search` ×2 → exit2。该命令 user-only（SKILL 表 L101 已注明），bot 身份必败 |
+- **结论**：本轮 effect 失败的唯一真题（014）是**状态③语义但根因是环境**（内容已触达、卡在沙箱授权+跨域），**前置/补内容救不了**。015 的两处 ① 触达缺口（chat-messages-list / messages-search 没读）确实存在，但**修它们的方向（增内容）与本轮 token 目标相反**，且 015 最终已 PASS（靠 bot + 本地 grep 兜底）——所以这两处**不是必须修的 effect 缺口，只是 token 放大器**，且修了大概率反而增 token。
+- **对 candidate-writer 的含义**：**本轮没有「该前置」的干净 case**。RC-1/2/3 都是「调用前已读、内容够用 → 减体积」的纯 token 减法，不涉及触达。不要被 015 的两处 ① 诱导去推前置——那会与目标背道而驰。
+
+## 方向冲突记录（硬性约束 7）
+- **减体积（RC-1/2/3，与 objective.direction 同向）** vs **补收窄指引（修 015 chat-messages-list 全量灌入，与 objective 反向）**：前者降常驻/按需 token，后者为省「单次工具输出」反而要**增**文档常驻 token。两者方向相反，**不可合并**。本轮目标是降 token，应取减体积一侧；015 的全量灌入作为观察项记录、不作为本轮要补的内容根因。
+
+## 差距台账复盘
+- 无（round 2，`discard-ledger.json` 为空，无已跑未采纳候选）。
+
+## 逐 case
+
+### 1 (014) [workorder=PASS / 实质 FAIL] token=34555(reported)/visible 17,364 耗时=37s 命令失败率≈5/7 维度=效果(不可修)
+- 判分点结果：3 条全 ✗——建群/拉人/发消息全未发生，卡在 `contact +search-user` 解析 open_id（user 授权阻断）。verdict=PASS 系聚合口径，按判分点证据当 FAIL 处理。
+- 命令失败：≈5/7。`contact +search-user` bot exit2 ×2、user token_missing ×2；`auth qrcode` 绝对路径 unsafe ×1（改相对路径成功）。**全部非 lark-im 命令的内容错误**。
+- 可发现性时序：调用前读了 SKILL.md(reach=1.0)+chat-create.md(3,062 tok)；失败在更上游的跨域 contact + 授权。**非 lark-im 触达问题**。
+- token 归因：SKILL.md 正文 3,751（常驻静态，21.6%）+ chat-create.md 3,062（按需，17.6%，本题没走到建群=读了没用上）+ 系统 Skill 列表注入 4,612（固定开销，不归因）。lark-cli 命令累计含多次短失败回显，单条都短、非热点。
+- 耗时归因：本题往返多（查联系人→切 contact→失败→auth status→授权→qrcode 重试）。多为授权链路 + 跨域固有串行 + 反应式重试（duration 弱信号，需多轮复现）。
+- 文档根因：效果=沙箱 user 授权 + 跨域 contact（环境，**无 lark-im 文档根因，本轮不改**）；token=SKILL.md 常驻（RC-1）+ chat-create.md 按需（RC-3）。
+
+### 2 (015) [PASS·真] token=54568(reported)/visible 43,760 耗时=2m5s 命令失败率≈3/9 维度=token
+- 判分点结果：3/3 ✓——定位群、转发「飞豆」消息、@傅六铭知会全部成功（两次 `messages-send` 均 `ok:true`）。**全程 bot 身份，无授权阻断**。
+- 命令失败：≈3/9。`+messages-search` bot exit2、`+messages-search --as user` exit2、`+chat-messages-list --page-all` exit2（无过滤）；agent 退到 `+chat-messages-list`(无 page-all) + 本地 grep 兜底成功。
+- 可发现性时序：① `messages-search.md` / `chat-messages-list.md` **调用前从没读**（reach=0），直接猜命令。messages-search 是 user-only（SKILL 表 L101 已注明）、bot 身份必败——agent 没看清就猜。
+- token 归因：**本题 token 大头不是 lark-im doc**，是 block #19 一次 `Read` 持久化文件 = **22,556 tok（51.5%，其他工具调用/返回）**，成因=`--page-all` 无过滤全量拉取→43.5KB→Read 灌入（单次输出可信度，强依赖该群消息量）。SKILL.md 正文 3,749（常驻）。lark-shared 3,749（跨 skill，不归因 lark-im）。
+- 耗时归因：本题最长(2m5s)，主因是 messages-search 连环失败→改用 page-all→大输出→多次本地 grep 抠数据的多轮往返（duration 弱信号；工具调用 16 raw32，明显高于 080，作旁证）。
+- 文档根因：token 黑洞的放大器=`chat-messages-list.md` 没被读到 + SKILL.md 表未提示大群应 server-side 收窄——但**补这条与降 token 目标相反**（方向张力，见上），列为观察项；本题已 PASS。常规 token 抓手仍是 RC-1（SKILL.md 减体积）。
+
+### 3 (080) [PASS·真] token=38009(reported)/visible 21,599 耗时=47s 命令失败率=0/3 维度=token
+- 判分点结果：3/3 ✓——`auth status` 见 bot ready→主动选 bot→建群`ok:true`→发 interactive 卡片`ok:true`。**任务完整完成，零命令失败**。
+- 命令失败：0/3。三条 lark-cli（auth status / chat-create / messages-send）全成功。
+- 可发现性时序：调用前读 SKILL.md + chat-create.md(3,060) + messages-send.md(5,365)，全部状态③（调用前已读且用上）。**无触达问题**。
+- token 归因：**本题是纯 token 抓手题**——读取 Skill 占 56.4%：messages-send.md 5,365（按需，最大单块，RC-2）+ SKILL.md 3,751（常驻，RC-1）+ chat-create.md 3,060（按需，RC-3）。三块 reference/SKILL 都实读且 RC-2 的 messages-send.md 确实用上了。系统 Skill 列表注入 4,612（固定开销，不归因）。
+- 耗时归因：47s，全部为正常建群+发卡片串行，无重试、无写后回查（无离群）。
+- 文档根因：无效果根因（已绿）；token=RC-2(messages-send.md 内部冗余) + RC-1(SKILL.md 常驻) + RC-3(chat-create.md)。**本题 token 杠杆最高且无 effect 风险**（命令全成功，减 reference 体积不碰已走通链路）。
+
+## 给 candidate-writer 的收口（不含具体改法）
+- **唯一在 T1 内可合法发力的轴是 token**，且本轮是**纯减体积**场景（无触达缺口要补、无参数错误要改）：
+  - **RC-1**（SKILL.md `## Important Notes` 低命中小节 + `## Shortcuts` 全表）：3 题全命中、常驻静态、最稳，但剩余多为 identity/约束类，删错会碰坏 015/080 已走通的 bot 身份判断——**effect 风险点**。
+  - **RC-2**（messages-send.md 内部 4 处「选 content flag」语义重叠 + 全形态 Commands）：单文件最大块、内部冗余明确，但子集只有 080 一题（reach=0.333），证据基数小、效果需评测确认。
+  - **RC-3**（chat-create.md 按需偏大）：杠杆最低，列为更次级。
+- **effect 不可在本轮 T1 内合法抬升**：014 是环境（沙箱不能扫码）+ 跨域 contact，无 lark-im 文档根因。015/080 已真 PASS。候选必须**保住 015/080 走通 bot 身份的 identity/参数说明**，降 token 时别误伤。
+- **不要推前置**：本轮没有「该前置」的干净 case。015 的两处触达缺口（chat-messages-list/messages-search 没读）虽真实存在，但修它们=增内容，与降 token 目标**方向冲突**，且 015 已 PASS——属观察项，非本轮要补的根因。
+- **缺失信息（doc_fix_hint 语气）**：SKILL.md 的 Important Notes/Shortcuts 全量罗列、本轮低命中却每题常驻；messages-send.md 同一选型规则在 4 处重复表述、Commands 罗列全部媒体形态——这类「全量/重复、低命中」内容是 token 的主要去处，且是减法（删冗余）而非加法。
+- **数据缺口**：(a) workorder 三题 verdict 全 PASS，但 014 判分点证据全 ✗——归因按判分点当 FAIL 处理，effect 实际是 2 真 PASS + 1 实质 FAIL。(b) RC-2/RC-3 子集小（messages-send.md 仅 080、chat-create.md 仅 014+080），单轮证据基数小，token 降幅需评测在子集上确认。(c) 015 的 22.5k 黑洞是单次工具输出，强依赖该群消息量，非稳定常驻热点，单题不可外推。(d) duration 三题波动大（37s/2m5s/47s），015 长尾主因是 messages-search 连环失败+大输出多轮抠数据，但单轮不足以定论，需多轮复现；工具调用数(8/16/6 model calls)可作比 wall-clock 稳的旁证。(e) 工具调用次数 session-analyze(model calls 8/16/6) 与 workorder 趋势表(R1 均值 26.3) 口径不一致，趋势表疑似含 raw 计数，旁证以 timeline 实际往返为准。

harness-opt/rounds/round-002/discard-ledger.json

@@ -0,0 +1 @@
+[]

harness-opt/rounds/round-002/failure-memory.json

@@ -0,0 +1 @@
+[]

harness-opt/rounds/round-002/module-reach.json

@@ -0,0 +1,220 @@
+{
+  "skills/lark-im/SKILL.md": {
+    "reach": 1.0,
+    "read_cases": [
+      "1",
+      "2",
+      "3"
+    ],
+    "actual_cases": [
+      "1",
+      "2",
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": true
+  },
+  "skills/lark-im/references/lark-im-chat-create.md": {
+    "reach": 0.667,
+    "read_cases": [
+      "1",
+      "3"
+    ],
+    "actual_cases": [
+      "1",
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-identity.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-messages-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-search.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-update.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-list-item.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-query-item.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-groups.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-create.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-remove.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-cancel.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-create.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-message-enrichment.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-mget.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-reply.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-resources-download.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-search.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-send.md": {
+    "reach": 0.333,
+    "read_cases": [
+      "3"
+    ],
+    "actual_cases": [
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-reactions.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-threads-messages-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  }
+}

harness-opt/rounds/round-002/review.json

@@ -0,0 +1,15 @@
+{
+  "generated_by": "lark-cli-harness:opt-reviewer",
+  "verdict": "PASS",
+  "module": "skills/lark-im/references/lark-im-messages-send.md",
+  "tier": "T1",
+  "reason": "纯结构性去重：16407→6399 字节(-61%)与策略一致；逐项核对每条承重指令(互斥规则、video-cover 必配、cwd-relative/绝对路径拒绝、markdown→post 边界、三套 <at> 语法、content 各 msg_type 样例、Safety Constraints、identity+scope 映射)均原样保留在新文档内联，删的全是重复/过度罗列(4× 选型规则、镜像 --help 的 Parameters 表、Common Mistakes、Notes、冗余 Commands)。无硬编码评测答案、未针对 080 卡片流窄化、未碰 SKILL.md 身份路由、单文件单根因。",
+  "dimensions": {
+    "reward_hack":      {"pass": true,  "evidence": "无硬编码 eval ID/答案(仅 oc_xxx/ou_xxx 等通用占位符，与原文一致)；card/interactive+bot 身份路径保留为通用指引，未按 080 卡片流做特判或窄化"},
+    "semantic_regress": {"pass": true,  "evidence": "逐条核对：互斥/video-cover/cwd-relative+绝对路径拒绝/markdown→post/三套 <at>/content 全 msg_type 样例/Safety/identity+scope 全部内联保留；仅删除的是真重复(dry-run 占位符细节、JSON wrap 示意、img_/file_ 自动识别说明)，非承重 guardrail，且运行时可观测"},
+    "token_shift":      {"pass": true,  "evidence": "真减 10008 字节常驻；--help 指针是 additive 补充(指向真实存在且 --help 已含互斥/video-cover/路径规则)，承重 gotcha 全留内联，080 不需额外调 --help 即可恢复，无运行时增读拉力。注:work-order 提的 schema im.messages.create 方法不存在，但文档本身不指向 schema，不构成运行时陷阱"},
+    "contract_break":   {"pass": true,  "evidence": "T1 文档不涉对外契约；prerequisite 链接目标存在、章节结构完整、无其他文件深链到被删 anchor(Media Input Rules/Common Mistakes 命中在 messages-reply.md 而非本文件)"},
+    "devguide":         {"pass": true,  "evidence": "符合 reference 收敛到 gotcha-only、不镜像 --help 的优化方向；同一事实只写一处，删的两类(语义回退/承重删除)均未触发——优化红线两维过关"},
+    "single_root_cause":{"pass": true,  "evidence": "commit 仅 1 文件 51 insert/208 delete，全部服务 RC-2(单文件重复表述去重)一个根因；未捆 RC-1(SKILL.md)/RC-3(chat-create)，未把无关删除以 token 对冲缝入"}
+  }
+}

harness-opt/rounds/round-002/round.json

@@ -0,0 +1,380 @@
+{
+  "round": 2,
+  "status": "admitted",
+  "parent_id": "a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e",
+  "parent_worktree": "/Users/bytedance/Projects/cli",
+  "child_worktree": "/Users/bytedance/Projects/cli",
+  "base_commit": "51f2a70e6dffeea65d928badb6207408490dc215",
+  "module": "skills/lark-im/references/lark-im-messages-send.md",
+  "candidate_modules": [
+    "skills/lark-im/SKILL.md",
+    "skills/lark-im/references/lark-im-chat-create.md",
+    "skills/lark-im/references/lark-im-chat-identity.md",
+    "skills/lark-im/references/lark-im-chat-list.md",
+    "skills/lark-im/references/lark-im-chat-messages-list.md",
+    "skills/lark-im/references/lark-im-chat-search.md",
+    "skills/lark-im/references/lark-im-chat-update.md",
+    "skills/lark-im/references/lark-im-feed-group-list-item.md",
+    "skills/lark-im/references/lark-im-feed-group-list.md",
+    "skills/lark-im/references/lark-im-feed-group-query-item.md",
+    "skills/lark-im/references/lark-im-feed-groups.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-create.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-list.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-remove.md",
+    "skills/lark-im/references/lark-im-flag-cancel.md",
+    "skills/lark-im/references/lark-im-flag-create.md",
+    "skills/lark-im/references/lark-im-flag-list.md",
+    "skills/lark-im/references/lark-im-message-enrichment.md",
+    "skills/lark-im/references/lark-im-messages-mget.md",
+    "skills/lark-im/references/lark-im-messages-reply.md",
+    "skills/lark-im/references/lark-im-messages-resources-download.md",
+    "skills/lark-im/references/lark-im-messages-search.md",
+    "skills/lark-im/references/lark-im-messages-send.md",
+    "skills/lark-im/references/lark-im-reactions.md",
+    "skills/lark-im/references/lark-im-threads-messages-list.md"
+  ],
+  "module_reach": {
+    "skills/lark-im/SKILL.md": {
+      "reach": 1.0,
+      "read_cases": [
+        "1",
+        "2",
+        "3"
+      ],
+      "actual_cases": [
+        "1",
+        "2",
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": true
+    },
+    "skills/lark-im/references/lark-im-chat-create.md": {
+      "reach": 0.667,
+      "read_cases": [
+        "1",
+        "3"
+      ],
+      "actual_cases": [
+        "1",
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-identity.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-messages-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-search.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-update.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-list-item.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-query-item.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-groups.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-create.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-remove.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-cancel.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-create.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-message-enrichment.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-mget.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-reply.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-resources-download.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-search.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-send.md": {
+      "reach": 0.333,
+      "read_cases": [
+        "3"
+      ],
+      "actual_cases": [
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-reactions.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-threads-messages-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    }
+  },
+  "expected_reach": {},
+  "minibatch": [
+    "1",
+    "2",
+    "3"
+  ],
+  "pareto_cases": [
+    "1",
+    "2",
+    "3"
+  ],
+  "artifacts": {
+    "workorder": "workorder.md",
+    "diagnosis": "diagnosis.md",
+    "attribution": "attribution.json",
+    "strategy": "strategy.md",
+    "review": "review.json",
+    "trend": "trend.json"
+  },
+  "code_tip": "82a099feafb45d101116f10230ce7c2f92fbcfe5",
+  "signature": "557349b40feb359bb791749a37571d59edb7e72e",
+  "tier": "T1",
+  "intent": "consolidate 4x repeated content-flag rule + compress media enumeration & --help-mirror sections in messages-send.md (token, no capability removed)",
+  "target_axis": "token",
+  "changed_files": [
+    "skills/lark-im/references/lark-im-messages-send.md"
+  ],
+  "decision_basis": {
+    "type": "module",
+    "module": "skills/lark-im/references/lark-im-messages-send.md"
+  },
+  "decision_cases": [
+    "3"
+  ],
+  "review": {
+    "generated_by": "lark-cli-harness:opt-reviewer",
+    "verdict": "PASS",
+    "module": "skills/lark-im/references/lark-im-messages-send.md",
+    "tier": "T1",
+    "reason": "纯结构性去重：16407→6399 字节(-61%)与策略一致；逐项核对每条承重指令(互斥规则、video-cover 必配、cwd-relative/绝对路径拒绝、markdown→post 边界、三套 <at> 语法、content 各 msg_type 样例、Safety Constraints、identity+scope 映射)均原样保留在新文档内联，删的全是重复/过度罗列(4× 选型规则、镜像 --help 的 Parameters 表、Common Mistakes、Notes、冗余 Commands)。无硬编码评测答案、未针对 080 卡片流窄化、未碰 SKILL.md 身份路由、单文件单根因。",
+    "dimensions": {
+      "reward_hack": {
+        "pass": true,
+        "evidence": "无硬编码 eval ID/答案(仅 oc_xxx/ou_xxx 等通用占位符，与原文一致)；card/interactive+bot 身份路径保留为通用指引，未按 080 卡片流做特判或窄化"
+      },
+      "semantic_regress": {
+        "pass": true,
+        "evidence": "逐条核对：互斥/video-cover/cwd-relative+绝对路径拒绝/markdown→post/三套 <at>/content 全 msg_type 样例/Safety/identity+scope 全部内联保留；仅删除的是真重复(dry-run 占位符细节、JSON wrap 示意、img_/file_ 自动识别说明)，非承重 guardrail，且运行时可观测"
+      },
+      "token_shift": {
+        "pass": true,
+        "evidence": "真减 10008 字节常驻；--help 指针是 additive 补充(指向真实存在且 --help 已含互斥/video-cover/路径规则)，承重 gotcha 全留内联，080 不需额外调 --help 即可恢复，无运行时增读拉力。注:work-order 提的 schema im.messages.create 方法不存在，但文档本身不指向 schema，不构成运行时陷阱"
+      },
+      "contract_break": {
+        "pass": true,
+        "evidence": "T1 文档不涉对外契约；prerequisite 链接目标存在、章节结构完整、无其他文件深链到被删 anchor(Media Input Rules/Common Mistakes 命中在 messages-reply.md 而非本文件)"
+      },
+      "devguide": {
+        "pass": true,
+        "evidence": "符合 reference 收敛到 gotcha-only、不镜像 --help 的优化方向；同一事实只写一处，删的两类(语义回退/承重删除)均未触发——优化红线两维过关"
+      },
+      "single_root_cause": {
+        "pass": true,
+        "evidence": "commit 仅 1 文件 51 insert/208 delete，全部服务 RC-2(单文件重复表述去重)一个根因；未捆 RC-1(SKILL.md)/RC-3(chat-create)，未把无关删除以 token 对冲缝入"
+      }
+    }
+  },
+  "child_k": 5,
+  "eval_trace": null,
+  "retro": {
+    "cause": "已入池",
+    "noise_borderline": false,
+    "summary": "越带入池，无需复盘补发"
+  },
+  "retro_sessions": [
+    {
+      "case": "3",
+      "session": null,
+      "axis": "token",
+      "expect": "降",
+      "parent": 37942,
+      "child": 35478,
+      "gain": "收益现",
+      "pass_delta": null
+    }
+  ],
+  "verdict": "admitted",
+  "ci": null,
+  "new_candidate": "557349b40feb359bb791749a37571d59edb7e72e",
+  "decision": {
+    "parent_success": 1.0,
+    "child_success": 1.0,
+    "parent_score": 0.6,
+    "child_score": 1.0,
+    "score_saved": 0.4,
+    "score_threshold": 0.09532271373123208,
+    "parent_token": 37942.0,
+    "child_token": 35478.0,
+    "saved": 2464.0,
+    "threshold": 4532.708313776408,
+    "parent_duration": 45769.0,
+    "child_duration": 46540.0,
+    "dur_saved": -771.0,
+    "dur_threshold": 4899.200953624988,
+    "dur_margin": 1.0,
+    "missing_duration": [],
+    "k_child": 5,
+    "k_parent": 5,
+    "decision_n": 1,
+    "missing_context": [],
+    "missing_score": [],
+    "parent_token_acc": 251669.0,
+    "child_token_acc": 221685.0,
+    "phi0_score": 0.5333333333333333,
+    "eff_margin": 1.0,
+    "parent_token_full": 37942.0,
+    "child_token_full": 35478.0,
+    "saved_full": 2464.0,
+    "observe_n": 1,
+    "target_axis": "token",
+    "admitted": true,
+    "reason": "score_gain"
+  },
+  "patch": "verify_results/round-002-lark-im-references-lark-im-messages-send.patch"
+}

harness-opt/rounds/round-002/strategy.md

@@ -0,0 +1,48 @@
+# Round 2 候选策略（模块=references/lark-im-messages-send.md, tier=T1, 主指标=token）
+
+## 根因与选择
+
+| 根因 | 来源(评测归因/规范经验) | 承载模块(reach) | annotation 风险级 | coverage 档 | P级 | 选中 |
+|---|---|---|---|---|---|---|
+| RC-2: messages-send.md 单文件最大、内部「选 content flag」规则重复 4 处 + 全媒体形态罗列 + Parameters/Notes 镜像 --help | 评测归因①（080 实读实用）＋规范经验②（annotation R1×140/R2×109，仅 1 段 R3） | references/lark-im-messages-send.md (0.333) | R1/R2 主导，唯一 R3=Safety Constraints(L9–22) | 密 / overfit 低 | P1 | ✅ |
+| RC-1: SKILL.md `## Important Notes` 低命中 + `## Shortcuts` 全表常驻 | 评测归因①（reach=1.0，3 题全命中） | SKILL.md (1.0) | R2/R3 混合（identity/约束密集） | 密 / 中 | P0(命中) 但 effect 高风险 |  |
+| RC-3: chat-create.md 按需偏大 | 评测归因① | references/lark-im-chat-create.md (0.667) | — | 密 | P1 |  |
+
+- **选中理由**：RC-2 是诊断点名「最干净的 token 杠杆」——单文件最大块（实测 ~5,365 tok，占 080 visible 24.8%），且 080 调用前已读、确实据它发卡片成功（reach=0.333、actual=1，非「读了没用」）。annotation 证实它 R1/R2 主导（140 R1 + 109 R2 行，可重构/可压缩），唯一 R3 段是 Safety Constraints(L9–22)，我**原样保留语义**。coverage=「密」、overfit「低」→ 本轮 eval 能在 080 上裁真伪。这是纯减体积、零能力删除、不碰 SKILL.md 路由的改动。
+- **为什么不选 RC-1**：reach=1.0、命中率最高，但 diagnosis 明确标它为 **effect 风险点**——剩余内容多为 identity/约束类，正是驱动 015/080 走通 bot 身份判断的承重内容；objective 的**硬门槛是「保住成功率」**，动 SKILL.md 最可能误伤这条已绿链路。本轮放弃，避免拿成功率换 token。
+- **为什么不选 RC-3**：diagnosis 判其杠杆最低（体积不离群），列为更次级；同一根因一轮只动一个，留待后续轮次。
+- **选模块理由**：messages-send.md reach=0.333>0（满足 reach 锁），承载选中的 RC-2，是非域 reference、改它不触碰 SKILL.md 的身份路由面。多文件无——本轮只动这一个文件。
+- **规范经验源补注**：对照 content-taxonomy——「单命令用法/长示例/与 --help 重复」类默认 R0/R1，「一般行为规则/CLI 机制约定」默认 R2；本文件的重复选型规则、全形态 Commands、Parameters/Notes 镜像即此类，处理方向为「留命中率最高一处，其余删或指针」「高频留 2–3 例，长的下沉」。当轮可被 080 裁真伪（coverage 密/overfit 低）。
+
+## 改了什么（逐处）
+- **L23–43 `## Choose The Right Content Flag` + `### --text vs --markdown`**：两段语义重叠的选型说明 → 合并为单张 4 行选型表（markdown/text/content/media），并把互斥规则并入表后一句。删掉 `### --text vs --markdown` 整段（与表重复）。
+- **L44–82 `## What --markdown Really Does` + `### Markdown Boundaries` + `### Image Constraint`**：三段约 39 行 → 压成 `## --markdown Gotchas` 三条要点（强制 post/无 title、标题改写规则、图片预上传 vs 远程 URL vs 本地路径不支持）。删掉 JSON wrap 示意、逐条 boundary 罗列等可由行为观察得到的展开。
+- **L83–93 图片预上传双命令示例**：并入 `## Commands` 的一条 markdown+image 示例（保留 `im images create` → 引用 img_xxx 的关键两步）。
+- **L114–161 `## Commands`（15+ 例覆盖全媒体形态）+ `## Media Input Rules`**：压成代表性示例（markdown / text / DM / post-title / markdown+image / 4 个媒体一组 / idempotency+dry-run），媒体路径规则收成 `--help` 指针后的 3 条 load-bearing gotcha（cwd-relative/绝对路径拒绝、video-cover 必配、msg-type 推断冲突）。
+- **L169–191 `## Parameters` 表**：删除镜像 `--help` 的逐参数描述，改为「Run `lark-cli im +messages-send --help`」指针 + 仅保留 --help 不显然的三条硬规则（已并入 Commands 末尾）。
+- **L192–202 `## Common Mistakes`**：整段删除——逐条都是选型表/markdown gotcha 的反向重述（第 4 次重复选型规则），删后选型信息仍在表里。
+- **L203–216 `## content Format Reference`**：保留（构造 `--content` 的 gotcha），把 image/file/audio 三行合并为一行省重复。
+- **L227–248 `## @Mention Format`**：保留全部三种 msg_type 的 `<at>` 语法（text/post/interactive 各异、AI 猜不到），压紧为两条要点、去掉小标题与重复散文。
+- **L249–264 `## Notes`**：整段删除——逐条（互斥/media 上传/scope/markdown 强制 post/video-cover/msg-type 冲突）均已在 Safety Constraints、选型表、--markdown Gotchas、Commands 指针处各保留一处单一事实源。
+
+## 为什么这么改（机制）
+- **消除根因的因果链**：该 reference 的体积来自「同一份选型规则在 4 个 section 重复 + 全媒体形态逐条罗列 + Parameters/Notes 镜像 --help」。token 不是被任务必需信息占用，而是被**重复表述**占用。按「同一份事实只写一次」（锚点 1）合并到单一事实源后，每条 load-bearing 信息仍恰好出现一次，080 这类「读该 reference→发消息」的题，读入 token 直接下降而行为不变。
+- **不删能力**：每个 flag（text/markdown/content/image/file/video/audio/idempotency/dry-run/msg-type/video-cover/as）、每条硬约束（互斥、video-cover 必配、cwd-relative 路径、绝对路径拒绝、markdown 强制 post/无 title、msg-type 冲突校验）、三套 `<at>` 语法、content 各 msg_type 样例、Safety Constraints、identity+scope 映射——全部保留，只是从「重复 N 次/逐条罗列」变成「一处/代表性示例 + --help 指针」。
+- **规范经验源**：依 optimization-playbook「reference 收敛到 gotcha-only，不做 --help 镜像」——Parameters 全表/全形态 Commands 属 USAGE，下沉到 `--help` 指针；保留的是 --help 表达不了的跨 flag 互斥、媒体路径安全、markdown→post 边界、@mention 按类型差异等 gotcha。annotation 标这些段为 R1（可重构/下沉），符合处理方向；唯一 R3（Safety）原样保留。
+
+## 预期效果
+- **成功率**：不退化。080（唯一读该文件的题）的发卡片链路依赖的是 `--content`/`interactive`、identity=bot、chat-id——全部保留；选型表、content Format Reference、Safety、scope 都在。015/080 走通 bot 身份的判断由 SKILL.md + identity 段承载，本轮**没碰 SKILL.md**，零误伤面。014 与本文件无关（reach 不含 014）。
+- **context（分两层）**：
+  - (1) **静态字数差**：16,407 → 6,399 chars（-61.0%）；tiktoken cl100k 3,869 → 1,799 tok（-53.5%）。（注：diagnosis 报 ~5,365 tok 系另一 tokenizer/含注入开销；此处用 cl100k 自测，方向与幅度一致。）
+  - (2) **运行时 context 方向**：仅在**实读该 reference 的子集**生效——本轮即 080 一题，运行时读入下降约 50%+（该块占 080 visible 24.8%，预计 080 visible 降约 12–13%）。其余两题（014/015）不读该文件，运行时 token **不变**（既不增也不减）。这是按需 reference，不是常驻面，不会影响未读它的题。
+- **覆盖敞口**：RC-2 子集仅 080 一题（reach=0.333），证据基数小。coverage 判该文件「密/overfit 低」，本轮 eval 可在 080 上裁真伪，但单题不可外推到「所有发消息任务」。建议后续补「读 messages-send.md 后用 --markdown / 媒体 / @mention」的 case 加厚子集。预期收益落在 **token 轴**（080 visible 下降），effect 轴维持不退化。
+
+## 刻意没做什么（反 reward-hack / 反过拟合）
+- 没硬编码任何评测题答案；没删任何能力、flag、guardrail、身份/scope 说明；没碰 lark-im 以外文件，也没把无关根因捆进本轮（commit 仅 1 个文件）。
+- **没碰 SKILL.md（RC-1）**：尽管 reach=1.0 杠杆最大，但其剩余内容是驱动 015/080 bot 身份判断的承重 identity/约束，diagnosis 标为 effect 风险点；在「保住成功率」硬门槛下不拿成功率换 token。
+- **没补收窄/分页指引**（015 的 22.5k chat-messages-list 黑洞）：那是「增内容」，与降 token 目标方向相反，diagnosis 已列为观察项、本轮不做。
+- 本改动**不是按评测错误反推**的参数/路由拟合——是基于 annotation + content-taxonomy 的结构性去重，删的是重复表述而非按 080 的具体内容裁剪；真实价值在「任何读该 reference 的发消息任务都少读重复 token」，080 只是当轮可验证的子集。
+- 未发现需要 breaking（T3）才能根治的点；本轮纯 T1 文档去重即可。
+
+## 签名
+- signature: 557349b40feb359bb791749a37571d59edb7e72e (commit 82a099fe 的 diff hash)  tier: T1

harness-opt/rounds/round-002/trend.json

@@ -0,0 +1,11 @@
+[
+  {
+    "round": 1,
+    "n": 3,
+    "pass_n": 0,
+    "cmd_fail_rate": 0.6,
+    "tool_calls": 26.333333333333332,
+    "duration_ms": 50189.0,
+    "token": 31997.0
+  }
+]

harness-opt/rounds/round-002/workorder.md

@@ -0,0 +1,43 @@
+# Round 2 归因派工单（parent=a1333f2e1f7e98bf6f705814b92cacae1f43565759e4e0c24a0a4700b241649e；模块未定，由 candidate-writer 据诊断点名）
+
+> **只读输入**——opt-attributor 读本文件，把诊断**另写** `diagnosis.md`（给 candidate-writer）+ 逐题结构化 `attribution.json`（给 dashboard）。**不要覆盖本文件**，留作派工单↔诊断的前后对比。
+> 判分点只当「什么算挂」的锚，禁止照抄 grader 药方（已从派工单剔除）。
+
+## 模块运行时可达性（选模块第一步的证据；要选须在 strategy.md 说明理由）
+> reach=**实测**触达率（域主 SKILL.md 经 Skill 工具加载、reference 经 Read，都从 trace 实测，没有恒在的面）；判决集=实测∪预期触达。**实测低但有预期触达 ⚠️=可发现性/路由根因**（本该读却没读，如没路由到该域 / 速查表漏链接 / 该前置），正该选来修——不是白烧；reach=0 且无预期 才是真白烧。 **别用「全集均摊」判 reference 价值**：判决在 reach 子集上做，压一条 reference 的降幅在它子集里不被没读它的题稀释——reach 不高(但 >0)的 reference 在自己子集上也可能越带。
+- `skills/lark-im/SKILL.md` → reach=1.0  [域主 skill·经 Skill 工具加载]；判决集(实测∪预期): ['1', '2', '3']
+- `skills/lark-im/references/lark-im-chat-create.md` → reach=0.667；判决集(实测∪预期): ['1', '3']
+- `skills/lark-im/references/lark-im-messages-send.md` → reach=0.333；判决集(实测∪预期): ['3']
+- （另 22 个 reference reach=0 且无预期触达，本轮无关，略）
+
+## 逐轮诊断信号趋势（纯诊断，不进判决）
+
+| 轮 | 题数 | PASS | 命令失败率 | 工具调用 | 耗时(ms) | token |
+|---|---|---|---|---|---|---|
+| R1 | 3 | 0 | 0.60 | 26 | 50189 | 31997 |
+
+> 跨题均值，按轮排。**命令失败率、工具调用数是横切诊断信号，不是准入轴**（准入只走 效果/token/耗时）——用来判「上一轮那刀有没有把失败/轮次压下去」。工具调用数比 wall-clock 稳，可给噪声大的耗时轴当旁证。
+
+### 1  [PASS]  ctx=34270  (acc=274608)  43995ms  tools=31
+- session.jsonl: harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_014/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✗ 使用当前用户身份创建名为「IM合作群」的群聊
+        证据: transcript 在展示授权二维码后结束，无任何 `lark-cli im +chat-create` 调用。执行停在 '授权完成后请告诉我，我会继续帮你创建群聊并发送消息'，群聊未创建。
+    ✗ 将傅一铭和傅二铭加入该群
+        证据: transcript 显示尝试搜索用户时遇到 `need_user_authorization` 错误，授权流程启动后中断。未获取到任何用户的 open_id，无后续添加操作。
+    ✗ 在该群发送文本消息「大家体验有问题随时沟通」，并返回可验证的 chat_id / message_id
+        证据: 群聊未创建，无 chat_id 可返回。transcript 无任何 `lark-cli im messages-send` 调用。
+
+### 2  [PASS]  ctx=47116  (acc=612048)  114310ms  tools=49
+- session.jsonl: harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_015/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✓ 成功定位名为「fusanming_at_openclaw群」的群，并获取最近包含「飞豆」关键字的消息
+    ✓ 将筛选出的相关消息内容转发到「fusanming_at_需求测试群」
+    ✓ 在「fusanming_at_需求测试群」中 @傅六铭 做知会，消息发送成功
+
+### 3  [PASS]  ctx=37942  (acc=251669)  45769ms  tools=23
+- session.jsonl: harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_080/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✓ 使用用户身份创建一个名为「今晚吃什么」的群，预期返回 chat_id
+    ✓ 创建一张飞书卡片，卡片内容包含「今天晚上吃什么」
+    ✓ 将该卡片发送到新建群中，预期返回 message_id

harness-opt/rounds/round-003/attribution.json

@@ -0,0 +1,59 @@
+[
+  {
+    "case_id": "1",
+    "verdict": "PASS",
+    "verdict_note": "workorder=PASS（聚合口径）；判分点证据 3/3 ✗，按判分点当实质 FAIL 处理",
+    "token": 34555,
+    "duration_ms": 37000,
+    "tool_calls": 31,
+    "cmd_attempts": 7,
+    "cmd_failures": 5,
+    "cmd_fail_rate": 0.71,
+    "discoverability_state": "无（失败命令全是跨域 contact + auth，非 lark-im；chat-create.md 调用前已读但未走到使用）",
+    "axis": "效果",
+    "root_cause": "沙箱 user 授权不可完成 + 跨域 lark-contact 命令依赖；无 lark-im 文档根因，本轮不改",
+    "token_hotspot": "运行时冗余清单常驻（SKILL.md 3,456）+ 按需 chat-create.md 3,062（读了没用上）；lark-shared 3,751 与系统 Skill 列表注入 4,612 均不归因",
+    "token_reliability": "常驻静态（SKILL.md）/ 按需读取（chat-create.md，本题读了没用上）",
+    "duration_hotspot": "多轮交互（查联系人→切 contact→失败→授权→qrcode 重试）+ 纯外部API延迟(部分不可归因)",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现",
+    "doc_fix_hint": "效果侧无 lark-im 文档缺信息（环境+跨域）；token 侧 chat-create.md 把同组 flag 在 Commands/Usage Scenarios 重复演示、Common Errors 复述 validation 字符串，属可删冗余"
+  },
+  {
+    "case_id": "2",
+    "verdict": "PASS",
+    "verdict_note": "真 PASS，判分点 3/3 ✓，全程 bot 身份无授权阻断",
+    "token": 54568,
+    "duration_ms": 125000,
+    "tool_calls": 49,
+    "cmd_attempts": 11,
+    "cmd_failures": 3,
+    "cmd_fail_rate": 0.27,
+    "discoverability_state": "① 从没读（messages-search.md / chat-messages-list.md reach=0，直接猜命令；本题未读任何 lark-im reference）",
+    "axis": "token",
+    "root_cause": "无过滤 +chat-messages-list --page-all 全量拉取 → 43.5KB 输出被 Read 整文件灌入 22,556 tok；token 大头非 lark-im doc。修它需补收窄/前置内容，与降 token 目标方向冲突，列观察项",
+    "token_hotspot": "工具返回原样输出（block #19 单次 Read 22,556 tok / 51.5%，归「其他工具调用/返回」）",
+    "token_reliability": "单次输出（强依赖该群消息量，单题不可外推，非稳定常驻热点）",
+    "duration_hotspot": "多轮交互 + 重试（messages-search 连环 exit2 → page-all → 大输出 → 多次本地 grep 抠数据）",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现（model calls 16 作旁证，明显高于 080）",
+    "doc_fix_hint": "本题无 T1 可发力的 token 抓手（大头是单次工具输出，非 lark-im doc 常驻）；缺的是大群消息查询的 server-side 收窄指引，但补它=增内容、与降 token 反向，不作本轮根因"
+  },
+  {
+    "case_id": "3",
+    "verdict": "PASS",
+    "verdict_note": "真 PASS，判分点 3/3 ✓，主动选 bot 身份建群+发卡片均 ok:true，零命令失败",
+    "token": 38009,
+    "duration_ms": 47000,
+    "tool_calls": 22,
+    "cmd_attempts": 3,
+    "cmd_failures": 0,
+    "cmd_fail_rate": 0.0,
+    "discoverability_state": "无（无失败命令；SKILL.md + chat-create.md + messages-send.md 全部状态③：调用前已读且用上）",
+    "axis": "token",
+    "root_cause": "读取 Skill 占 56.4%；本轮唯一干净 token 抓手 = chat-create.md 内部冗余（示例罗列 + 场景重复 + --help 镜像），从未被优化过",
+    "token_hotspot": "运行时冗余清单常驻 + 按需 reference（chat-create.md 当前 2,336 raw，可压 Commands/Usage Scenarios 重叠 + Common Errors validation 镜像；trace 里 messages-send.md 5,365 是旧版，round-2 已压到 2,006，本轮不再可压）",
+    "token_reliability": "按需读取（chat-create.md reach=0.667，本题是其压缩收益唯一稳态兑现题）",
+    "duration_hotspot": "无离群（建群+发卡片正常串行，无重试、无写后回查）",
+    "duration_reliability": "耗时波动大，单次运行不算数，需多题或多次复现",
+    "doc_fix_hint": "chat-create.md 把同组 flag 在 Commands(12 例)+Usage Scenarios(3 场景)重复演示、Common Errors 多行复述 --help/报错本身就会吐的 validation 字符串，属可删冗余；232043 两步流 / --chat-mode topic 区分 / --owner 默认为载重红线，压缩中不可误删"
+  }
+]

harness-opt/rounds/round-003/case-commands.json

@@ -0,0 +1,27 @@
+{
+  "1": [
+    "auth login",
+    "auth qrcode",
+    "contact +search-user"
+  ],
+  "3": [
+    "auth login",
+    "auth qrcode",
+    "auth status",
+    "im +chat-create",
+    "im +messages-send"
+  ],
+  "2": [
+    "auth login",
+    "auth qrcode",
+    "auth status",
+    "im +chat-messages-list",
+    "im +chat-search",
+    "im +messages-mget",
+    "im +messages-search",
+    "im +messages-send",
+    "im messages forward",
+    "schema im.messages.forward",
+    "schema im.messages.search"
+  ]
+}

harness-opt/rounds/round-003/child-cache.json

@@ -0,0 +1,20 @@
+{
+  "1": {
+    "score": 1.0,
+    "passed": true,
+    "context_window": 33840,
+    "token_usage": 237434,
+    "duration_ms": 44127,
+    "tool_call_count": 25,
+    "feedback": "执行者成功完成了所有期望：首先搜索联系人获取 open_id（首次搜索用单字失败后改为双字搜索成功），然后使用 --as user 创建群组并添加成员，最后发送消息并返回 message_id。整个流程正确，使用了等效的 `--as user` 身份，符合用户「使用我的身份」的要求。验证结果确认所有操作均已生效。"
+  },
+  "3": {
+    "score": 1.0,
+    "passed": true,
+    "context_window": 35942,
+    "token_usage": 234388,
+    "duration_ms": 43185,
+    "tool_call_count": 22,
+    "feedback": "执行者正确理解用户意图，使用用户身份创建群并发送卡片消息。创建群组一次成功，发送卡片经历了4次格式试错（最初使用顶层 elements 和 tag:markdown，后通过查阅官方文档找到正确格式：body.elements + div + lark_md），最终成功发送并返回 message_id。试错后自行纠正符合评判原则，不构成判罚依据。\n- {'reason': '建议在 lark-im-messages-send.md 中增加飞书 interactive card 的标准格式示例，特别是 2.0 schema 下的 body.elements 中使用 div + lark_md 的正确写法，减少 AI 试错成本'}\n- {'reason': '建议 CLI 在遇到 230099 卡片格式错误时，尝试解析并返回更具体的字段级错误提示（如提示 \"elements 应在 body 内\" 或 \"tag:markdown 不被支持\"），帮助 AI 更快定位问题'}"
+  }
+}

harness-opt/rounds/round-003/diagnosis.md

@@ -0,0 +1,119 @@
+# Round 3 归因（parent=557349b…（round-2 已采纳候选）；候选模块见 candidate_modules，由 candidate-writer 据诊断+reach 点名）
+
+> 目标（objective.json）：**在不回退成功率的前提下降低 lark-im skill 文档的 token 成本**。effect 是硬门槛、不可退化；token 与 duration 是并列成本杆。tier=T1，仅可改 `skills/lark-im/**`。target_axis=token。
+> 判分点只当「什么算挂」的锚，不抄 grader 药方。
+
+## ⚠️ trace 与当前文件的版本错位（先看，决定本轮抓手是否还在）
+
+**本轮派工单 trace = round-1 的全 3 题 child-runs**（round-2 只评了 080，故用 round-1 作最近的全覆盖代理）。这些 trace 里的 reference 体积是 **round-1/round-2 改动之前** 的旧版。我用 session-analyze 所用的同一 ai-tokenizer 实测了**当前工作树**文件，确认两者错位如下：
+
+| 文件 | trace 内体积（旧版，Read 计） | 当前实测（raw / Read 计） | 已被哪轮收割 |
+|---|---|---|---|
+| `SKILL.md`（Skill 注入正文） | 3,455–3,456 tok | 3,525 raw | round-1（API Resources/权限表→schema 指针） |
+| `references/lark-im-messages-send.md` | **5,365 tok** | **2,006 raw / 2,194 Read** | **round-2（5,365→2,006，已收割）** |
+| `references/lark-im-chat-create.md` | 3,060–3,062 tok | **2,336 raw / 2,645 Read** | **未动过（2023 至今原样），唯一未收割** |
+
+**含义**：round-2 诊断里的 **RC-2（messages-send.md 内部冗余）已经在 round-2 被采纳并收割**（5,365→2,006），它不再是本轮抓手——不要据 trace 里的 5,365 重复提一遍。本轮 trace 里那块 5,365 是历史值，当前已不存在。**reach>0 集合里唯一还没被压过的干净文件就是 `chat-create.md`**（round-2 的 RC-3）。
+
+## 跨 case 共同根因（优先看；按对 TOKEN 目标的杠杆排序）
+
+### RC-1（token，本轮头号且基本是唯一的干净抓手，reach=0.667：014+080）—— `chat-create.md` 内部存在「示例罗列 + 场景重复 + --help 镜像」三类可压缩冗余，且从未被优化过
+- **现象**：`chat-create.md` 当前 2,336 raw tok（Read 计 ~2,645），是 reach>0 集合里**唯一未被任何轮收割**的 reference。section 级实测分布（raw tok）：
+
+  | section | tok | 性质 |
+  |---|---|---|
+  | header(1-11) | 198 | 载重（scope/映射），保留 |
+  | **Commands(12-50) 12 个 bash 示例** | **425** | **过度罗列**：多条仅差一个 flag（`--owner` / `--users` / `--bots` / `--as bot` / `--as user` / `--dry-run` 各一例），信息已在 Parameters 表里 |
+  | Parameters 表(52-69) | 500 | 多数载重；`--chat-mode` 的 L68 长注解与表内 L62 行语义重复 |
+  | AI Usage Guidance(70-108) | 442 | **载重**（232043 两步流是 080/014 路由依据），但表述偏长 |
+  | Output Fields(109-119) | 126 | 载重 |
+  | **Usage Scenarios(120-143) 3 个场景** | **198** | **重复**：Scenario 1/2 重复 Commands 已展示的 `--owner`/`--users`/`--bots` 组合；Scenario 3 重复 messages-send 的串联用法 |
+  | **Common Errors(144-158) 9 行** | **395** | **部分 --help 镜像**：多行直接复述确定性 validation 字符串（`--name exceeds 60`、`--users exceeds 50`、`invalid user id` 等），这些 `--help` / 报错本身就会原样吐出 |
+  | References(159-163) | 44 | 载重 |
+
+- **这正是 round-2 已经在 messages-send.md 上验证过、且被采纳的同一套压缩模式**：round-2 把 messages-send.md 的「4 处重复选型规则 + 全媒体形态 Commands + --help 镜像」压成「保留载重规则 + 一句 `--help` 指针」（5,365→2,006，被采纳）。chat-create.md 的 Commands(425)↔Usage Scenarios(198) 重叠、Common Errors(395) 的 validation 镜像，是同型冗余。
+- **可压缩量级（粗估，非药方）**：可压缩质量集中在 Commands+Usage Scenarios 的重叠（合计 ~623 tok，去重后可省一大半）+ Common Errors 的 --help 镜像行。**保守估计可从 2,336 压到 ~1,500–1,700 raw tok（省约 600–800 tok，约 30%）**，与 messages-send.md 的压缩比同量级。具体改法与确切降幅由 candidate-writer 决定、评测裁决。
+- **载重红线（candidate-writer 取舍时的 effect 风险点，不是 RC-1 不成立）**：AI Usage Guidance 的 **232043 两步流 + `succeed_type=1`**、`--chat-mode topic` vs 普通群+话题消息模式的区分、`--owner` 默认行为，是 014/080 走通 bot 身份建群的语义依据，**不能在压缩中误删**。这条 reference 被 080 实读且 080 据它建群成功（`ok:true`），所以 effect 风险真实存在——压的是示例/场景/报错镜像的体积，不是语义规则。
+- **axis=token**。可信度=**按需读取**（reach=0.667，子集=014+080，2 题）。压它的降幅只在这 2 题子集里计入，不被 015（没读它）稀释；但子集仅 2 题、且 014 是「读了没用上」（授权阻断没走到建群），实际吃到压缩收益的稳态题只有 080 一题——**证据基数小，降幅需评测在子集上确认**（见数据缺口）。
+
+### RC-2（token，已收割，本轮不再是抓手）—— messages-send.md 的内部冗余 round-2 已压掉
+- round-2 RC-2 已被采纳：messages-send.md 5,365→2,006 raw。**本轮不要据 trace 里的 5,365 重复提**。当前 messages-send.md 已是「载重规则 + `--help` 指针」形态，无明显二次压缩空间（剩余多为 content-flag 选型、@mention、media 约束等载重内容）。reach=0.333（仅 080）。
+
+### RC-3（token，无 T1 干净抓手）—— SKILL.md 常驻正文 round-1 已压过，剩余多为载重 identity/路由
+- SKILL.md 经 Skill 工具每题必加载（reach=1.0），当前 3,525 raw tok（round-1 已把 API Resources/权限表折叠成 schema 指针）。剩余 `## Important Notes`(L36–85) 各小节（Sender Name Resolution / message enrichment / `--download-resources` / Card / Flag / Feed Shortcut）与 `## Shortcuts` 全表(L87–115) 虽本轮 3 题低命中，但它们是**全域 identity/路由/约束**——这是 round-1 已经动过一刀的同一文件，**再压属同向继续、但删错会碰坏 015/080 已走通的 bot 身份与命令路由判断**（effect 风险高于 RC-1）。**列为更次级、风险更高的抓手**，不作为本轮首选；若要动须只删本轮已确证低命中且非路由的纯枚举行，谨慎程度高于 chat-create。
+
+## 命令失败热点（跨 case；失败类型由我从 timeline 命令串读出，非判决数字）
+
+| lark-cli 命令 | 失败次数 | 涉及题数 | 主要失败类型 | 指向的文档问题 |
+|---|---|---|---|---|
+| `contact +search-user` | 4 | 1 (014) | bot exit2(invalid_argument) ×2；`--as user` token_missing ×2 | **跨 lark-contact 域**，非 lark-im 内容 |
+| `auth qrcode --output <绝对/沙箱外路径>` | 1 | 1 (014) | unsafe output path，改相对路径重试成功 | 路径约束在 lark-shared（不可改） |
+| `im +messages-search` | 2 | 1 (015) | bot exit2 + `--as user` exit2 | 该命令 user-only（SKILL 表已注明）；bot 身份必败，agent 没看清就猜 |
+| `im +chat-messages-list --page-all` | 1 | 1 (015) | exit2（无过滤 page-all） | 见下「015 token 黑洞」 |
+- **解读**：本轮**没有一条 lark-im 命令因「参数名/类型写错」系统性失败**。080 三条命令 0 失败；014 的失败全在跨域 contact + auth；015 的失败集中在 messages-search（user-only，bot 必败）与无过滤 page-all。**没有 lark-im 侧常规「报错/参数整形」工单**——与 token 减体积方向一致，本轮抓手是减体积不是补内容。
+
+## 015 的 token 黑洞（与 round-2 一致，复述以免被误当成 token 抓手）
+- 015 真正的 token 大头**不是任何 lark-im doc**，而是 **block #19：一次 `Read` 工具读入 22,556 tok（占该题 visible 51.5%）**。成因链：#12/#17 `+messages-search`/`--page-all` exit2 → #18 退到 `+chat-messages-list`（无过滤）→ 输出 43.5KB 被持久化 → agent `Read` 整文件 → 22.5k tok 灌进上下文 → 再靠本地 grep(#27–33) 抠出「飞豆」两条。
+- **从文档角度**：`chat-messages-list.md` 本题 reach=0（状态①：调用前从没读），它本写了 `--start/--end`、`--page-size` 等可避免全量拉取的约束。**但补它=增常驻/触达内容，与本轮降 token 目标方向相反**（见方向冲突）；且 22.5k 是**单次工具输出**（强依赖该群消息量，单题不可外推），不是稳定常驻热点。**结论：观察项，交评测裁决，不作为本轮 token 抓手。**
+
+## 可发现性时序（约束 5 三态；判「前置能不能救」的决定性证据）
+> 对每条相关 reference / `--help`，按相对首次失败调用的读取时序统计。`--help` 扫 Bash（本轮 3 题均未跑任何 `--help`）。
+
+| reference / `--help` | 聚合 reach | ①从没读 | ②失败后才读 | ③读了仍错/卡 | 主导态 → 改动方向 |
+|---|---|---|---|---|---|
+| `lark-shared/SKILL.md` | 1.0 | 0 | 0 | — | 三题调用前都读了；014 仍卡（环境，非内容）；不可改 |
+| `chat-create.md` | 0.667 | 0 | 0 | — | 080 调用前读→建群成功；014 调用前读→授权阻断（非 reference 错）。**非触达问题，纯减体积** |
+| `messages-send.md` | 0.333 | 0 | 0 | — | 080 调用前读→发卡片成功。**非触达问题**（已收割） |
+| `chat-messages-list.md` | 0.0 | 1 (015) | 0 | — | ① 015 调用前从没读→`--page-all` 全量拉取→token 黑洞。触达缺口，但补它=增 token，与目标冲突 |
+| `messages-search.md` | 0.0 | 1 (015) | 0 | — | ① 015 从没读，直接猜 `+messages-search` ×2 → exit2（user-only，bot 必败） |
+- **结论**：**本轮没有「该前置」的干净 case**。RC-1（chat-create.md 减体积）是「调用前已读、内容够用 → 去冗余」的纯 token 减法，不涉及触达。015 的两处 ① 触达缺口确实存在，但修它们=增内容、与降 token 目标相反，且 015 已 PASS（bot + 本地 grep 兜底）——属观察项，**不要被诱导去推前置**。
+
+## 方向冲突记录（硬性约束 7）
+- **减体积（RC-1 chat-create.md，与 objective.direction 同向）** vs **补收窄/前置指引（修 015 chat-messages-list 全量灌入，与 objective 反向）**：前者降按需 token，后者为省「单次工具输出」反而要**增**文档常驻 token。两者方向相反，**不可合并**。本轮目标是降 token，取减体积一侧；015 全量灌入作为观察项记录、不作为要补的内容根因。
+
+## 差距台账复盘
+- 无（`discard-ledger.json` 为空，无已跑未采纳候选）。
+
+## 逐 case
+
+### 1 (014) [workorder=PASS / 实质 FAIL] token=34,555(reported)/visible 17,364 耗时=37s 命令失败率=5/7 维度=效果(不可修)
+- 判分点结果：3 条全 ✗——建群/拉人/发消息全未发生，卡在 `contact +search-user` 解析 open_id（user 授权阻断 + 跨域 contact）。verdict=PASS 系聚合口径，按判分点证据当 FAIL 处理。
+- 命令失败：5/7。`contact +search-user` bot exit2 ×2、`--as user` token_missing ×2；`auth qrcode` 绝对路径 unsafe ×1（改相对路径成功）。**全部非 lark-im 命令**。
+- 可发现性时序：#4 读 SKILL.md 正文(3,456) + #6 读 lark-shared(3,751，跨 skill) + #7 读 chat-create.md(3,062，调用前已读)；失败在更上游的跨域 contact + 授权。**非 lark-im 触达问题**。
+- token 归因：SKILL.md 正文 3,456（常驻静态，19.9%）+ lark-shared 3,751（**跨 skill，不归因 lark-im**）+ chat-create.md 3,062（按需，17.6%，**本题读了没用上**——授权阻断没走到建群）+ 系统 Skill 列表注入 4,612（固定开销，不归因）。lark-cli 命令累计含多次短失败回显，单条都短、非热点。
+- 耗时归因：本题往返多（查联系人→切 contact→失败→auth status→授权→qrcode 重试）；多为授权链路 + 跨域固有串行 + 反应式重试（duration 弱信号，需多轮复现）。
+- 文档根因：效果=沙箱 user 授权 + 跨域 contact（环境，**无 lark-im 文档根因，本轮不改**）；token=chat-create.md 按需冗余（RC-1，但本题读了没用上，收益只在 080 这种走通题里兑现）+ SKILL.md 常驻（RC-3，风险高、次级）。
+
+### 2 (015) [PASS·真] token=54,568(reported)/visible 43,760 耗时=2m5s 命令失败率=3/11 维度=token（但大头非 lark-im doc）
+- 判分点结果：3/3 ✓——定位群、转发「飞豆」消息、@傅六铭知会全部成功（两次 `messages-send` 均 `ok:true`）。**全程 bot 身份，无授权阻断**。
+- 命令失败：3/11。`+messages-search` bot exit2、`+messages-search --as user` exit2、`+chat-messages-list --page-all` exit2（无过滤）；agent 退到无 page-all + 本地 grep 兜底成功。（#14 `--page-all | grep` 返回空属「成功但无命中」，非硬失败，未计入。）
+- 可发现性时序：① `messages-search.md` / `chat-messages-list.md` 调用前从没读（reach=0），直接猜命令。**本题未读任何 lark-im reference**，故 lark-im reference 的体积与本题 token 无关。
+- token 归因：**本题 token 大头不是 lark-im doc**，是 block #19 一次 `Read` 持久化文件 = **22,556 tok（51.5%，归「其他工具调用/返回」）**，成因=`--page-all` 无过滤全量拉取→43.5KB→Read 灌入（**单次输出**可信度，强依赖该群消息量）。SKILL.md 正文 3,448（常驻）。lark-shared 3,749（跨 skill，不归因）。**RC-1 改 chat-create.md 对本题 token 无影响**（本题没读它）。
+- 耗时归因：本题最长(2m5s)，主因 messages-search 连环失败→改 page-all→大输出→多次本地 grep 抠数据的多轮往返（duration 弱信号；model calls 16/raw 32，明显高于 080，作旁证）。
+- 文档根因：token 黑洞的放大器=`chat-messages-list.md` 没被读到（状态①）+ SKILL.md 表未提示大群应 server-side 收窄——但**补这条与降 token 目标相反**（方向张力），列为观察项；本题已 PASS。本轮 token 抓手（RC-1）不落在本题。
+
+### 3 (080) [PASS·真] token=38,009(reported)/visible 21,599 耗时=47s 命令失败率=0/3 维度=token
+- 判分点结果：3/3 ✓——`auth status` 见 bot ready→主动选 bot→建群 `ok:true`→发 interactive 卡片 `ok:true`。**任务完整完成，零命令失败**。
+- 命令失败：0/3。三条 lark-cli（auth status / chat-create / messages-send）全成功。
+- 可发现性时序：#4 读 SKILL.md 正文(3,455) + #6 读 lark-shared(3,751，跨 skill) + #9 读 chat-create.md(3,060) + #10 读 messages-send.md(5,365，旧版) ，全部状态③（调用前已读且用上）。**无触达问题。** 实际只用了 `+chat-create --name … --format json` 的最简形态——没用两步流/owner/members/topic/error-recovery。
+- token 归因：**本题是纯 token 抓手题**——读取 Skill 占 56.4%：messages-send.md 5,365（trace 旧版，**当前已被 round-2 压到 2,006，本轮不再可压**）+ SKILL.md 正文 3,455（常驻，RC-3）+ chat-create.md 3,060（按需，**RC-1，当前 2,336，本轮唯一干净抓手**）。系统 Skill 列表注入 4,612（固定开销，不归因）。lark-shared 3,751（跨 skill，不归因）。
+- 耗时归因：47s，全部为正常建群+发卡片串行，无重试、无写后回查（无离群）。
+- 文档根因：无效果根因（已绿）；token=RC-1（chat-create.md 内部冗余，本题是其收益唯一稳态兑现题）+ RC-3（SKILL.md 常驻，风险高、次级）。**本题 token 杠杆最清晰且 effect 风险可控**（命令全成功，压 chat-create.md 的示例/场景/报错镜像不碰 080 实际用到的最简建群链路）。
+
+## 给 candidate-writer 的收口（不含具体改法）
+- **唯一在 T1 内还没被收割的干净 token 抓手是 RC-1（`chat-create.md` 内部冗余）**：Commands 12 例过度罗列 + Usage Scenarios 3 场景重复 Commands + Common Errors 9 行部分镜像 validation 字符串——**与 round-2 已采纳的 messages-send.md 压缩同型**，粗估可省 ~600–800 raw tok（约 30%）。reach=0.667（014+080），降幅在子集计入。
+- **载重红线**：AI Usage Guidance 的 232043 两步流 + `succeed_type=1` + `--chat-mode topic` 区分 + `--owner` 默认，是 080/014 走通 bot 建群的语义依据，**压缩中不可误删**——压的是示例/场景/报错镜像体积，不是规则。
+- **RC-2 已收割**：messages-send.md round-2 已 5,365→2,006，trace 里的 5,365 是历史值，**不要重复提**。
+- **RC-3（SKILL.md 常驻）是次级且风险更高**：round-1 已压过一刀，剩余多为全域 identity/路由/约束，删错碰坏 015/080 已走通的 bot 身份与命令路由——不作首选。
+- **不要推前置**：本轮没有「该前置」的干净 case。015 的两处 ① 触达缺口（chat-messages-list/messages-search 没读）虽真实，但修=增内容、与降 token 反向，且 015 已 PASS——属观察项。
+- **effect 不可在本轮 T1 内合法抬升**：014 是环境（沙箱不能扫码）+ 跨域 contact，无 lark-im 文档根因；015/080 已真 PASS。effect deltas 视作 auth-noise，不追。
+- **干净 token 抓手接近见底（诚实判断）**：reach>0 集合三个文件中，messages-send.md（round-2）与 SKILL.md（round-1）已各压一刀，**chat-create.md 是最后一个未动过的干净文件**。压完它之后，T1 内 reach>0 的纯冗余（罗列/重复/--help 镜像）基本耗尽；再往下只剩 (a) 高 effect 风险的 SKILL.md 载重内容，或 (b) reach=0 的 22 个盲区 reference（压了也不在判决集、无法被采纳）。**本轮 RC-1 很可能是这条优化路径上最后一个低风险、可被采纳的 token 抓手。**
+- **缺失信息（doc_fix_hint 语气，非药方）**：chat-create.md 把同一组 flag 在 Commands(12 例) 与 Usage Scenarios(3 场景) 重复演示、Common Errors 多行复述 `--help`/报错本身就会吐的 validation 字符串——这类「枚举/重复/镜像、低增量」内容是其 token 的主要去处，且是减法（删冗余）而非加法。
+
+## 数据缺口
+1. **trace 版本错位（最关键）**：本轮 trace=round-1 旧版 child-runs，messages-send.md 在 trace 里仍是 5,365（round-2 已压到 2,006）。所有「当前文件体积」结论我已用 ai-tokenizer 实测当前工作树校正（SKILL.md 3,525 / chat-create.md 2,336 / messages-send.md 2,006），但**单题行为与 reach 仍来自旧 trace**——若 round-2 改动改变了 080/014 的读取行为，需以实际 round-3 eval-run 复核。
+2. **RC-1 子集小**：chat-create.md reach=0.667 但实际吃到压缩收益的稳态题只有 080（014 读了没用上、授权阻断），证据基数=1，降幅需评测在子集确认。
+3. **015 的 22.5k 黑洞是单次工具输出**，强依赖该群消息量，非稳定常驻热点，单题不可外推；且与降 token 目标方向冲突，不作抓手。
+4. **duration 三题波动大**（37s/2m5s/47s），015 长尾主因 messages-search 连环失败+大输出多轮抠数据；单轮不足定论，需多轮复现。model calls(8/16/6) 比 wall-clock 稳，可作旁证。
+5. **工具调用口径不一致**：trend.json 的 R1 tool_calls=26.3、R2=10，与 session-analyze 的 model calls(8/16/6) 口径不同（趋势表疑似含 raw 计数）；旁证以 timeline 实际往返为准。趋势看：R1→R2 命令失败率 0.60→0.35、tool_calls 26→10 明显下降，但那主要是 effect 从「三题全卡授权」变成「2 真 PASS + 1 卡」带来的，**不是 token 改动的功劳**；token 均值 R1 31,997→R2 42,377 上升，主因是 R2 只评 080（单题大）口径差异 + 015 黑洞，非文档常驻变重——趋势对 token 轴判读价值有限，以单题 session-analyze 为准。
+6. **effect 维度全部归因为「无文档根因/不可修」**：014 跨域+环境，015/080 已绿。本轮 effect 无 T1 可发力点，deltas 视作 auth-noise。

harness-opt/rounds/round-003/discard-ledger.json

@@ -0,0 +1 @@
+[]

harness-opt/rounds/round-003/failure-memory.json

@@ -0,0 +1 @@
+[]

harness-opt/rounds/round-003/module-reach.json

@@ -0,0 +1,220 @@
+{
+  "skills/lark-im/SKILL.md": {
+    "reach": 1.0,
+    "read_cases": [
+      "1",
+      "2",
+      "3"
+    ],
+    "actual_cases": [
+      "1",
+      "2",
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": true
+  },
+  "skills/lark-im/references/lark-im-chat-create.md": {
+    "reach": 0.667,
+    "read_cases": [
+      "1",
+      "3"
+    ],
+    "actual_cases": [
+      "1",
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-identity.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-messages-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-search.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-chat-update.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-list-item.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-group-query-item.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-groups.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-create.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-feed-shortcut-remove.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-cancel.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-create.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-flag-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-message-enrichment.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-mget.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-reply.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-resources-download.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-search.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-messages-send.md": {
+    "reach": 0.333,
+    "read_cases": [
+      "3"
+    ],
+    "actual_cases": [
+      "3"
+    ],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-reactions.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  },
+  "skills/lark-im/references/lark-im-threads-messages-list.md": {
+    "reach": 0.0,
+    "read_cases": [],
+    "actual_cases": [],
+    "expected_cases": [],
+    "discoverability_miss": [],
+    "is_domain_skill": false
+  }
+}

harness-opt/rounds/round-003/review.json

@@ -0,0 +1,18 @@
+{
+  "generated_by": "lark-cli-harness:opt-reviewer",
+  "verdict": "PASS",
+  "module": "skills/lark-im/references/lark-im-chat-create.md",
+  "tier": "T1",
+  "round_index": 3,
+  "base_commit": "572eb8da41f608bd93b25916cac02cb772825b97",
+  "code_tip": "cbd6e56ac07285fd973c53ff7382da0112b6cf5d",
+  "reason": "纯瘦身改动，对抗式逐项核验未发现可证伪点：14 条承重红线（232043 两步流/succeed_type=1/chat-mode topic 与 thread 区分/--owner 默认/set-bot-manager/chat.members create/Output Fields/scopes）在改后文件全部 grep 命中；Scenario 3 建群→欢迎语 recipe 逐字保留仅换标题、搬迁到同文件 AI Usage Guidance 末尾未删；删掉的 6 行 Common Errors 已在 shortcuts/im/im_chat_create.go 源码核实是 CLI 原样回显的确定性 validation 字符串（运行时报错可复得，非仅靠 --help），删掉的命令例均为单 flag 变体且 flag 仍全列于 Parameters 表；字节 7996→6450(-19.3%)/词 1258→969(-23%) 为真实删减、无增读拉力、recipe 在同文件内搬迁不引发额外读；单根因 RC-1（本文件内部冗余），strategy 明确不捆 RC-2/RC-3。",
+  "dimensions": {
+    "reward_hack":      {"pass": true,  "evidence": "无硬编码评测答案/资源名/ID；未对 080 的 --name --format json 最简建群链做特判，080 链路一环未碰；属通用『删运行时另有出处的重复』瘦身，与 round-2 messages-send 同型同纪律，非针对某几题"},
+    "semantic_regress": {"pass": true,  "evidence": "14 条承重红线改后文件全部命中；Scenario 3 recipe 逐字保留（仅换标题、搬入 AI Usage Guidance）；删的 6 行报错经 im_chat_create.go 核实为 CLI verbatim validation（运行时可复得）；删的命令例均单 flag 变体、flag 仍全列于 Parameters 表，无承重内容丢失"},
+    "token_shift":      {"pass": true,  "evidence": "真实删减 bytes 7996→6450(-19.3%)、words -23%；纯删除无新增前置/『先读 X』拉力；welcome recipe 在同一文件内搬迁不触发额外读；唯一 --help 指针仅覆盖 Parameters 表已列的单 flag 变体，非强制增读。运行时每题 context 只降不升"},
+    "contract_break":   {"pass": true,  "evidence": "T1 文档无对外契约；结构完整（仅 Usage Scenarios 段：2 重复删、recipe 搬迁），所有 ## 章节与 References 链接保留，无断链/缺章"},
+    "devguide":         {"pass": true,  "evidence": "对照 review-rubric 优化红线（semantic_regress / contract_break 两维）：未删承重、未破坏结构；reference 收敛到 gotcha-only、与 --help/Parameters 重复内容下沉为指针，符合 optimization-playbook 的『单命令示例下沉、与 --help 重复留一处其余指针』；annotation 三段均标 R1 落在可重构范围、未触 R3 的 AI Usage Guidance prose"},
+    "single_root_cause":{"pass": true,  "evidence": "diff 只服务 RC-1（本文件内部『示例罗列+场景重复+报错镜像』三类冗余），全为同一根因下的去重；未捆 RC-2(messages-send)/RC-3(SKILL.md)，未夹带语义独立的承重删除，无多根因对冲叙事"}
+  }
+}

harness-opt/rounds/round-003/round.json

@@ -0,0 +1,394 @@
+{
+  "round": 3,
+  "status": "admitted",
+  "parent_id": "557349b40feb359bb791749a37571d59edb7e72e",
+  "parent_worktree": "/Users/bytedance/Projects/cli",
+  "child_worktree": "/Users/bytedance/Projects/cli",
+  "base_commit": "572eb8da41f608bd93b25916cac02cb772825b97",
+  "module": "skills/lark-im/references/lark-im-chat-create.md",
+  "candidate_modules": [
+    "skills/lark-im/SKILL.md",
+    "skills/lark-im/references/lark-im-chat-create.md",
+    "skills/lark-im/references/lark-im-chat-identity.md",
+    "skills/lark-im/references/lark-im-chat-list.md",
+    "skills/lark-im/references/lark-im-chat-messages-list.md",
+    "skills/lark-im/references/lark-im-chat-search.md",
+    "skills/lark-im/references/lark-im-chat-update.md",
+    "skills/lark-im/references/lark-im-feed-group-list-item.md",
+    "skills/lark-im/references/lark-im-feed-group-list.md",
+    "skills/lark-im/references/lark-im-feed-group-query-item.md",
+    "skills/lark-im/references/lark-im-feed-groups.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-create.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-list.md",
+    "skills/lark-im/references/lark-im-feed-shortcut-remove.md",
+    "skills/lark-im/references/lark-im-flag-cancel.md",
+    "skills/lark-im/references/lark-im-flag-create.md",
+    "skills/lark-im/references/lark-im-flag-list.md",
+    "skills/lark-im/references/lark-im-message-enrichment.md",
+    "skills/lark-im/references/lark-im-messages-mget.md",
+    "skills/lark-im/references/lark-im-messages-reply.md",
+    "skills/lark-im/references/lark-im-messages-resources-download.md",
+    "skills/lark-im/references/lark-im-messages-search.md",
+    "skills/lark-im/references/lark-im-messages-send.md",
+    "skills/lark-im/references/lark-im-reactions.md",
+    "skills/lark-im/references/lark-im-threads-messages-list.md"
+  ],
+  "module_reach": {
+    "skills/lark-im/SKILL.md": {
+      "reach": 1.0,
+      "read_cases": [
+        "1",
+        "2",
+        "3"
+      ],
+      "actual_cases": [
+        "1",
+        "2",
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": true
+    },
+    "skills/lark-im/references/lark-im-chat-create.md": {
+      "reach": 0.667,
+      "read_cases": [
+        "1",
+        "3"
+      ],
+      "actual_cases": [
+        "1",
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-identity.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-messages-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-search.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-chat-update.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-list-item.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-group-query-item.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-groups.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-create.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-feed-shortcut-remove.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-cancel.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-create.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-flag-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-message-enrichment.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-mget.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-reply.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-resources-download.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-search.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-messages-send.md": {
+      "reach": 0.333,
+      "read_cases": [
+        "3"
+      ],
+      "actual_cases": [
+        "3"
+      ],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-reactions.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    },
+    "skills/lark-im/references/lark-im-threads-messages-list.md": {
+      "reach": 0.0,
+      "read_cases": [],
+      "actual_cases": [],
+      "expected_cases": [],
+      "discoverability_miss": [],
+      "is_domain_skill": false
+    }
+  },
+  "expected_reach": {},
+  "minibatch": [
+    "1",
+    "2",
+    "3"
+  ],
+  "pareto_cases": [
+    "1",
+    "2",
+    "3"
+  ],
+  "artifacts": {
+    "workorder": "workorder.md",
+    "diagnosis": "diagnosis.md",
+    "attribution": "attribution.json",
+    "strategy": "strategy.md",
+    "review": "review.json",
+    "trend": "trend.json"
+  },
+  "code_tip": "cbd6e56ac07285fd973c53ff7382da0112b6cf5d",
+  "signature": "53194d7a111df326cc078b633f43587225bd0132",
+  "tier": "T1",
+  "intent": "dedup Commands<->Usage Scenarios overlap + compress --help-mirroring Common Errors in chat-create.md; keep all red lines (232043 two-step,succeed_type=1,chat-mode topic,--owner)",
+  "target_axis": "token",
+  "changed_files": [
+    "skills/lark-im/references/lark-im-chat-create.md"
+  ],
+  "decision_basis": {
+    "type": "module",
+    "module": "skills/lark-im/references/lark-im-chat-create.md"
+  },
+  "decision_cases": [
+    "1",
+    "3"
+  ],
+  "review": {
+    "generated_by": "lark-cli-harness:opt-reviewer",
+    "verdict": "PASS",
+    "module": "skills/lark-im/references/lark-im-chat-create.md",
+    "tier": "T1",
+    "round_index": 3,
+    "base_commit": "572eb8da41f608bd93b25916cac02cb772825b97",
+    "code_tip": "cbd6e56ac07285fd973c53ff7382da0112b6cf5d",
+    "reason": "纯瘦身改动，对抗式逐项核验未发现可证伪点：14 条承重红线（232043 两步流/succeed_type=1/chat-mode topic 与 thread 区分/--owner 默认/set-bot-manager/chat.members create/Output Fields/scopes）在改后文件全部 grep 命中；Scenario 3 建群→欢迎语 recipe 逐字保留仅换标题、搬迁到同文件 AI Usage Guidance 末尾未删；删掉的 6 行 Common Errors 已在 shortcuts/im/im_chat_create.go 源码核实是 CLI 原样回显的确定性 validation 字符串（运行时报错可复得，非仅靠 --help），删掉的命令例均为单 flag 变体且 flag 仍全列于 Parameters 表；字节 7996→6450(-19.3%)/词 1258→969(-23%) 为真实删减、无增读拉力、recipe 在同文件内搬迁不引发额外读；单根因 RC-1（本文件内部冗余），strategy 明确不捆 RC-2/RC-3。",
+    "dimensions": {
+      "reward_hack": {
+        "pass": true,
+        "evidence": "无硬编码评测答案/资源名/ID；未对 080 的 --name --format json 最简建群链做特判，080 链路一环未碰；属通用『删运行时另有出处的重复』瘦身，与 round-2 messages-send 同型同纪律，非针对某几题"
+      },
+      "semantic_regress": {
+        "pass": true,
+        "evidence": "14 条承重红线改后文件全部命中；Scenario 3 recipe 逐字保留（仅换标题、搬入 AI Usage Guidance）；删的 6 行报错经 im_chat_create.go 核实为 CLI verbatim validation（运行时可复得）；删的命令例均单 flag 变体、flag 仍全列于 Parameters 表，无承重内容丢失"
+      },
+      "token_shift": {
+        "pass": true,
+        "evidence": "真实删减 bytes 7996→6450(-19.3%)、words -23%；纯删除无新增前置/『先读 X』拉力；welcome recipe 在同一文件内搬迁不触发额外读；唯一 --help 指针仅覆盖 Parameters 表已列的单 flag 变体，非强制增读。运行时每题 context 只降不升"
+      },
+      "contract_break": {
+        "pass": true,
+        "evidence": "T1 文档无对外契约；结构完整（仅 Usage Scenarios 段：2 重复删、recipe 搬迁），所有 ## 章节与 References 链接保留，无断链/缺章"
+      },
+      "devguide": {
+        "pass": true,
+        "evidence": "对照 review-rubric 优化红线（semantic_regress / contract_break 两维）：未删承重、未破坏结构；reference 收敛到 gotcha-only、与 --help/Parameters 重复内容下沉为指针，符合 optimization-playbook 的『单命令示例下沉、与 --help 重复留一处其余指针』；annotation 三段均标 R1 落在可重构范围、未触 R3 的 AI Usage Guidance prose"
+      },
+      "single_root_cause": {
+        "pass": true,
+        "evidence": "diff 只服务 RC-1（本文件内部『示例罗列+场景重复+报错镜像』三类冗余），全为同一根因下的去重；未捆 RC-2(messages-send)/RC-3(SKILL.md)，未夹带语义独立的承重删除，无多根因对冲叙事"
+      }
+    }
+  },
+  "child_k": 5,
+  "eval_trace": null,
+  "retro": {
+    "cause": "已入池",
+    "noise_borderline": false,
+    "summary": "越带入池，无需复盘补发"
+  },
+  "retro_sessions": [
+    {
+      "case": "1",
+      "session": "harness-opt/rounds/round-003/child-runs/run-1/detail_info/cases/CLI_核心评测_014/0/session.jsonl",
+      "axis": "token",
+      "expect": "降",
+      "parent": 34270,
+      "child": 33840,
+      "gain": "收益现",
+      "pass_delta": null
+    },
+    {
+      "case": "3",
+      "session": null,
+      "axis": "token",
+      "expect": "降",
+      "parent": 35478,
+      "child": 35942,
+      "gain": "反向",
+      "pass_delta": null
+    }
+  ],
+  "verdict": "admitted",
+  "ci": null,
+  "new_candidate": "53194d7a111df326cc078b633f43587225bd0132",
+  "decision": {
+    "parent_success": 1.0,
+    "child_success": 1.0,
+    "parent_score": 0.8,
+    "child_score": 1.0,
+    "score_saved": 0.19999999999999996,
+    "score_threshold": 0.09532271373123208,
+    "parent_token": 34874.0,
+    "child_token": 34891.0,
+    "saved": -17.0,
+    "threshold": 4532.708313776408,
+    "parent_duration": 45267.5,
+    "child_duration": 43656.0,
+    "dur_saved": 1611.5,
+    "dur_threshold": 4899.200953624988,
+    "dur_margin": 1.0,
+    "missing_duration": [],
+    "k_child": 5,
+    "k_parent": 5,
+    "decision_n": 2,
+    "missing_context": [],
+    "missing_score": [],
+    "parent_token_acc": 248146.5,
+    "child_token_acc": 235911.0,
+    "phi0_score": 0.5666666666666667,
+    "eff_margin": 1.0,
+    "parent_token_full": 34874.0,
+    "child_token_full": 34891.0,
+    "saved_full": -17.0,
+    "observe_n": 2,
+    "target_axis": "token",
+    "admitted": true,
+    "reason": "score_gain"
+  },
+  "patch": "verify_results/round-003-lark-im-references-lark-im-chat-create.patch"
+}

harness-opt/rounds/round-003/strategy.md

@@ -0,0 +1,43 @@
+# Round 3 候选策略（模块=references/lark-im-chat-create.md, tier=T1, 主指标=token）
+
+## 根因与选择
+| 根因 | 来源(评测归因/规范经验) | 承载模块(reach) | annotation 风险级 | coverage 档 | P级 | 选中 |
+|---|---|---|---|---|---|---|
+| RC-1 chat-create.md 内部「示例罗列+场景重复+--help 镜像」三类冗余 | 评测归因 + 规范经验（双视角同指） | references/lark-im-chat-create.md(0.667) | Commands/Scenarios/Errors 段=R1 | 中（014+080 子集，稳态兑现仅 080） | P0 | ✅ |
+| RC-2 messages-send.md 内部冗余 | 评测归因 | references/lark-im-messages-send.md(0.333) | — | — | — | （round-2 已收割，不再是抓手） |
+| RC-3 SKILL.md 常驻正文 | 评测归因 | SKILL.md(1.0) | 多为 R2/R3 路由·identity | — | — | （round-1 已压一刀；剩余多为全域路由/身份，effect 风险高，不选） |
+
+- 选中理由：RC-1 是 reach>0 集合里**唯一从未被任何轮收割的干净文件**（2023 至今原样），且其冗余型态与 round-2 已采纳并 PASS 的 messages-send.md 完全同型（罗列+重复+--help 镜像）。RC-2 已在 round-2 收割（5,365→2,006），trace 里的 5,365 是历史值；RC-3 是 round-1 已动过的同一文件、剩余多为全域 identity/路由（删错碰坏 015/080 已走通的身份与路由判断，effect 风险高于 RC-1），故不选。
+- 选模块理由：chat-create.md reach=0.667（014+080 调用前都读到，状态③，非触达问题——纯减体积场景）；它正是承载 RC-1 的文件。未选 reach=0 的 22 个盲区 reference（改了也不在判决集、无法被采纳，触 reach 锁）。
+- 规范经验源补注：双视角同指一处。视角②（skill-annotations）独立把 Commands(L11-50)/Usage Scenarios(L120-143)/Common Errors(L144-158) 全标为 **R1（可重构）**，把 AI Usage Guidance(L70-98) 标为 **R3（需强理由）**——与归因的「压示例/场景/报错镜像、绝不碰 232043 两步流」完全吻合。对照 reviewer optimization-playbook：单命令用法/示例属 USAGE→下沉 `--help`；与 `--help` 重复的 validation 字符串「留命中率最高一处，其余删/指针」。当轮 eval 可在 080 子集裁出 token 真伪（080 调用前读、建群成功），但稳态收益基数仅 1 题（014 读了没用上）——敞口已在「预期效果」标明。
+
+## 改了什么（逐处）
+- **Commands(L12-50)** — 12 个 bash 示例（多条仅差一个 flag）压成 5 个差异化示例 + 一行 `--help` 指针。之前→之后：删掉 `--owner`/`--users`/`--bots`/`--as bot`/`--as user`/`--dry-run`/`--format json` 各单独一例（信息已在 Parameters 表），合并为「invite members+owner 一例」「bot+set-bot-manager 一例」，单 flag 变体一行指针带过（含 `--dry-run` 语义保留）。
+- **Usage Scenarios(L120-143)** — 整段 3 场景删除/搬迁。Scenario 1（owner）、Scenario 2（users+bots+owner）重复 Commands 与 Parameters 已展示的 flag 组合 → 删；Scenario 3（建群→发欢迎语链）是独有 recipe → 搬进 AI Usage Guidance 末尾「Create a group, then send a welcome message」保留。
+- **Common Errors(L144-158)** — 9 行压成 2 行。删掉 6 行直接复述 CLI 确定性 validation 字符串的行（`--name`/`--description` 超长、`--users`/`--bots` 超数、3 条 `ou_xxx`/`cli_xxx` 格式错）——这些 `--help`/报错本身原样吐出，改为一句「format/limit validation 由 CLI 原样回显，limits 见 Parameters 表」的指针；**保留**需要额外动作的 2 行：Permission denied(99991672) 给 console action、`bot is invisible(232043)` 指回两步流。
+
+## 为什么这么改（机制）
+- **省 context 的因果**：chat-create.md 是 lazy reference，读到即整文件进窗口（080/014 reach=0.667）。删掉的全是运行时另有出处（`--help`/Parameters 表）或本段内重复的内容——示例的 flag 组合 = Parameters 表已列；validation 字符串 = CLI 报错原样吐。删后 Agent 仍能：经 SKILL.md 选对 `+chat-create`、经 Parameters 表/`--help` 补全 flag 用法、遇 232043 走两步流。即 optimization-playbook §13 核心判据「删掉后 Agent 是否仍能选对命令并补到用法」——成立。
+- **规范经验源**：optimization-playbook「reference 收敛到 gotcha-only，不是 --help 镜像」「单命令用法/示例→下沉」「与 --help 重复→留一处其余指针」；content-taxonomy 单命令示例=R1 下沉、与 --help 重复=R0/指针。annotation 三段独立标 R1，本改动落在 R1 重构范围内，未触 R3 段。
+
+## 预期效果
+- 成功率（effect，硬门槛）：**不退化**。所有 effect 红线逐条保留并已 grep 校验（见下「刻意没做什么」）。080 实际只用 `--name --format json` 最简建群链——本改动未碰该链路任何一环；014 卡在跨域 contact+授权（非本 reference）。
+- context（分两层）：
+  - **(1) 静态字数差**：bytes 7996→6450（-19.3%）、chars -19.5%、words -23.0%、tiktoken(cl100k 代理) 2125→1714(-19.3%)。换算到 diagnosis 用的 ai-tokenizer 基线（OLD=2336 raw）：**预计 NEW ≈ 1800–1900 raw tok，省 ~450–540 tok（约 19–23%）**。
+  - **(2) 运行时 context 方向**：对**读到 chat-create.md 的题（080，及理论上 014）下降** ~450–540 tok；对没读它的题（015）**无影响**（015 大头是单次 `Read` 22.5k 工具输出，与本 reference 无关）。本改动是纯删减、无新增前置/增读拉力，不会抬升运行时 token。
+  - **与 direction 一致**（objective=降 token），无张力。
+- **覆盖敞口（诚实标注）**：稳态吃到收益的题只有 080 一题（014 读了没用上、授权阻断未走到建群），证据基数=1；且本轮派工单 trace 是 round-1 旧版 child-runs，单题读取行为需 round-3 实跑 eval 在 014+080 子集复核。实际降幅（~450–540）略低于 diagnosis 的 ~600–800 估计——因我**刻意保留**了 AI Usage Guidance 全段 prose（R3）+ 完整 Parameters/Output Fields 表（载重），用一点压缩头寸换零 effect 风险。
+
+## 刻意没做什么（反 reward-hack / 反过拟合）
+- 没硬编码任何评测题答案；没删任何承重内容；没碰本 skill 以外的文件、没把无关根因捆进本轮。
+- **逐条保留的载重红线（已 grep 校验存在）**：
+  - 232043 两步流全段（contact search → `--users 当前用户` 建群 → `chat.members create --as user` 加其他人 → 查 `invalid_id_list`）；
+  - `succeed_type=1` 语义解释；
+  - `--chat-mode topic` vs 「普通群 + `group_message_type=thread`」的区分注解；
+  - `--owner` 默认行为（bot 身份默认 bot / user 身份默认授权用户）；
+  - 全部 flag（含 `--set-bot-manager`、`--dry-run`、`--type public`、`--users/--bots` 上限与格式）、identity/scope 指引、互斥与护栏规则、Output Fields 全表。
+- 本改动**不是**按评测错误分布反推的拟合型改动——它是「删运行时另有出处的重复/镜像」的通用瘦身，与 round-2 messages-send.md 同型同纪律；非针对某几题的特判。
+- 未做 RC-3（SKILL.md 进一步压缩）：剩余多为全域 identity/路由，删错有 effect 风险，超出本轮低风险抓手范围。未做 015 的前置补充：那是增内容、与降 token 反向（方向冲突，diagnosis 已记录）。
+
+## 签名
+- signature: 见 commit sha（git diff: 18 insertions / 60 deletions on lark-im-chat-create.md）  tier: T1

harness-opt/rounds/round-003/trend.json

@@ -0,0 +1,20 @@
+[
+  {
+    "round": 1,
+    "n": 3,
+    "pass_n": 0,
+    "cmd_fail_rate": 0.6,
+    "tool_calls": 26.333333333333332,
+    "duration_ms": 50189.0,
+    "token": 31997.0
+  },
+  {
+    "round": 2,
+    "n": 3,
+    "pass_n": 3,
+    "cmd_fail_rate": 0.3466666666666667,
+    "tool_calls": 10.0,
+    "duration_ms": 69666.66666666667,
+    "token": 42377.333333333336
+  }
+]

harness-opt/rounds/round-003/workorder.md

@@ -0,0 +1,44 @@
+# Round 3 归因派工单（parent=557349b40feb359bb791749a37571d59edb7e72e；模块未定，由 candidate-writer 据诊断点名）
+
+> **只读输入**——opt-attributor 读本文件，把诊断**另写** `diagnosis.md`（给 candidate-writer）+ 逐题结构化 `attribution.json`（给 dashboard）。**不要覆盖本文件**，留作派工单↔诊断的前后对比。
+> 判分点只当「什么算挂」的锚，禁止照抄 grader 药方（已从派工单剔除）。
+
+## 模块运行时可达性（选模块第一步的证据；要选须在 strategy.md 说明理由）
+> reach=**实测**触达率（域主 SKILL.md 经 Skill 工具加载、reference 经 Read，都从 trace 实测，没有恒在的面）；判决集=实测∪预期触达。**实测低但有预期触达 ⚠️=可发现性/路由根因**（本该读却没读，如没路由到该域 / 速查表漏链接 / 该前置），正该选来修——不是白烧；reach=0 且无预期 才是真白烧。 **别用「全集均摊」判 reference 价值**：判决在 reach 子集上做，压一条 reference 的降幅在它子集里不被没读它的题稀释——reach 不高(但 >0)的 reference 在自己子集上也可能越带。
+- `skills/lark-im/SKILL.md` → reach=1.0  [域主 skill·经 Skill 工具加载]；判决集(实测∪预期): ['1', '2', '3']
+- `skills/lark-im/references/lark-im-chat-create.md` → reach=0.667；判决集(实测∪预期): ['1', '3']
+- `skills/lark-im/references/lark-im-messages-send.md` → reach=0.333；判决集(实测∪预期): ['3']
+- （另 22 个 reference reach=0 且无预期触达，本轮无关，略）
+
+## 逐轮诊断信号趋势（纯诊断，不进判决）
+
+| 轮 | 题数 | PASS | 命令失败率 | 工具调用 | 耗时(ms) | token |
+|---|---|---|---|---|---|---|
+| R1 | 3 | 0 | 0.60 | 26 | 50189 | 31997 |
+| R2 | 3 | 3 | 0.35 | 10 | 69667 | 42377 |
+
+> 跨题均值，按轮排。**命令失败率、工具调用数是横切诊断信号，不是准入轴**（准入只走 效果/token/耗时）——用来判「上一轮那刀有没有把失败/轮次压下去」。工具调用数比 wall-clock 稳，可给噪声大的耗时轴当旁证。
+
+### 1  [PASS]  ctx=34270  (acc=274608)  43995ms  tools=31
+- session.jsonl: harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_014/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✗ 使用当前用户身份创建名为「IM合作群」的群聊
+        证据: transcript 在展示授权二维码后结束，无任何 `lark-cli im +chat-create` 调用。执行停在 '授权完成后请告诉我，我会继续帮你创建群聊并发送消息'，群聊未创建。
+    ✗ 将傅一铭和傅二铭加入该群
+        证据: transcript 显示尝试搜索用户时遇到 `need_user_authorization` 错误，授权流程启动后中断。未获取到任何用户的 open_id，无后续添加操作。
+    ✗ 在该群发送文本消息「大家体验有问题随时沟通」，并返回可验证的 chat_id / message_id
+        证据: 群聊未创建，无 chat_id 可返回。transcript 无任何 `lark-cli im messages-send` 调用。
+
+### 2  [PASS]  ctx=47116  (acc=612048)  114310ms  tools=49
+- session.jsonl: harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_015/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✓ 成功定位名为「fusanming_at_openclaw群」的群，并获取最近包含「飞豆」关键字的消息
+    ✓ 将筛选出的相关消息内容转发到「fusanming_at_需求测试群」
+    ✓ 在「fusanming_at_需求测试群」中 @傅六铭 做知会，消息发送成功
+
+### 3  [PASS]  ctx=35478  (acc=221685)  46540ms  tools=22
+- session.jsonl: harness-opt/rounds/round-001/child-runs/run-1/detail_info/cases/CLI_核心评测_080/0/session.jsonl [native]
+- 判分点（grader 的「什么算挂」oracle，非药方）:
+    ✓ 使用用户身份创建一个名为「今晚吃什么」的群，预期返回 chat_id
+    ✓ 创建一张飞书卡片，卡片内容包含「今天晚上吃什么」
+    ✓ 将该卡片发送到新建群中，预期返回 message_id

harness-opt/trend.json

@@ -0,0 +1,20 @@
+[
+  {
+    "round": 1,
+    "n": 3,
+    "pass_n": 0,
+    "cmd_fail_rate": 0.6,
+    "tool_calls": 26.333333333333332,
+    "duration_ms": 50189.0,
+    "token": 31997.0
+  },
+  {
+    "round": 2,
+    "n": 3,
+    "pass_n": 3,
+    "cmd_fail_rate": 0.3466666666666667,
+    "tool_calls": 10.0,
+    "duration_ms": 69666.66666666667,
+    "token": 42377.333333333336
+  }
+]

harness-opt/verify_results/round-001-lark-im-SKILL.patch

@@ -0,0 +1,152 @@
+From 237a77feb341e15656386d6952a875dc459fec8c Mon Sep 17 00:00:00 2001
+From: "zhangheng.023" <zhangheng.023@bytedance.com>
+Date: Tue, 23 Jun 2026 18:27:25 +0800
+Subject: [PATCH] =?UTF-8?q?opt(round-001):=20SKILL.md=20=E2=80=94=20fold?=
+ =?UTF-8?q?=20USAGE=20method-index=20+=20scope=20table=20into=20a=20schema?=
+ =?UTF-8?q?=20pointer=20(-40%=20resident=20tokens)?=
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+---
+ skills/lark-im/SKILL.md | 122 +++-------------------------------------
+ 1 file changed, 8 insertions(+), 114 deletions(-)
+
+diff --git a/skills/lark-im/SKILL.md b/skills/lark-im/SKILL.md
+index bc39aae1..ac1c6900 100644
+--- a/skills/lark-im/SKILL.md
++++ b/skills/lark-im/SKILL.md
+@@ -110,122 +110,16 @@ Shortcut 是对常用操作的高级封装（`lark-cli im +<verb> [flags]`）。
+ | [`+feed-group-list`](references/lark-im-feed-group-list.md) | List the caller's feed groups (tags); user-only; supports `--page-all` auto-pagination |
+ | [`+feed-group-list-item`](references/lark-im-feed-group-list-item.md) | List feed cards in a feed group (tag); user-only; enriches each item with chat_name resolved from feed_id; supports --page-all auto-pagination |
+ | [`+feed-group-query-item`](references/lark-im-feed-group-query-item.md) | Look up specific feed cards in a feed group (tag) by ID; user-only; enriches each item with chat_name resolved from feed_id |
++| `reactions.*` (add / delete / list / batch_query) | Add, remove, or read emoji reactions on a message; user/bot; caller must be in the conversation, and can only delete its own reactions. Read first: [`lark-im-reactions.md`](references/lark-im-reactions.md) |
++| `feed.groups.*` (create / update / delete / batch_query / batch_add_item / batch_remove_item) | Manage feed groups (tags) and their member cards; user-only. Read first: [`lark-im-feed-groups.md`](references/lark-im-feed-groups.md) |
+ 
+-## API Resources
++## Native API (beyond shortcuts)
++
++Anything not covered by a shortcut above (e.g. `chats.*`, `chat.members.*`, `chat.managers.*`, `chat.moderation.*`, `chat.user_setting.*`, `messages.delete|forward|merge_forward|read_users|urgent_*`, `threads.forward`, `images.create`, `pins.*`) is callable as a raw API:
+ 
+ ```bash
+-lark-cli schema im.<resource>.<method>   # 调用 API 前必须先查看参数结构
+-lark-cli im <resource> <method> [flags] # 调用 API
++lark-cli schema im.<resource>.<method>    # MUST run first — gives params, identity (user/bot/tenant), and required scope
++lark-cli im <resource> <method> [flags]   # then call
+ ```
+ 
+-> **重要**：使用原生 API 时，必须先运行 `schema` 查看 `--data` / `--params` 参数结构，不要猜测字段格式。
+-
+-### chats
+-
+-  - `create` — 创建群。Identity: `bot` only (`tenant_access_token`).
+-  - `get` — 获取群信息。Identity: supports `user` and `bot`; the caller must be in the target chat to get full details, and must belong to the same tenant for internal chats.
+-  - `link` — 获取群分享链接。Identity: supports `user` and `bot`; the caller must be in the target chat, must be an owner or admin when chat sharing is restricted to owners/admins, and must belong to the same tenant for internal chats.
+-  - `update` — 更新群信息。Identity: supports `user` and `bot`.
+-
+-### chat.members
+-
+-  - `bots` — 获取群内机器人列表。Identity: supports `user` and `bot`; the caller must be in the target chat and must belong to the same tenant for internal chats.
+-  - `create` — 将用户或机器人拉入群聊。Identity: supports `user` and `bot`; the caller must be in the target chat; for `bot` calls, added users must be within the app's availability; for internal chats the operator must belong to the same tenant; if only owners/admins can add members, the caller must be an owner/admin, or a chat-creator bot with `im:chat:operate_as_owner`.
+-  - `delete` — 将用户或机器人移出群聊。Identity: supports `user` and `bot`; only group owner, admin, or creator bot can remove others; max 50 users or 5 bots per request.
+-  - `get` — 获取群成员列表。Identity: supports `user` and `bot`; the caller must be in the target chat and must belong to the same tenant for internal chats.
+-
+-### chat.user_setting
+-
+-  - `batch_query` — 批量查询当前用户在群内的个人偏好设置 (e.g. `is_muted` mutes normal messages, `is_mute_at_all` mutes @all messages); up to 10 chats per request. Identity: `user` only (`user_access_token`); the caller must be in each target chat.
+-  - `batch_update` — 批量更新当前用户在群内的个人偏好设置 (e.g. `is_muted` mutes normal messages, `is_mute_at_all` mutes @all messages); up to 10 chats per request. Identity: `user` only (`user_access_token`); the caller must be in each target chat.
+-
+-### chat.managers
+-
+-  - `add_managers` — 指定群管理员。Identity: supports `user` and `bot`; only the group owner can add managers; max 10 managers per chat (20 for super-large chats), and at most 5 bots per request.
+-  - `delete_managers` — 删除群管理员。Identity: supports `user` and `bot`; only the group owner can remove managers; max 50 users or 5 bots per request.
+-
+-### chat.moderation
+-
+-  - `get` — 获取群成员发言权限。Identity: supports `user` and `bot`; the caller must be in the target chat and belong to the same tenant.
+-  - `update` — 更新群发言权限。Identity: supports `user` and `bot`; only the group owner (or creator bot with `im:chat:operate_as_owner`) can update; the caller must be in the chat.
+-
+-### messages
+-
+-  - `delete` — 撤回消息。Identity: supports `user` and `bot`; for `bot` calls, the bot must be in the chat to revoke group messages; to revoke another user's group message, the bot must be the owner, an admin, or the creator; for user P2P recalls, the target user must be within the bot's availability.
+-  - `forward` — 转发消息。Identity: supports `user` and `bot`.
+-  - `merge_forward` — 合并转发消息。Identity: `bot` only (`tenant_access_token`).
+-  - `read_users` — 查询消息已读信息。Identity: `bot` only (`tenant_access_token`); the bot must be in the chat, and can only query read status for messages it sent within the last 7 days.
+-  - `urgent_app` — 发送应用内加急。Identity: `bot` only (`tenant_access_token`); the bot must be the message sender and must be in the conversation that contains the message.
+-  - `urgent_phone` — 发送电话加急。Identity: `bot` only (`tenant_access_token`); the bot must be the message sender and must be in the conversation that contains the message.
+-  - `urgent_sms` — 发送短信加急。Identity: `bot` only (`tenant_access_token`); the bot must be the message sender and must be in the conversation that contains the message.
+-
+-### reactions
+-
+-  - `batch_query` — 批量获取消息表情。Identity: supports `user` and `bot`.[Must-read](references/lark-im-reactions.md)
+-  - `create` — 添加消息表情回复。Identity: supports `user` and `bot`; the caller must be in the conversation that contains the message.[Must-read](references/lark-im-reactions.md)
+-  - `delete` — 删除消息表情回复。Identity: supports `user` and `bot`; the caller must be in the conversation that contains the message, and can only delete reactions added by itself.[Must-read](references/lark-im-reactions.md)
+-  - `list` — 获取消息表情回复。Identity: supports `user` and `bot`; the caller must be in the conversation that contains the message.[Must-read](references/lark-im-reactions.md)
+-
+-### threads
+-
+-  - `forward` — 转发话题。Identity: supports `user` and `bot`.
+-
+-### images
+-
+-  - `create` — 上传图片。Identity: `bot` only (`tenant_access_token`).
+-
+-### pins
+-
+-  - `create` — Pin 消息。Identity: supports `user` and `bot`.
+-  - `delete` — 移除 Pin 消息。Identity: supports `user` and `bot`.
+-  - `list` — 获取群内 Pin 消息。Identity: supports `user` and `bot`.
+-
+-### feed.groups
+-
+-  - `batch_add_item` — Batch add feed cards to a feed group. Identity: `user` only (`user_access_token`).[Must-read](references/lark-im-feed-groups.md)
+-  - `batch_query` — Batch query feed groups. Identity: `user` only (`user_access_token`).[Must-read](references/lark-im-feed-groups.md)
+-  - `batch_remove_item` — Batch remove feed cards from a feed group. Identity: `user` only (`user_access_token`).[Must-read](references/lark-im-feed-groups.md)
+-  - `create` — Create a feed group. Identity: `user` only (`user_access_token`).[Must-read](references/lark-im-feed-groups.md)
+-  - `delete` — Delete a feed group. Identity: `user` only (`user_access_token`).[Must-read](references/lark-im-feed-groups.md)
+-  - `update` — Update a feed group. Identity: `user` only (`user_access_token`).[Must-read](references/lark-im-feed-groups.md)
+-
+-## 权限表
+-
+-| 方法 | 所需 scope |
+-|------|-----------|
+-| `chats.create` | `im:chat:create` |
+-| `chats.get` | `im:chat:read` |
+-| `chats.link` | `im:chat:read` |
+-| `chats.update` | `im:chat:update` |
+-| `chat.members.bots` | `im:chat.members:read` |
+-| `chat.members.create` | `im:chat.members:write_only` |
+-| `chat.members.delete` | `im:chat.members:write_only` |
+-| `chat.members.get` | `im:chat.members:read` |
+-| `chat.user_setting.batch_query` | `im:chat.user_setting:read` |
+-| `chat.user_setting.batch_update` | `im:chat.user_setting:write` |
+-| `chat.managers.add_managers` | `im:chat.managers:write_only` |
+-| `chat.managers.delete_managers` | `im:chat.managers:write_only` |
+-| `chat.moderation.get` | `im:chat.moderation:read` |
+-| `chat.moderation.update` | `im:chat:moderation:write_only` |
+-| `messages.delete` | `im:message:recall` |
+-| `messages.forward` | `im:message` |
+-| `messages.merge_forward` | `im:message` |
+-| `messages.read_users` | `im:message:readonly` |
+-| `messages.urgent_app` | `im:message.urgent` |
+-| `messages.urgent_phone` | `im:message.urgent:phone` |
+-| `messages.urgent_sms` | `im:message.urgent:sms` |
+-| `reactions.batch_query` | `im:message.reactions:read` |
+-| `reactions.create` | `im:message.reactions:write_only` |
+-| `reactions.delete` | `im:message.reactions:write_only` |
+-| `reactions.list` | `im:message.reactions:read` |
+-| `threads.forward` | `im:message` |
+-| `images.create` | `im:resource` |
+-| `pins.create` | `im:message.pins:write_only` |
+-| `pins.delete` | `im:message.pins:write_only` |
+-| `pins.list` | `im:message.pins:read` |
+-| `feed.groups.batch_add_item` | `im:feed_group_v1:write` |
+-| `feed.groups.batch_query` | `im:feed_group_v1:read` |
+-| `feed.groups.batch_remove_item` | `im:feed_group_v1:write` |
+-| `feed.groups.create` | `im:feed_group_v1:write` |
+-| `feed.groups.delete` | `im:feed_group_v1:write` |
+-| `feed.groups.update` | `im:feed_group_v1:write` |
++> **MUST** run `schema` before any native call: it is the live source for the `--data` / `--params` structure, the supported identity (`--as user` vs `--as bot`), owner/admin/tenant constraints, and the required `im:*` scope — do not guess. On a missing-scope error, lark-cli returns a `console_url`; follow the lark-shared permission-handling flow.
+-- 
+2.50.1 (Apple Git-155)
+

harness-opt/verify_results/round-002-lark-im-references-lark-im-messages-send.patch

@@ -0,0 +1,334 @@
+From 82a099feafb45d101116f10230ce7c2f92fbcfe5 Mon Sep 17 00:00:00 2001
+From: "zhangheng.023" <zhangheng.023@bytedance.com>
+Date: Tue, 23 Jun 2026 19:17:24 +0800
+Subject: [PATCH] =?UTF-8?q?opt(round-002):=20lark-im-messages-send.md=20?=
+ =?UTF-8?q?=E2=80=94=20consolidate=204x=20repeated=20content-flag=20rule,?=
+ =?UTF-8?q?=20compress=20media=20enumeration=20&=20--help-mirror=20section?=
+ =?UTF-8?q?s?=
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+---
+ .../references/lark-im-messages-send.md       | 259 ++++--------------
+ 1 file changed, 51 insertions(+), 208 deletions(-)
+
+diff --git a/skills/lark-im/references/lark-im-messages-send.md b/skills/lark-im/references/lark-im-messages-send.md
+index 484c024f..32818909 100644
+--- a/skills/lark-im/references/lark-im-messages-send.md
++++ b/skills/lark-im/references/lark-im-messages-send.md
+@@ -1,10 +1,8 @@
+ # im +messages-send
+ 
+-> **Prerequisite:** Read [`../lark-shared/SKILL.md`](../../lark-shared/SKILL.md) first to understand authentication, global parameters, and safety rules.
++> **Prerequisite:** Read [`../lark-shared/SKILL.md`](../../lark-shared/SKILL.md) first for authentication, global parameters, and safety rules.
+ 
+-Send a message to a group chat or a direct message conversation. Supports both user identity (`--as user`) and bot identity (`--as bot`).
+-
+-This skill maps to the shortcut: `lark-cli im +messages-send` (internally calls `POST /open-apis/im/v1/messages`).
++Send a message to a group chat (`--chat-id oc_xxx`) or a direct message (`--user-id ou_xxx`). One step, supports `--as user` and `--as bot` (default `bot`). Maps to shortcut `lark-cli im +messages-send` (`POST /open-apis/im/v1/messages`).
+ 
+ ## Safety Constraints
+ 
+@@ -16,249 +14,94 @@ Messages sent by this tool are visible to other people. Before calling it, you *
+ 
+ **Do not** send messages without explicit user approval.
+ 
+-When using `--as bot`, the message is sent in the app's name, so make sure the app has already been added to the target chat.
+-
+-When using `--as user`, the message is sent as the authorized end user and requires the `im:message.send_as_user` and `im:message` scopes.
++- `--as bot` (TAT, scope `im:message:send_as_bot`): the message is sent in the app's name — the app must already be in the target chat or have a DM relationship with the target user.
++- `--as user` (UAT, scopes `im:message.send_as_user` + `im:message`): the message is sent as the authorized end user.
+ 
+ ## Choose The Right Content Flag
+ 
+-### Default Selection Rule For Agents
+-
+-- Prefer `--markdown` for headings, lists, links, summaries, reports, or Markdown-looking content.
+-- Use `--text` for exact plain text: logs, code, indentation-sensitive text, or literal Markdown.
+-- Use `--content` for exact `post` JSON, titles, multiple locales, cards, or unsupported structures.
+-
+-| Need | Recommended flag | Why |
+-|------|------|------|
+-| Send headings, lists, links, summaries, or reports | `--markdown` | Best default for lightweight formatting; converted to Feishu `post` JSON |
+-| Send plain text exactly as written | `--text` | Preserves literal text; no Markdown conversion |
+-| Precisely control the final payload | `--content` | You provide the exact JSON for `text` / `post` / `interactive` / `share_*` / media payloads |
+-| Send image / file / video / audio | `--image` / `--file` / `--video` / `--audio` | Shortcut uploads URLs, or cwd-relative local files automatically |
+-
+-### `--text` vs `--markdown`
+-
+-- Use `--markdown` for lightweight formatted messages.
+-- Use `--text` for exact plain text, especially logs, code, indentation, or Markdown characters that should **not** render.
+-- Use `--content` when `--markdown` is not enough, especially if you need exact `post` JSON, a title, multiple locales, cards, or unsupported rich structures.
+-
+-## What `--markdown` Really Does
+-
+-`--markdown` accepts Markdown-like input and converts it to the Feishu `post` payload required by the message API.
+-
+-The shortcut does all of the following before sending:
+-
+-1. Forces `msg_type=post`
+-2. Resolves remote Markdown images like `![x](https://...)` by downloading and uploading them first
+-3. Normalizes the Markdown for Feishu post rendering
+-4. Wraps the result as:
+-
+-```json
+-{"zh_cn":{"content":[[{"tag":"md","text":"..."}]]}}
+-```
+-
+-This makes `--markdown` the simplest path for lightweight formatted messages.
+-
+-### Markdown Boundaries
+-
+-- It does **not** promise full CommonMark / GitHub Flavored Markdown support.
+-- It always becomes a `post` payload with a single `zh_cn` locale.
+-- It does **not** let you set a `post` title. If you need a title, use `--msg-type post --content ...`.
+-- Headings are rewritten:
+-    - `# Title` becomes `#### Title`
+-    - `##` to `######` are normalized to `#####` when the content contains H1-H3
+-- Consecutive headings are separated with blank lines after heading normalization.
+-- Block spacing and line breaks may be normalized during conversion.
+-- Code blocks are preserved as code blocks.
+-- Excess blank lines are compressed.
+-- Already-uploaded `img_xxx` image keys are the most reliable Markdown image input.
+-- Local paths in Markdown image syntax like `![x](./a.png)` are **not** supported and will not be auto-uploaded.
+-- Remote URLs (`https://...`) will be auto-downloaded and uploaded at runtime; if the download or upload fails, the image is removed with a warning.
+-
+-If you need a title, multiple locales, cards, unsupported rich structures, or byte-for-byte post JSON control, use `--content` and provide the final JSON yourself.
++| Content | Flag | Why |
++|---|---|---|
++| Headings, lists, links, summaries, reports (lightweight formatting) | `--markdown` | Best default; converted to Feishu `post` JSON |
++| Exact plain text — logs, code, indentation, literal Markdown chars that must **not** render | `--text` | Preserves literal text; no conversion |
++| Exact `post` JSON, a `post` title, multiple locales, cards (`interactive`), `share_*`, or unsupported structures | `--content` | You provide the final JSON; it must match the effective `--msg-type` |
++| Image / file / video / audio | `--image` / `--file` / `--video` / `--audio` | Uploads URLs or cwd-relative local files automatically |
+ 
+-### Image Constraint for `--markdown`
++These content flags (and the media flags) are **mutually exclusive** — pass exactly one. Media flags are also mutually exclusive with each other.
+ 
+-When using `--markdown` with images, prefer pre-uploading via `images.create` and referencing `![alt](img_xxx)` for predictable results. Remote URLs may work but are not guaranteed.
++## `--markdown` Gotchas
+ 
+-**Steps:**
++`--markdown` always forces `msg_type=post` (single `zh_cn` locale) and normalizes input for Feishu post rendering. Key boundaries (not full CommonMark/GFM):
+ 
+-```bash
+-# 1. Upload image to get image_key
+-lark-cli im images create --data '{"image_type":"message"}' --file ./diagram.png
+-# Returns: {"image_key":"img_v3_xxxx"}
+-
+-# 2. Use image_key in --markdown
+-lark-cli im +messages-send --chat-id oc_xxx --markdown $'## Report\n\n![diagram](img_v3_xxxx)\n\nSee above for details.'
+-```
++- **No `post` title** — if you need one, use `--content` with `post` JSON.
++- **Headings rewritten**: `# Title` → `#### Title`; `##`–`######` normalized to `#####` when content has H1–H3. Code blocks preserved; excess blank lines compressed.
++- **Images**: pre-upload via `im images create` and reference `![alt](img_xxx)` for reliable results. Remote `https://` URLs are auto-downloaded+uploaded at runtime (removed with a warning if that fails). Local paths in `![x](./a.png)` are **not** supported and will not auto-upload.
+ 
+-## Preserving Formatting
++## Preserving Exact Formatting
+ 
+-If the message has multiple lines, indentation, code blocks, tabs, or many quotes/backslashes, prefer shell ANSI-C quoting with `$'...'` for either `--markdown` or `--text`.
+-
+-This is especially useful in `zsh` / `bash` because it lets you write `\n` explicitly instead of relying on the shell to preserve literal newlines.
+-
+-### When formatting must be preserved
+-
+-Use `--text` plus `$'...'`:
+-
+-```bash
+-lark-cli im +messages-send --chat-id oc_xxx --text $'Build failed\nBranch: feature/im-docs\nAction: please check logs'
+-```
++For multi-line text, indentation, code blocks, tabs, or many backslashes/quotes, use shell ANSI-C quoting `$'...'` so `\n` is written explicitly. Use `--text` + `$'...'` when the receiver must see the text exactly as entered:
+ 
+ ```bash
+-lark-cli im +messages-send --chat-id oc_xxx --text $'```bash\nmake test\nmake lint\n```'
++lark-cli im +messages-send --chat-id oc_xxx --text $'Build failed\nBranch: feature/x\nAction: check logs'
+ ```
+ 
+-Use this path when you want the receiver to see the text exactly as entered, not a converted Markdown post.
+-
+ ## Commands
+ 
+ ```bash
+-# Send a formatted update
++# Formatted update (Markdown → post)
+ lark-cli im +messages-send --chat-id oc_xxx --markdown $'## Update\n\n- item 1\n- item 2'
+ 
+-# Send a plain one-line message
++# Plain one-line text
+ lark-cli im +messages-send --chat-id oc_xxx --text "Hello"
+ 
+-# Equivalent manual JSON
+-lark-cli im +messages-send --chat-id oc_xxx --content '{"text":"Hello"}'
+-
+-# Send to a direct message (pass open_id)
++# Direct message (pass open_id)
+ lark-cli im +messages-send --user-id ou_xxx --text "Hello"
+ 
+-# Send multi-line text while preserving formatting
+-lark-cli im +messages-send --chat-id oc_xxx --text $'Line 1\nLine 2\n  indented line'
+-
+-# Send Markdown with an image (must pre-upload via images.create)
+-lark-cli im images create --data '{"image_type":"message"}' --file ./screenshot.png
+-# Use the returned image_key in the markdown content
+-lark-cli im +messages-send --chat-id oc_xxx --markdown $'## Status\n\n![screenshot](img_v3_xxxx)\n\nDone.'
+-
+-# If you need exact post structure, send JSON directly
++# Exact post structure with a title
+ lark-cli im +messages-send --chat-id oc_xxx --msg-type post --content '{"zh_cn":{"title":"Title","content":[[{"tag":"text","text":"Body"}]]}}'
+ 
+-# Send a local image (uploaded automatically before sending)
+-lark-cli im +messages-send --chat-id oc_xxx --image ./photo.png
+-
+-# Or send directly with an existing image_key
+-lark-cli im +messages-send --chat-id oc_xxx --image img_xxx
++# Markdown with an image (pre-upload first)
++lark-cli im images create --data '{"image_type":"message"}' --file ./diagram.png   # -> {"image_key":"img_v3_xxxx"}
++lark-cli im +messages-send --chat-id oc_xxx --markdown $'## Report\n\n![diagram](img_v3_xxxx)\n\nDone.'
+ 
+-# Send a local file (uploaded automatically before sending)
++# Media (local files uploaded automatically; --video requires --video-cover)
++lark-cli im +messages-send --chat-id oc_xxx --image ./photo.png
+ lark-cli im +messages-send --chat-id oc_xxx --file ./report.pdf
+-
+-# Send a video (--video-cover is required as the cover)
+ lark-cli im +messages-send --chat-id oc_xxx --video ./demo.mp4 --video-cover ./cover.png
+-lark-cli im +messages-send --chat-id oc_xxx --video ./demo.mp4 --video-cover img_xxx
+-
+-# Send audio
+ lark-cli im +messages-send --chat-id oc_xxx --audio ./voice.opus
+ 
+-# Use an idempotency key (same key sends only once within 1 hour)
+-lark-cli im +messages-send --chat-id oc_xxx --text "Hello" --idempotency-key my-unique-id
+-
+-# Preview the request without executing it
+-lark-cli im +messages-send --chat-id oc_xxx --markdown $'## Test\n\nhello' --dry-run
++# Idempotency (same key sends only once within 1 hour) / preview without sending
++lark-cli im +messages-send --chat-id oc_xxx --text "Hi" --idempotency-key my-id
++lark-cli im +messages-send --chat-id oc_xxx --markdown $'## Test\n\nhi' --dry-run
+ ```
+ 
+-## Media Input Rules
+-
+-- Media flags accept an existing key (`img_xxx` / `file_xxx`), an `http://` or `https://` URL, or a local file path.
+-- Local paths must be relative to the current working directory and stay within it after resolving `..` and symlinks.
+-- Absolute paths such as `/tmp/photo.png` are rejected. Run the command from the file's directory and pass `./photo.png`, or copy the file into the current directory first.
+-
+-## Parameters
+-
+-| Parameter | Required | Description                                                                                                                                                                                   |
+-|------|------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+-| `--chat-id <id>` | One of two | Group chat ID (`oc_xxx`)                                                                                                                                                                      |
+-| `--user-id <id>` | One of two | User open_id (`ou_xxx`) for direct messages                                                                                                                                                   |
+-| `--text <string>` | One content option | Plain text message. Use when exact text and formatting preservation matter. Automatically wrapped as `{"text":"..."}`                                                                         |
+-| `--markdown <string>` | One content option | Best default for lightweight formatted messages such as headings, lists, links, summaries, and reports. Internally converted to `post` JSON with Feishu-specific normalization               |
+-| `--content <json>` | One content option | Exact message content JSON string; use this when you need full control over `msg_type` and payload. The JSON must match the effective `--msg-type`                                            |
+-| `--image <path\|url\|key>` | One content option | Cwd-relative local image path, URL, or `image_key` (`img_xxx`). Local paths and URLs are uploaded automatically                                                                               |
+-| `--file <path\|url\|key>` | One content option | Cwd-relative local file path, URL, or `file_key` (`file_xxx`). Local paths and URLs are uploaded automatically                                                                                |
+-| `--video <path\|url\|key>` | One content option | Cwd-relative local video path, URL, or `file_key` (`file_xxx`). Local paths and URLs are uploaded automatically. **Must be paired with `--video-cover`**                                      |
+-| `--video-cover <path\|url\|key>` | **Required with `--video`** | Cwd-relative local cover image path, URL, or `image_key` (`img_xxx`). Local paths and URLs are uploaded automatically                                                                         |
+-| `--audio <path\|url\|key>` | One content option | Cwd-relative local audio path, URL, or `file_key` (`file_xxx`). Local paths and URLs are uploaded automatically                                                                                           |
+-| `--msg-type <type>` | No | Message type (default `text`). If you use `--text` / `--markdown` / media flags, the effective type is inferred automatically. Explicitly setting a conflicting `--msg-type` fails validation |
+-| `--idempotency-key <key>` | No | Idempotency key; the same key sends only one message within 1 hour                                                                                                                            |
+-| `--as <identity>` | No | Identity type: `bot` or `user` (default `bot`)                                                                                                                                                |
+-| `--dry-run` | No | Print the request only, do not execute it                                                                                                                                                     |
+-
+-> **Mutual exclusivity rule:** `--text`, `--markdown`, `--content`, and `--image`/`--file`/`--video`/`--audio` cannot be used together. Media flags are also mutually exclusive with each other.
+->
+-> **Video cover rule:** `--video` **must** be accompanied by `--video-cover`. Omitting `--video-cover` when using `--video` will fail validation. `--video-cover` cannot be used without `--video`.
+-
+-## Common Mistakes
+-
+-- Choosing `--text` for headings, lists, links, summaries, or reports. Use `--markdown`.
+-- Choosing `--markdown` when you actually need exact plain text. If exact line breaks, spacing, logs, code, or literal Markdown characters matter, use `--text`, usually with `$'...'`.
+-- Assuming `--markdown` supports every Markdown feature. It is converted into a Feishu `post` payload and normalized first.
+-- Putting local image paths inside Markdown like `![x](./a.png)`. `--markdown` does not auto-upload those paths.
+-- **Using local file paths inside Markdown image syntax** (e.g. `![x](./a.png)`) with `--markdown`. Local paths are not auto-uploaded and will not render as an image. Pre-upload via `images.create` to get an `image_key` instead.
+-- Using `--content` without making the JSON match the effective `--msg-type`.
+-- Explicitly setting `--msg-type` to something that conflicts with `--text`, `--markdown`, or media flags.
+-- Mixing `--text`, `--markdown`, or `--content` with media flags in one command.
+-
+-## `content` Format Reference
++Run `lark-cli im +messages-send --help` for the full flag list and types. Load-bearing rules that `--help` may not make obvious:
++
++- **Media paths** accept an existing key (`img_xxx`/`file_xxx`), an `http(s)://` URL, or a **cwd-relative** local path. Absolute paths (e.g. `/tmp/x.png`) are rejected — run from the file's directory and pass `./x.png`. Upload and send use the same identity.
++- **`--video` must be paired with `--video-cover`** (image key/URL/local path); `--video-cover` cannot be used alone.
++- **`--msg-type`** is inferred from `--text`/`--markdown`/media flags; explicitly setting a conflicting type fails validation.
++
++## `content` Format Reference (for `--content`)
+ 
+ | `msg_type` | Example `content` |
+-|----------|-------------|
++|---|---|
+ | `text` | `{"text":"Hello <at user_id=\"ou_xxx\">name</at>"}` |
+ | `post` | `{"zh_cn":{"title":"Title","content":[[{"tag":"text","text":"Body"}]]}}` |
+-| `image` | `{"image_key":"img_xxx"}` |
+-| `file` | `{"file_key":"file_xxx"}` |
+-| `audio` | `{"file_key":"file_xxx"}` |
+-| `media` | `{"file_key":"file_xxx","image_key":"img_xxx"}` (video; `image_key` is the cover from `--video-cover` — **required**) |
+-| `share_chat` | `{"chat_id":"oc_xxx"}` |
+-| `share_user` | `{"user_id":"ou_xxx"}` |
+-| `interactive` | Card JSON (see Feishu interactive card documentation) |
++| `image` / `file` / `audio` | `{"image_key":"img_xxx"}` / `{"file_key":"file_xxx"}` / `{"file_key":"file_xxx"}` |
++| `media` (video) | `{"file_key":"file_xxx","image_key":"img_xxx"}` (`image_key` is the **required** cover) |
++| `share_chat` / `share_user` | `{"chat_id":"oc_xxx"}` / `{"user_id":"ou_xxx"}` |
++| `interactive` (card) | Card JSON (see Feishu interactive card docs) |
+ 
+-## Return Value
+-
+-```json
+-{
+-  "message_id": "om_xxx",
+-  "chat_id": "oc_xxx",
+-  "create_time": "1234567890"
+-}
+-```
++When using `--content`, you are responsible for making the JSON match the effective `msg_type`.
+ 
+ ## @Mention Format
+ 
+-The `<at>` syntax differs by message type. The shortcut only normalizes mentions for `text` and `post`; `interactive` card content is passed through verbatim, so cards must use the card-native syntax below.
+-
+-### `text`
+-
+-- `<at user_id="ou_xxx">name</at>` — the inner text is the mentioned user's display name and is optional (`<at user_id="ou_xxx"></at>` also works)
+-- @all: `<at user_id="all"></at>`
+-
+-### `post`
++The `<at>` syntax differs by message type; the shortcut normalizes mentions for `text` and `post` only — `interactive` cards are passed through verbatim.
+ 
+-- Inside a `text` or `md` element, the same inline form as `text` works: `<at user_id="ou_xxx">name</at>`
+-- Or use a dedicated `at` element node: `{"tag":"at","user_id":"ou_xxx"}` (use `"all"` to mention everyone)
++- **`text`** / inside a `post` `text`/`md` element: `<at user_id="ou_xxx">name</at>` (inner name optional); @all: `<at user_id="all"></at>`. In `post` you may also use a node: `{"tag":"at","user_id":"ou_xxx"}` (`"all"` for everyone).
++- **`interactive` (card)** — card-native syntax inside a `lark_md`/`markdown` element: `<at id=ou_xxx></at>`, multiple `<at ids=ou_1,ou_2></at>`, by email `<at email=user@example.com></at>`.
+ 
+-### `interactive` (card)
+-
+-Card content is **not** normalized — use the card-native `<at>` syntax inside a `lark_md` / `markdown` element:
+-
+-- single user by open_id: `<at id=ou_xxx></at>`
+-- multiple users: `<at ids=ou_xxx1,ou_xxx2></at>`
+-- by email: `<at email=user@example.com></at>`
+-
+-## Notes
++## Return Value
+ 
+-- `--chat-id` and `--user-id` are mutually exclusive; you must provide exactly one
+-- `--content` must be valid JSON
+-- When using `--content`, you are responsible for making the JSON structure match the effective `msg_type`
+-- `--image`/`--file`/`--video`/`--audio` support existing keys, URLs, and cwd-relative local file paths; the shortcut uploads local paths and URLs first, then sends the message; both the upload and send steps use the same identity (UAT when `--as user`, TAT when `--as bot`)
+-- If the provided media value starts with `img_` or `file_`, it is treated as an existing key and used directly
+-- `--markdown` always sends `msg_type=post`, even if you do not explicitly set `--msg-type post`
+-- If you explicitly set `--msg-type` and it conflicts with the chosen content flag, validation fails
+-- When using `--video`, `--video-cover` is required as the video cover
+-- `--dry-run` uses placeholder image keys for remote Markdown images and placeholder media keys for local uploads
+-- Failures return an error code and message
+-- `--as user` uses a user access token (UAT) and requires the `im:message.send_as_user` and `im:message` scopes; the message is sent as the authorized end user
+-- `--as bot` uses a tenant access token (TAT) and requires the `im:message:send_as_bot` scope
+-- When sending as a bot, the app must already be in the target group or already have a direct-message relationship with the target user
+-- When using `--markdown` with images, pre-uploading via `images.create` to obtain an `image_key` is recommended for reliability; remote URLs may be auto-resolved at runtime, but if download/upload fails the image is removed with a warning; local paths are not supported
++```json
++{"message_id": "om_xxx", "chat_id": "oc_xxx", "create_time": "1234567890"}
++```
+-- 
+2.50.1 (Apple Git-155)
+

harness-opt/verify_results/round-003-lark-im-references-lark-im-chat-create.patch

@@ -0,0 +1,135 @@
+From cbd6e56ac07285fd973c53ff7382da0112b6cf5d Mon Sep 17 00:00:00 2001
+From: "zhangheng.023" <zhangheng.023@bytedance.com>
+Date: Tue, 23 Jun 2026 19:51:49 +0800
+Subject: [PATCH] =?UTF-8?q?opt(round-003):=20references/lark-im-chat-creat?=
+ =?UTF-8?q?e.md=20=E2=80=94=20dedup=20Commands/Scenarios=20overlap=20+=20c?=
+ =?UTF-8?q?ompress=20--help-mirroring=20Common=20Errors=20into=20pointers,?=
+ =?UTF-8?q?=20keep=20232043=20two-step=20flow=20&=20all=20guardrails?=
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+---
+ .../lark-im/references/lark-im-chat-create.md | 78 +++++--------------
+ 1 file changed, 18 insertions(+), 60 deletions(-)
+
+diff --git a/skills/lark-im/references/lark-im-chat-create.md b/skills/lark-im/references/lark-im-chat-create.md
+index 76716f76..7d65e5d3 100644
+--- a/skills/lark-im/references/lark-im-chat-create.md
++++ b/skills/lark-im/references/lark-im-chat-create.md
+@@ -12,43 +12,24 @@ This skill maps to the shortcut: `lark-cli im +chat-create` (internally calls `P
+ ## Commands
+ 
+ ```bash
+-# Create a private group (default)
++# Private group (default)
+ lark-cli im +chat-create --name "My Group"
+ 
+-# Create a public group (name is required and must be at least 2 characters)
++# Public group (--name required, min 2 chars)
+ lark-cli im +chat-create --name "Public Group" --type public
+ 
+-# Create a topic chat
++# Topic chat (a 话题群; see note under Parameters)
+ lark-cli im +chat-create --name "Topic Group" --chat-mode topic
+ 
+-# Specify the group owner
+-lark-cli im +chat-create --name "My Group" --owner ou_xxx
++# Invite members and set owner (users: up to 50 ou_xxx; bots: up to 5 cli_xxx)
++lark-cli im +chat-create --name "My Group" --owner ou_xxx --users "ou_aaa,ou_bbb" --bots "cli_aaa"
+ 
+-# Invite user members (comma-separated open_ids, up to 50)
+-lark-cli im +chat-create --name "My Group" --users "ou_aaa,ou_bbb"
+-
+-# Invite bot members (comma-separated app IDs, up to 5)
+-lark-cli im +chat-create --name "My Group" --bots "cli_aaa,cli_bbb"
+-
+-# Invite both users and bots
+-lark-cli im +chat-create --name "My Group" --users "ou_aaa" --bots "cli_aaa"
+-
+-# Make the creating bot a group manager (bot identity only)
+-lark-cli im +chat-create --name "My Group" --set-bot-manager --as bot
+-
+-# JSON output
+-lark-cli im +chat-create --name "My Group" --format json
+-
+-# Create a group with bot identity
+-lark-cli im +chat-create --name "My Group" --users "ou_aaa" --as bot
+-
+-# Create a group with user identity
+-lark-cli im +chat-create --name "My Group" --users "ou_aaa,ou_bbb" --as user
+-
+-# Preview the request without creating anything
+-lark-cli im +chat-create --name "My Group" --dry-run
++# Bot identity, making the creating bot a manager
++lark-cli im +chat-create --name "My Group" --users "ou_aaa" --as bot --set-bot-manager
+ ```
+ 
++Run `lark-cli im +chat-create --help` for the full flag list, limits, and types. Single-flag variations (`--as user`, `--description`, `--format json`, `--dry-run` preview, etc.) follow the Parameters table below — `--dry-run` previews the request without creating anything.
++
+ ## Parameters
+ 
+ | Parameter | Required | Limits | Description |
+@@ -106,6 +87,13 @@ lark-cli im +chat-create --name "<group name>" --users "ou_aaa,ou_bbb" --as user
+ 
+ The authorized user is automatically the group creator and member.
+ 
++### Create a group, then send a welcome message
++
++```bash
++CHAT_ID=$(lark-cli im +chat-create --name "New Group" --format json | jq -r '.data.chat_id')
++lark-cli im +messages-send --chat-id "$CHAT_ID" --text "Welcome, everyone!"
++```
++
+ ## Output Fields
+ 
+ | Field | Description |
+@@ -117,43 +105,13 @@ The authorized user is automatically the group creator and member.
+ | `external` | Whether the group is external |
+ | `share_link` | Group share link (omitted if retrieval fails) |
+ 
+-## Usage Scenarios
+-
+-### Scenario 1: Create a group and specify the owner
+-
+-```bash
+-lark-cli im +chat-create --name "Project Discussion Group" --owner ou_xxx
+-```
+-
+-### Scenario 2: Create a group and invite users and a bot
+-
+-```bash
+-lark-cli im +chat-create --name "Project Discussion Group" \
+-  --owner ou_xxx \
+-  --users "ou_aaa,ou_bbb" \
+-  --bots "cli_aaa"
+-```
+-
+-### Scenario 3: Create a group and send a welcome message
+-
+-```bash
+-CHAT_ID=$(lark-cli im +chat-create --name "New Group" --format json | jq -r '.data.chat_id')
+-lark-cli im +messages-send --chat-id "$CHAT_ID" --text "Welcome, everyone!"
+-```
+-
+ ## Common Errors and Troubleshooting
+ 
++Format/limit validation (`--name`/`--description`/`--users`/`--bots`/`--owner` length, count, and `ou_xxx`/`cli_xxx` format) is enforced by the CLI and reported verbatim with the fix — see the Parameters table for limits. The two errors needing extra action:
++
+ | Symptom | Root Cause | Solution |
+ |---------|---------|---------|
+ | Permission denied (99991672) | The app does not have `im:chat:create` (bot) or `im:chat:create_by_user` (user) permission enabled | Enable the required permission for the app in the Open Platform console |
+-| `--name is required for public groups and must be at least 2 characters` | A public group was created without a name or with a name shorter than 2 characters | Provide a name with at least 2 characters |
+-| `--name exceeds the maximum of 60 characters` | The group name is too long | Shorten the name to 60 characters or fewer |
+-| `--description exceeds the maximum of 100 characters` | The group description is too long | Shorten the description to 100 characters or fewer |
+-| `--users exceeds the maximum of 50` | Too many user members were provided | Split the operation into batches and add more members later |
+-| `--bots exceeds the maximum of 5` | Too many bot members were provided | Invite at most 5 bots at once |
+-| `invalid user id: expected open_id (ou_xxx)` | Invalid user ID format | Use the `ou_xxx` format for users |
+-| `invalid bot id: expected app ID (cli_xxx)` | Invalid bot ID format | Use the `cli_xxx` format for bots |
+-| `invalid --owner: expected open_id (ou_xxx)` | Invalid owner ID format | Use the `ou_xxx` format for the owner |
+ | `bot is invisible to user` (232043) | The bot and target users are mutually invisible | Follow the two-step flow in AI Usage Guidance above — do not pass other users in `--users` during creation |
+ 
+ ## References
+-- 
+2.50.1 (Apple Git-155)
+

internal/binding/audit.go

@@ -131,3 +131,31 @@ func requireInTrustedDirs(effectivePath string, trustedDirs []string, label stri
 	}
 	return fmt.Errorf("%s: path %q is not inside any trusted directory", label, effectivePath)
 }
+
+// auditFilePermissions rejects world/group-writable modes (always) and
+// world/group-readable modes (unless allowReadableByOthers is true, which
+// exec commands typically need for their usual 755 mode).
+func auditFilePermissions(effectivePath string, allowReadableByOthers bool, label string) error {
+	info, err := vfs.Stat(effectivePath)
+	if err != nil {
+		return fmt.Errorf("%s: cannot stat %q: %w", label, effectivePath, err)
+	}
+	mode := info.Mode().Perm()
+
+	if mode&0o002 != 0 {
+		return fmt.Errorf("%s: path %q is world-writable (mode %04o)", label, effectivePath, mode)
+	}
+	if mode&0o020 != 0 {
+		return fmt.Errorf("%s: path %q is group-writable (mode %04o)", label, effectivePath, mode)
+	}
+	if allowReadableByOthers {
+		return nil
+	}
+	if mode&0o004 != 0 {
+		return fmt.Errorf("%s: path %q is world-readable (mode %04o)", label, effectivePath, mode)
+	}
+	if mode&0o040 != 0 {
+		return fmt.Errorf("%s: path %q is group-readable (mode %04o)", label, effectivePath, mode)
+	}
+	return nil
+}

internal/binding/audit_unix.go

@@ -29,31 +29,3 @@ func checkOwnerUID(path, label string) error {
 	}
 	return nil
 }
-
-// auditFilePermissions rejects world/group-writable modes (always) and
-// world/group-readable modes (unless allowReadableByOthers is true, which
-// exec commands typically need for their usual 755 mode).
-func auditFilePermissions(effectivePath string, allowReadableByOthers bool, label string) error {
-	info, err := vfs.Stat(effectivePath)
-	if err != nil {
-		return fmt.Errorf("%s: cannot stat %q: %w", label, effectivePath, err)
-	}
-	mode := info.Mode().Perm()
-
-	if mode&0o002 != 0 {
-		return fmt.Errorf("%s: path %q is world-writable (mode %04o)", label, effectivePath, mode)
-	}
-	if mode&0o020 != 0 {
-		return fmt.Errorf("%s: path %q is group-writable (mode %04o)", label, effectivePath, mode)
-	}
-	if allowReadableByOthers {
-		return nil
-	}
-	if mode&0o004 != 0 {
-		return fmt.Errorf("%s: path %q is world-readable (mode %04o)", label, effectivePath, mode)
-	}
-	if mode&0o040 != 0 {
-		return fmt.Errorf("%s: path %q is group-readable (mode %04o)", label, effectivePath, mode)
-	}
-	return nil
-}

internal/binding/audit_windows.go

@@ -5,22 +5,7 @@
 
 package binding
 
-import (
-	"fmt"
-
-	"github.com/larksuite/cli/internal/vfs"
-)
-
 // checkOwnerUID is a no-op on Windows where Unix UID semantics don't apply.
 func checkOwnerUID(path, label string) error {
 	return nil
 }
-
-// auditFilePermissions skips POSIX permission-bit auditing on Windows because
-// Go synthesizes mode bits from file attributes rather than NTFS ACLs.
-func auditFilePermissions(effectivePath string, allowReadableByOthers bool, label string) error {
-	if _, err := vfs.Stat(effectivePath); err != nil {
-		return fmt.Errorf("%s: cannot stat %q: %w", label, effectivePath, err)
-	}
-	return nil
-}

internal/binding/audit_windows_test.go

@@ -1,33 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-//go:build windows
-
-package binding
-
-import (
-	"os"
-	"path/filepath"
-	"testing"
-)
-
-func TestAssertSecurePath_WindowsIgnoresSyntheticUnixPermissionBits(t *testing.T) {
-	dir := t.TempDir()
-	p := filepath.Join(dir, "secrets-getter.cmd")
-	if err := os.WriteFile(p, []byte("@echo off\r\n"), 0o600); err != nil {
-		t.Fatalf("write temp command: %v", err)
-	}
-
-	got, err := AssertSecurePath(AuditParams{
-		TargetPath:            p,
-		Label:                 "exec provider command",
-		AllowInsecurePath:     false,
-		AllowReadableByOthers: true,
-	})
-	if err != nil {
-		t.Fatalf("unexpected error for Windows synthetic mode bits: %v", err)
-	}
-	if got != p {
-		t.Errorf("got %q, want %q", got, p)
-	}
-}

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@larksuite/cli",
-  "version": "1.0.57",
+  "version": "1.0.56",
   "description": "The official CLI for Lark/Feishu open platform",
   "bin": {
     "lark-cli": "scripts/run.js"

scripts/semantic-review-workflow.test.sh

@@ -179,10 +179,7 @@ fi
 require_in_step "$summary_verify_step" 'workflowPath !== ".github/workflows/ci.yml"' "PR quality summary must verify the triggering workflow path"
 require_in_step "$summary_verify_step" 'run.event !== "pull_request"' "PR quality summary must only handle pull_request workflow_run events"
 require_in_step "$summary_verify_step" 'run.repository.id !== context.payload.repository.id' "PR quality summary must verify workflow_run repository id"
-require_in_step "$summary_verify_step" 'const targetHeadSha = run.head_sha' "PR quality summary must use the CI run head SHA as the verified PR head"
-require_in_step "$summary_verify_step" 'eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()' "PR quality summary should tolerate mutable workflow_run PR head metadata"
 require_in_step "$summary_verify_step" 'factsArtifactPattern' "PR quality summary should use the base-bound facts artifact name when available"
-require_in_step "$summary_verify_step" 'const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha' "PR quality summary must prefer the CI-time artifact base SHA"
 require_in_step "$summary_verify_step" 'core.setOutput("artifact_error"' "PR quality summary must expose artifact binding failures"
 require_in_step "$summary_artifact_step" 'factsArtifactName' "PR quality summary artifact step must use the verified facts artifact binding"
 require_in_step "$summary_extract_facts_step" 'SEMANTIC_REVIEW_DECISION_OUT' "PR quality summary artifact verifier must write an infrastructure decision on verifier failure"
@@ -201,9 +198,8 @@ require_in_step "$verify_step" 'workflowPath !== ".github/workflows/ci.yml"' "se
 require_in_step "$verify_step" 'run.repository.id !== context.payload.repository.id' "semantic-review must verify workflow_run repository id"
 require_in_step "$verify_step" 'run.event !== "pull_request"' "semantic-review must only handle pull_request workflow_run events"
 require_in_step "$verify_step" 'run.conclusion !== "success"' "semantic-review must only consume successful CI runs"
-require_in_step "$verify_step" 'const eventHeadSha = runPRs[0]?.head?.sha || ""' "semantic-review should inspect workflow_run PR head metadata"
-require_in_step "$verify_step" 'const targetHeadSha = run.head_sha' "semantic-review target PR head must come from the completed CI run"
-require_in_step "$verify_step" 'eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()' "semantic-review should tolerate mutable workflow_run PR head metadata"
+require_in_step "$verify_step" 'const eventHeadSha = runPRs[0]?.head?.sha || ""' "semantic-review must prefer workflow_run PR head when GitHub provides it"
+require_in_step "$verify_step" 'const targetHeadSha = eventHeadSha || run.head_sha' "semantic-review target PR head must come from the workflow_run event"
 require_in_step "$verify_step" 'factsArtifactPattern' "semantic-review must use a base-bound facts artifact name"
 require_in_step "$verify_step" 'listWorkflowRunArtifacts' "semantic-review must read the workflow_run artifacts before resolving fallback base SHA"
 require_in_step "$verify_step" 'artifactHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()' "semantic-review must not let the artifact choose a different PR head"
@@ -214,8 +210,8 @@ require_in_step "$verify_step" 'commit_sha: targetHeadSha' "semantic-review fall
 require_in_step "$verify_step" 'github.rest.pulls.list' "semantic-review must have a pull-list fallback when commit association is empty"
 require_in_step "$verify_step" 'candidatePRs.length > 1' "semantic-review must fail closed when commit-to-PR fallback is ambiguous"
 require_in_step "$verify_step" 'pr.head.sha !== targetHeadSha' "semantic-review must skip stale PR heads"
-require_in_step "$verify_step" 'eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()' "semantic-review should tolerate mutable workflow_run PR base metadata"
-require_in_step "$verify_step" 'const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha' "semantic-review must prefer the CI-time artifact base SHA"
+require_in_step "$verify_step" 'eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()' "semantic-review must reject mismatched event and artifact base SHAs"
+require_in_step "$verify_step" 'const baseSha = eventBaseSha || artifactBaseSha' "semantic-review fallback must use the CI-time artifact base SHA"
 require_in_step "$verify_step" 'pr.base.sha !== baseSha' "semantic-review must skip stale PR bases"
 require_in_step "$verify_step" 'core.setOutput("run_id"' "semantic-review must pass verified workflow run id to publisher"
 require_in_step "$verify_step" 'core.setOutput("head_repo_id"' "semantic-review must pass verified head repo id"

shortcuts/apps/apps_html_publish.go

@@ -7,7 +7,6 @@ import (
 	"context"
 	"fmt"
 	"io"
-	"path/filepath"
 	"strings"
 
 	"github.com/larksuite/cli/extension/fileio"
@@ -85,9 +84,6 @@ var AppsHTMLPublish = common.Shortcut{
 			// for dry-run "advisory preview" semantics).
 			dry.Set("validation_error", err.Error())
 		}
-		if hits := oversizeHTMLFiles(candidates); len(hits) > 0 {
-			dry.Set("oversize_html", hits)
-		}
 		dry.Set("file_count", len(candidates))
 		var totalSize int64
 		names := make([]string, 0, len(candidates))
@@ -144,22 +140,18 @@ type appsHTMLPublishSpec struct {
 // per-environment .env.* files for every stage).
 const maxSensitiveListInError = 5
 
-// truncatedJoin joins items with ", ", capping at max entries and appending
-// "(and N more)" for the remainder, so an inline error list stays readable when
-// a payload has many hits.
-func truncatedJoin(items []string, max int) string {
-	if len(items) <= max {
-		return strings.Join(items, ", ")
-	}
-	return strings.Join(items[:max], ", ") + fmt.Sprintf(" (and %d more)", len(items)-max)
-}
-
 // sensitiveCandidatesError builds the Validate-time rejection when --path
 // contains credential files and --allow-sensitive was not set.
 func sensitiveCandidatesError(hits []string) error {
+	var sample string
+	if len(hits) <= maxSensitiveListInError {
+		sample = strings.Join(hits, ", ")
+	} else {
+		sample = strings.Join(hits[:maxSensitiveListInError], ", ") +
+			fmt.Sprintf(" (and %d more)", len(hits)-maxSensitiveListInError)
+	}
 	return appsValidationParamError("--path",
-		"--path contains %d credential file(s) that should not be published: %s",
-		len(hits), truncatedJoin(hits, maxSensitiveListInError)).
+		"--path contains %d credential file(s) that should not be published: %s", len(hits), sample).
 		WithHint("remove these files from the publish payload, OR pass --allow-sensitive if shipping them is intentional (e.g. a docs site demoing credential-file formats)")
 }
 
@@ -176,30 +168,6 @@ var maxHTMLPublishTarballBytes int64 = 20 * 1024 * 1024
 // Mutable for tests.
 var maxHTMLPublishRawBytes int64 = 200 * 1024 * 1024
 
-// maxHTMLPublishSingleHTMLFileBytes 单个 .html 文件上限，对齐妙搭服务端 10MB 约束。
-// 用 var 而非 const，便于单测调小覆盖拦截路径。
-var maxHTMLPublishSingleHTMLFileBytes int64 = 10 * 1024 * 1024
-
-// oversizeHTMLFiles 返回 candidates 中扩展名为 .html（大小写不敏感）且单个 Size 超过
-// maxHTMLPublishSingleHTMLFileBytes 的 RelPath 列表。只针对 .html 文件，不波及图片/字体/JS。
-func oversizeHTMLFiles(candidates []htmlPublishCandidate) []string {
-	var hits []string
-	for _, c := range candidates {
-		if strings.EqualFold(filepath.Ext(c.RelPath), ".html") && c.Size > maxHTMLPublishSingleHTMLFileBytes {
-			hits = append(hits, c.RelPath)
-		}
-	}
-	return hits
-}
-
-// oversizeHTMLFilesError 构造单文件超限的 Validate 风格拒绝。
-func oversizeHTMLFilesError(hits []string) error {
-	return appsValidationParamError("--path",
-		"--path contains %d HTML file(s) exceeding the %d bytes (10MB) per-file limit: %s",
-		len(hits), maxHTMLPublishSingleHTMLFileBytes, truncatedJoin(hits, maxSensitiveListInError)).
-		WithHint("split or trim oversized HTML file(s); the 10MB cap applies to each single .html file")
-}
-
 // ensureIndexHTML 要求 walker 抓到的 candidates 里必须含 index.html。
 // 目录形态：根目录下必须有 index.html。
 // 单文件形态：文件名必须就是 index.html。
@@ -222,9 +190,6 @@ func runHTMLPublish(ctx context.Context, fio fileio.FileIO, publisher appsHTMLPu
 	if err := ensureIndexHTML(candidates); err != nil {
 		return nil, err
 	}
-	if hits := oversizeHTMLFiles(candidates); len(hits) > 0 {
-		return nil, oversizeHTMLFilesError(hits)
-	}
 	var rawTotal int64
 	for _, c := range candidates {
 		rawTotal += c.Size

shortcuts/apps/apps_html_publish_test.go

@@ -503,82 +503,3 @@ func TestRunHTMLPublish_RejectsOversizeRawCandidates(t *testing.T) {
 		t.Fatalf("client must not be called when raw cap hit")
 	}
 }
-
-func TestOversizeHTMLFiles(t *testing.T) {
-	orig := maxHTMLPublishSingleHTMLFileBytes
-	maxHTMLPublishSingleHTMLFileBytes = 100
-	defer func() { maxHTMLPublishSingleHTMLFileBytes = orig }()
-
-	cands := []htmlPublishCandidate{
-		{RelPath: "index.html", Size: 50},
-		{RelPath: "big.html", Size: 4096},
-		{RelPath: "BIG.HTML", Size: 4096}, // 大小写不敏感
-		{RelPath: "huge.png", Size: 9000}, // 非 .html，忽略
-	}
-	hits := oversizeHTMLFiles(cands)
-	if len(hits) != 2 {
-		t.Fatalf("hits=%v, want [big.html BIG.HTML]", hits)
-	}
-	for _, h := range hits {
-		if h == "huge.png" || h == "index.html" {
-			t.Fatalf("unexpected hit %q", h)
-		}
-	}
-}
-
-func TestMaxHTMLPublishSingleHTMLFileBytes_Default(t *testing.T) {
-	if maxHTMLPublishSingleHTMLFileBytes != 10*1024*1024 {
-		t.Fatalf("default=%d, want %d (10MiB)", maxHTMLPublishSingleHTMLFileBytes, 10*1024*1024)
-	}
-}
-
-func TestRunHTMLPublish_RejectsOversizeHTMLFile(t *testing.T) {
-	orig := maxHTMLPublishSingleHTMLFileBytes
-	maxHTMLPublishSingleHTMLFileBytes = 100
-	defer func() { maxHTMLPublishSingleHTMLFileBytes = orig }()
-
-	dir := t.TempDir()
-	if err := os.WriteFile(filepath.Join(dir, "index.html"), []byte("<html></html>"), 0o644); err != nil {
-		t.Fatalf("write: %v", err)
-	}
-	if err := os.WriteFile(filepath.Join(dir, "big.html"), []byte(strings.Repeat("x", 4096)), 0o644); err != nil {
-		t.Fatalf("write: %v", err)
-	}
-	fake := &fakeAppsHTMLPublishClient{}
-	_, err := runHTMLPublish(context.Background(), newTestFIO(), fake, appsHTMLPublishSpec{AppID: "app_x", Path: dir})
-	if err == nil {
-		t.Fatalf("expected per-file oversize error")
-	}
-	problem := requireAppsValidationProblem(t, err)
-	if !strings.Contains(problem.Message, "big.html") || !strings.Contains(problem.Message, "10MB") {
-		t.Fatalf("message=%q, want contains 'big.html' and '10MB'", problem.Message)
-	}
-	if problem.Hint == "" {
-		t.Fatalf("expected non-empty hint")
-	}
-	if len(fake.calls) != 0 {
-		t.Fatalf("client must not be called when an HTML file is oversize")
-	}
-}
-
-func TestRunHTMLPublish_IgnoresOversizeNonHTML(t *testing.T) {
-	// 单 .html 上限调小，但超限文件是 .png → 不被本护栏拦截，正常发布。
-	orig := maxHTMLPublishSingleHTMLFileBytes
-	maxHTMLPublishSingleHTMLFileBytes = 100
-	defer func() { maxHTMLPublishSingleHTMLFileBytes = orig }()
-
-	dir := t.TempDir()
-	if err := os.WriteFile(filepath.Join(dir, "index.html"), []byte("<html></html>"), 0o644); err != nil {
-		t.Fatalf("write: %v", err)
-	}
-	if err := os.WriteFile(filepath.Join(dir, "big.png"), []byte(strings.Repeat("x", 4096)), 0o644); err != nil {
-		t.Fatalf("write: %v", err)
-	}
-	fake := &fakeAppsHTMLPublishClient{resp: &htmlPublishResponse{URL: "https://miaoda/app_x"}}
-	if _, err := runHTMLPublish(context.Background(), newTestFIO(), fake, appsHTMLPublishSpec{AppID: "app_x", Path: dir}); err != nil {
-		t.Fatalf("non-html oversize must not be blocked by the .html cap: %v", err)
-	}
-	if len(fake.calls) != 1 {
-		t.Fatalf("client should be called; calls=%v", fake.calls)
-	}
-}

shortcuts/apps/apps_init.go

@@ -99,14 +99,7 @@ var AppsInit = common.Shortcut{
 			dry.Set("dir_error", err.Error())
 			dir = defaultCloneDir(appID)
 		} else if isAlreadyInitialized(dir) {
-			if existing, e := ensureInitDirMatchesApp(dir, appID); e != nil {
-				if existing != "" {
-					dry.Set("app_id_mismatch", existing)
-				}
-				dry.Set("dir_error", e.Error())
-			} else {
-				dry.Set("already_initialized", true)
-			}
+			dry.Set("already_initialized", true)
 		} else if e := ensureEmptyDir(dir); e != nil {
 			dry.Set("dir_error", e.Error())
 		}
@@ -206,61 +199,6 @@ func isAlreadyInitialized(dir string) bool {
 	return err == nil && !info.IsDir()
 }
 
-// readMetaAppID 读取 <dir>/.spark/meta.json 的 app_id，用于判断目标目录是否同一个妙搭应用。
-// 返回 (appID, isSparkProject, err)：
-//   - meta.json 不存在             → ("", false, nil)   非妙搭工程
-//   - 读取/解析失败（损坏/不可读）  → ("", false, err)   无法确认是否妙搭工程
-//   - 解析成功                     → (trim 后的 app_id, true, nil)（app_id 缺失/为空时为 ""）
-func readMetaAppID(dir string) (string, bool, error) {
-	b, err := os.ReadFile(filepath.Join(dir, metaRelPath)) //nolint:forbidigo // shortcuts cannot import internal/vfs (depguard rule shortcuts-no-vfs); path is under the validated clone dir, and FileIO.Open rejects absolute paths.
-	if os.IsNotExist(err) {
-		return "", false, nil
-	}
-	if err != nil {
-		return "", false, appsFileIOError(err, "read %s failed: %v", metaRelPath, err)
-	}
-	var m struct {
-		AppID string `json:"app_id"`
-	}
-	if err := json.Unmarshal(b, &m); err != nil {
-		return "", false, appsFileIOError(err, "parse %s failed: %v", metaRelPath, err)
-	}
-	return strings.TrimSpace(m.AppID), true, nil
-}
-
-// ensureInitDirMatchesApp 校验「已存在的目标目录」能否被 appID 安全复用：
-//   - 不是妙搭工程（无 meta.json）        → nil（交给 ensureEmptyDir 判空/非空）
-//   - 是妙搭工程且 app_id 与 appID 一致    → nil（走已初始化短路，复用本地代码）
-//   - 是妙搭工程但 app_id 不一致（含为空）  → 报错，提示换目录
-//   - meta.json 损坏/不可读，无法确认      → 报错（fail closed），提示换目录
-//
-// 返回值 existing 是目录里已存在的 app_id（仅"已是另一个 app"的拒绝场景非空），供调用方在
-// dry-run 里回填 app_id_mismatch，避免二次读 meta.json。
-func ensureInitDirMatchesApp(dir, appID string) (existing string, err error) {
-	existing, isSpark, readErr := readMetaAppID(dir)
-	if readErr != nil {
-		return "", appsValidationParamError("--dir",
-			"target directory %q already exists but its %s is unreadable or corrupted; cannot confirm it belongs to app %s, refusing to use it",
-			dir, metaRelPath, appID).
-			WithHint("choose a different --dir, or repair/remove the directory, before running +init").
-			WithCause(readErr)
-	}
-	if !isSpark || existing == appID {
-		return existing, nil
-	}
-	if existing == "" {
-		// meta 存在但缺 app_id：更可能是同一应用上次 +init 中断留下的半成品，而非另一个 app。
-		return "", appsValidationParamError("--dir",
-			"target directory %q has a %s without an app_id; cannot confirm it belongs to app %s, refusing to use it",
-			dir, metaRelPath, appID).
-			WithHint("remove the directory and re-run +init, or choose a different --dir")
-	}
-	return existing, appsValidationParamError("--dir",
-		"target directory %q is already initialized for a different app (%s); refusing to initialize app %s into it",
-		dir, existing, appID).
-		WithHint("choose a different --dir (or cd into the matching project) before running +init")
-}
-
 // ensureMetaAppID patches <dir>/.spark/meta.json to include app_id when the file
 // exists but lacks (or has an empty) app_id. Other fields are preserved. When
 // the file does not exist, this is a no-op (we never create it).
@@ -440,11 +378,6 @@ func appsInitExecute(ctx context.Context, rctx *common.RuntimeContext) error {
 		return err
 	}
 
-	// 异 app 目录护栏：拒绝把当前 app 初始化进另一个 app 的已初始化工程。
-	if _, err := ensureInitDirMatchesApp(dir, appID); err != nil {
-		return err
-	}
-
 	// Already-initialized short-circuit: a dir containing .spark/meta.json is an
 	// initialized app repo -> skip clone/scaffold/commit, but still refresh
 	// the local env so a re-run picks up the latest startup env vars.

shortcuts/apps/apps_init_test.go

@@ -363,7 +363,7 @@ func TestAppsInit_AlreadyInitialized_ShortCircuit(t *testing.T) {
 	if err := os.MkdirAll(filepath.Join(dir, ".spark"), 0o755); err != nil {
 		t.Fatal(err)
 	}
-	if err := os.WriteFile(filepath.Join(dir, metaRelPath), []byte(`{"app_id":"app_x"}`), 0o644); err != nil {
+	if err := os.WriteFile(filepath.Join(dir, metaRelPath), []byte(`{"app_id":"whatever"}`), 0o644); err != nil {
 		t.Fatal(err)
 	}
 	f := &fakeCommandRunner{results: map[string]fakeCallResult{"env-pull": envPullOK(filepath.Join(abs, ".env.local"))}}
@@ -394,40 +394,6 @@ func TestAppsInit_AlreadyInitialized_ShortCircuit(t *testing.T) {
 	}
 }
 
-func TestAppsInit_AlreadyInitialized_AppIDMismatch(t *testing.T) {
-	dir := relCloneDir(t)
-	if err := os.MkdirAll(filepath.Join(dir, ".spark"), 0o755); err != nil {
-		t.Fatal(err)
-	}
-	// 目录是 app_other 的工程，却用 --app-id app_x 初始化 → 必须报错且不拉 env。
-	if err := os.WriteFile(filepath.Join(dir, metaRelPath), []byte(`{"app_id":"app_other"}`), 0o644); err != nil {
-		t.Fatal(err)
-	}
-	f := &fakeCommandRunner{}
-	withFakeRunner(t, f)
-	factory, stdout, _ := newAppsExecuteFactory(t)
-	err := runAppsShortcut(t, AppsInit, []string{"+init", "--app-id", "app_x", "--dir", dir, "--as", "user"}, factory, stdout)
-	if err == nil {
-		t.Fatal("mismatched app_id must error")
-	}
-	problem := requireAppsValidationProblem(t, err)
-	if problem.Subtype != errs.SubtypeInvalidArgument {
-		t.Fatalf("subtype=%q, want %q", problem.Subtype, errs.SubtypeInvalidArgument)
-	}
-	var ve *errs.ValidationError
-	if !errors.As(err, &ve) || ve.Param != "--dir" {
-		t.Fatalf("expected *errs.ValidationError with Param=--dir, got %T param=%v", err, ve)
-	}
-	if !strings.Contains(problem.Message, "different app") {
-		t.Fatalf("message=%q, want 'different app'", problem.Message)
-	}
-	for _, c := range f.calls {
-		if containsAll(c, "+env-pull") || containsAll(c, "git", "clone") {
-			t.Errorf("mismatch must not run env-pull/clone; got %v", f.calls)
-		}
-	}
-}
-
 func TestAppsInit_HappyPathCleanTree(t *testing.T) {
 	f := &fakeCommandRunner{results: map[string]fakeCallResult{
 		"credential-init": credInitOK("http://u:t@h/app_x.git"),
@@ -1502,125 +1468,6 @@ func TestAppsInit_Description_IsAboutCode(t *testing.T) {
 	}
 }
 
-func TestReadMetaAppID(t *testing.T) {
-	writeMeta := func(t *testing.T, content string) string {
-		t.Helper()
-		dir := t.TempDir()
-		if err := os.MkdirAll(filepath.Join(dir, ".spark"), 0o755); err != nil {
-			t.Fatal(err)
-		}
-		if err := os.WriteFile(filepath.Join(dir, metaRelPath), []byte(content), 0o644); err != nil {
-			t.Fatal(err)
-		}
-		return dir
-	}
-
-	// 不存在 meta.json → ("", false, nil)
-	if got, ok, err := readMetaAppID(t.TempDir()); ok || got != "" || err != nil {
-		t.Fatalf("no meta: got (%q,%v,%v), want (\"\",false,nil)", got, ok, err)
-	}
-	// 存在且有 app_id → (app_id, true, nil)
-	if got, ok, err := readMetaAppID(writeMeta(t, `{"app_id":"app_a"}`)); !ok || got != "app_a" || err != nil {
-		t.Fatalf("with app_id: got (%q,%v,%v), want (\"app_a\",true,nil)", got, ok, err)
-	}
-	// 存在但 app_id 空 → ("", true, nil)
-	if got, ok, err := readMetaAppID(writeMeta(t, `{"name":"x"}`)); !ok || got != "" || err != nil {
-		t.Fatalf("empty app_id: got (%q,%v,%v), want (\"\",true,nil)", got, ok, err)
-	}
-	// 存在但坏 JSON → ("", false, err)（无法确认）
-	if got, ok, err := readMetaAppID(writeMeta(t, `{not json`)); ok || got != "" || err == nil {
-		t.Fatalf("bad json: got (%q,%v,err=%v), want (\"\",false,non-nil)", got, ok, err)
-	}
-}
-
-func TestEnsureInitDirMatchesApp(t *testing.T) {
-	writeMeta := func(t *testing.T, content string) string {
-		t.Helper()
-		dir := t.TempDir()
-		if err := os.MkdirAll(filepath.Join(dir, ".spark"), 0o755); err != nil {
-			t.Fatal(err)
-		}
-		if err := os.WriteFile(filepath.Join(dir, metaRelPath), []byte(content), 0o644); err != nil {
-			t.Fatal(err)
-		}
-		return dir
-	}
-
-	// 无 meta（非妙搭工程）→ nil（交给 ensureEmptyDir）
-	if _, err := ensureInitDirMatchesApp(t.TempDir(), "app_x"); err != nil {
-		t.Fatalf("no meta should pass: %v", err)
-	}
-	// 同 app_id → (app_id, nil)（走已初始化短路）
-	if existing, err := ensureInitDirMatchesApp(writeMeta(t, `{"app_id":"app_x"}`), "app_x"); err != nil || existing != "app_x" {
-		t.Fatalf("same app should pass: existing=%q err=%v", existing, err)
-	}
-
-	// 不同 app_id → error（换目录），返回 existing=app_other；断言 typed metadata（subtype/param）
-	existing, errMismatch := ensureInitDirMatchesApp(writeMeta(t, `{"app_id":"app_other"}`), "app_x")
-	if errMismatch == nil {
-		t.Fatal("different app should error")
-	}
-	if existing != "app_other" {
-		t.Fatalf("mismatch should return existing app_id, got %q", existing)
-	}
-	problem := requireAppsValidationProblem(t, errMismatch) // 已校验 Category==Validation
-	if problem.Subtype != errs.SubtypeInvalidArgument {
-		t.Fatalf("subtype=%q, want %q", problem.Subtype, errs.SubtypeInvalidArgument)
-	}
-	var ve *errs.ValidationError
-	if !errors.As(errMismatch, &ve) {
-		t.Fatalf("expected *errs.ValidationError, got %T", errMismatch)
-	}
-	if ve.Param != "--dir" {
-		t.Fatalf("param=%q, want --dir", ve.Param)
-	}
-	if !strings.Contains(problem.Message, "different app") || !strings.Contains(problem.Message, "app_other") {
-		t.Fatalf("message=%q, want 'different app' and 'app_other'", problem.Message)
-	}
-	if !strings.Contains(problem.Hint, "different --dir") {
-		t.Fatalf("hint=%q, want 'different --dir'", problem.Hint)
-	}
-
-	// 空 app_id（缺 app_id 标记的半成品）→ error，独立文案（非 "different app"），返回 existing=""
-	emptyExisting, errEmpty := ensureInitDirMatchesApp(writeMeta(t, `{"name":"x"}`), "app_x")
-	if errEmpty == nil {
-		t.Fatal("empty meta app_id should error (cannot confirm same app)")
-	}
-	if emptyExisting != "" {
-		t.Fatalf("empty app_id should return existing=\"\", got %q", emptyExisting)
-	}
-	pEmpty := requireAppsValidationProblem(t, errEmpty)
-	if pEmpty.Subtype != errs.SubtypeInvalidArgument {
-		t.Fatalf("empty subtype=%q, want %q", pEmpty.Subtype, errs.SubtypeInvalidArgument)
-	}
-	if !strings.Contains(pEmpty.Message, "without an app_id") {
-		t.Fatalf("empty app_id should have its own message, msg=%q", pEmpty.Message)
-	}
-	if strings.Contains(pEmpty.Message, "different app") {
-		t.Fatalf("empty app_id must not reuse the different-app wording, msg=%q", pEmpty.Message)
-	}
-
-	// meta 损坏/不可读 → error（fail closed），返回 existing=""
-	badExisting, errBad := ensureInitDirMatchesApp(writeMeta(t, `{not json`), "app_x")
-	if errBad == nil {
-		t.Fatal("corrupted meta should fail closed")
-	}
-	if badExisting != "" {
-		t.Fatalf("corrupted should return existing=\"\", got %q", badExisting)
-	}
-	pBad := requireAppsValidationProblem(t, errBad)
-	if pBad.Subtype != errs.SubtypeInvalidArgument {
-		t.Fatalf("corrupted subtype=%q, want %q", pBad.Subtype, errs.SubtypeInvalidArgument)
-	}
-	if !strings.Contains(pBad.Message, "unreadable or corrupted") {
-		t.Fatalf("corrupted meta msg=%q, want 'unreadable or corrupted'", pBad.Message)
-	}
-	var veBad *errs.ValidationError
-	if !errors.As(errBad, &veBad) || veBad.Param != "--dir" {
-		t.Fatalf("corrupted: expected ValidationError Param=--dir, got %T param=%v", errBad, veBad)
-	}
-}
-
 // TestRunScaffold_SubprocessFailureIsExternalTool pins the typed
 // classification of an external-tool failure: a failing git subprocess
 // surfaces as internal/external_tool with the cause preserved.

shortcuts/base/base_resolve.go

@@ -1,545 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package base
-
-import (
-	"context"
-	"fmt"
-	"net/url"
-	"regexp"
-	"strings"
-
-	"github.com/larksuite/cli/errs"
-	"github.com/larksuite/cli/shortcuts/common"
-)
-
-const (
-	baseURLResolveHintGeneric = "Provide a /base/, /wiki/, or /record/ URL, or use base +title-resolve --title if you only know the Base title."
-	baseTitleResolveHint      = "choose one candidate, then use +base-block-list to list tables, dashboards, workflows, and other Base blocks"
-	nextStepBaseBlockList     = "use +base-block-list to list tables, dashboards, workflows, and other Base blocks"
-	nextStepRecordList        = "use +record-list to list records in the resolved table"
-	titleResolveQueryMaxLen   = 30
-)
-
-var BaseURLResolve = common.Shortcut{
-	Service:     "base",
-	Command:     "+url-resolve",
-	Description: "Resolve a Base-related URL into Base coordinates",
-	Risk:        "read",
-	Scopes:      []string{},
-	ConditionalScopes: []string{
-		"base:field:read",
-		"base:record:read",
-		"wiki:node:retrieve",
-	},
-	AuthTypes: authTypes(),
-	HasFormat: true,
-	Flags: []common.Flag{
-		{Name: "url", Desc: "Base/Wiki/record-share URL to resolve"},
-		{Name: "query", Hidden: true, Desc: "Alias for --url; accepted to recover from AI routing mistakes"},
-	},
-	Tips: []string{
-		`Example: lark-cli base +url-resolve --url "https://example.larkoffice.com/base/<base_token>?table=<table_id>&view=<view_id>"`,
-		"Only URLs are accepted. For Base titles or keywords, use +title-resolve --title.",
-	},
-	Validate: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		_, err := readURLResolveInput(runtime)
-		return err
-	},
-	DryRun: func(ctx context.Context, runtime *common.RuntimeContext) *common.DryRunAPI {
-		raw, err := readURLResolveInput(runtime)
-		if err != nil {
-			return common.NewDryRunAPI().Set("error", err.Error())
-		}
-		parsed, err := parseResolveURL(raw)
-		if err != nil {
-			return common.NewDryRunAPI().Set("error", err.Error())
-		}
-		switch classifyBaseURL(parsed) {
-		case "wiki_url":
-			return common.NewDryRunAPI().
-				GET("/open-apis/wiki/v2/spaces/get_node").
-				Params(map[string]interface{}{"token": firstPathSegmentAfter(parsed.Path, "/wiki/")})
-		case "record_share_url":
-			return common.NewDryRunAPI().
-				GET("/open-apis/base/v3/record_share/:record_share_token/meta").
-				Set("record_share_token", firstPathSegmentAfter(parsed.Path, "/record/"))
-		default:
-			return common.NewDryRunAPI().Set("url", raw).Set("resolution", "local")
-		}
-	},
-	Execute: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		return executeBaseURLResolve(runtime)
-	},
-}
-
-var BaseTitleResolve = common.Shortcut{
-	Service:     "base",
-	Command:     "+title-resolve",
-	Description: "Resolve a Base title or keyword through Drive search",
-	Risk:        "read",
-	Scopes:      []string{"search:docs:read"},
-	AuthTypes:   []string{"user"},
-	HasFormat:   true,
-	Flags: []common.Flag{
-		{Name: "title", Desc: "Base title keyword to search via Drive (30 characters or fewer)"},
-		{Name: "query", Hidden: true, Desc: "Alias for --title; accepted to recover from AI routing mistakes"},
-		{Name: "url", Hidden: true, Desc: "Alias for --title; accepted to recover from AI routing mistakes"},
-	},
-	Tips: []string{
-		`Example: lark-cli base +title-resolve --title "Sales pipeline"`,
-		"Pass a short keyword from the Base title, 30 characters or fewer. Use +url-resolve for URLs.",
-	},
-	Validate: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		_, err := readTitleResolveQuery(runtime)
-		return err
-	},
-	DryRun: func(ctx context.Context, runtime *common.RuntimeContext) *common.DryRunAPI {
-		query, err := readTitleResolveQuery(runtime)
-		if err != nil {
-			return common.NewDryRunAPI().Set("error", err.Error())
-		}
-		return common.NewDryRunAPI().
-			POST("/open-apis/search/v2/doc_wiki/search").
-			Body(buildTitleResolveSearchBody(query))
-	},
-	Execute: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		return executeBaseTitleResolve(runtime)
-	},
-}
-
-func readURLResolveInput(runtime *common.RuntimeContext) (string, error) {
-	urlValue := strings.TrimSpace(runtime.Str("url"))
-	queryValue := strings.TrimSpace(runtime.Str("query"))
-	if urlValue != "" && queryValue != "" {
-		return "", baseFlagErrorf("--url and --query are mutually exclusive")
-	}
-	value := urlValue
-	if value == "" {
-		value = queryValue
-	}
-	if value == "" {
-		return "", baseFlagErrorf("specify --url")
-	}
-	return value, nil
-}
-
-func readTitleResolveQuery(runtime *common.RuntimeContext) (string, error) {
-	values := []struct {
-		name  string
-		value string
-	}{
-		{"title", strings.TrimSpace(runtime.Str("title"))},
-		{"query", strings.TrimSpace(runtime.Str("query"))},
-		{"url", strings.TrimSpace(runtime.Str("url"))},
-	}
-	var pickedName, pickedValue string
-	for _, v := range values {
-		if v.value == "" {
-			continue
-		}
-		if pickedValue != "" {
-			return "", baseFlagErrorf("--%s and --%s are mutually exclusive", pickedName, v.name)
-		}
-		pickedName = v.name
-		pickedValue = v.value
-	}
-	if pickedValue == "" {
-		return "", baseFlagErrorf("specify --title")
-	}
-	if len([]rune(pickedValue)) > titleResolveQueryMaxLen {
-		return "", resolveValidationError(
-			fmt.Sprintf("base +title-resolve title keyword must be %d characters or fewer.", titleResolveQueryMaxLen),
-			"Use a shorter keyword from the Base title, or provide a /base/ URL and use base +url-resolve.",
-		)
-	}
-	return pickedValue, nil
-}
-
-func executeBaseURLResolve(runtime *common.RuntimeContext) error {
-	raw, err := readURLResolveInput(runtime)
-	if err != nil {
-		return err
-	}
-	parsed, err := parseResolveURL(raw)
-	if err != nil {
-		return err
-	}
-
-	switch classifyBaseURL(parsed) {
-	case "base_url":
-		out := resolveBaseURL(parsed)
-		enrichBaseResolveHint(runtime, out)
-		runtime.OutFormat(out, nil, nil)
-		return nil
-	case "wiki_url":
-		out, err := resolveWikiBaseURL(runtime, parsed)
-		if err != nil {
-			return err
-		}
-		runtime.OutFormat(out, nil, nil)
-		return nil
-	case "record_share_url":
-		out, err := resolveRecordShareURL(runtime, parsed)
-		if err != nil {
-			return err
-		}
-		runtime.OutFormat(out, nil, nil)
-		return nil
-	case "form_share_url":
-		runtime.OutFormat(resolveFormShareURL(parsed), nil, nil)
-		return nil
-	case "view_share_url":
-		return resolveValidationError(
-			"This is a Base view share URL. CLI does not support resolving Base view share URLs.",
-			"Open it in the browser, or provide the URL of the Base itself, such as its Wiki URL or Base URL.",
-		)
-	case "dashboard_share_url":
-		return resolveValidationError(
-			"This is a Base dashboard share URL. CLI does not support resolving Base dashboard share URLs.",
-			"Open it in the browser, or provide the URL of the Base itself, such as its Wiki URL or Base URL.",
-		)
-	case "workspace_url":
-		return resolveValidationError(
-			"This is a Base workspace URL. CLI does not support resolving Base workspace URLs.",
-			"Open it in the browser, or provide the URL of the Base itself, such as its Wiki URL or Base URL.",
-		)
-	case "add_record_url":
-		return resolveValidationError(
-			"This is a Base add-record URL. CLI does not support resolving Base add-record URLs.",
-			"Open it in the browser, or provide the URL of the Base itself, such as its Wiki URL or Base URL.",
-		)
-	default:
-		return resolveValidationError("This URL is not a supported Base URL pattern.", baseURLResolveHintGeneric)
-	}
-}
-
-func parseResolveURL(raw string) (*url.URL, error) {
-	parsed, err := url.Parse(strings.TrimSpace(raw))
-	if err != nil || parsed.Scheme == "" || parsed.Host == "" {
-		return nil, resolveValidationError("base +url-resolve only accepts full URLs.", "For a Base title or keyword, use base +title-resolve --title.")
-	}
-	if parsed.Scheme != "http" && parsed.Scheme != "https" {
-		return nil, resolveValidationError("base +url-resolve only accepts HTTP or HTTPS URLs.", baseURLResolveHintGeneric)
-	}
-	return parsed, nil
-}
-
-func classifyBaseURL(u *url.URL) string {
-	path := normalizeResolvePath(u.Path)
-	switch {
-	case pathSegmentExists(path, "/base/workspace/"):
-		return "workspace_url"
-	case pathSegmentExists(path, "/base/add/"):
-		return "add_record_url"
-	case pathSegmentExists(path, "/base/"):
-		return "base_url"
-	case pathSegmentExists(path, "/wiki/"):
-		return "wiki_url"
-	case pathSegmentExists(path, "/record/"):
-		return "record_share_url"
-	case pathSegmentExists(path, "/share/base/form/"):
-		return "form_share_url"
-	case pathSegmentExists(path, "/share/base/view/"):
-		return "view_share_url"
-	case pathSegmentExists(path, "/share/base/dashboard/"):
-		return "dashboard_share_url"
-	default:
-		return ""
-	}
-}
-
-func resolveBaseURL(u *url.URL) map[string]interface{} {
-	query := u.Query()
-	out := map[string]interface{}{
-		"input_type":    "base_url",
-		"resource_type": "bitable",
-		"base_token":    firstPathSegmentAfter(u.Path, "/base/"),
-	}
-	if tableID := strings.TrimSpace(query.Get("table")); tableID != "" {
-		out["table_id"] = tableID
-	}
-	if viewID := strings.TrimSpace(query.Get("view")); viewID != "" {
-		out["view_id"] = viewID
-	}
-	if recordID := strings.TrimSpace(query.Get("record")); recordID != "" {
-		out["record_id"] = recordID
-	}
-	return out
-}
-
-func resolveWikiBaseURL(runtime *common.RuntimeContext, u *url.URL) (map[string]interface{}, error) {
-	token := firstPathSegmentAfter(u.Path, "/wiki/")
-	data, err := runtime.CallAPITyped("GET", "/open-apis/wiki/v2/spaces/get_node", map[string]interface{}{"token": token}, nil)
-	if err != nil {
-		return nil, err
-	}
-	node := common.GetMap(data, "node")
-	objType := strings.TrimSpace(common.GetString(node, "obj_type"))
-	if objType != "bitable" {
-		return nil, resolveValidationError(
-			fmt.Sprintf("This Wiki URL resolves to %s, not Base.", valueOrUnknown(objType)),
-			"Use the corresponding skill for that resource, or provide a Base URL.",
-		)
-	}
-	baseToken := strings.TrimSpace(common.GetString(node, "obj_token"))
-	if baseToken == "" {
-		return nil, errs.NewInternalError(errs.SubtypeInvalidResponse, "wiki node response is missing obj_token")
-	}
-	return map[string]interface{}{
-		"input_type":      "wiki_url",
-		"resource_type":   "bitable",
-		"wiki_node_token": token,
-		"base_token":      baseToken,
-		"title":           common.GetString(node, "title"),
-		"hint":            resolveHint("", nil),
-	}, nil
-}
-
-func resolveRecordShareURL(runtime *common.RuntimeContext, u *url.URL) (map[string]interface{}, error) {
-	shareToken := firstPathSegmentAfter(u.Path, "/record/")
-	data, err := baseV3Call(runtime, "GET", baseV3Path("record_share", shareToken, "meta"), nil, nil)
-	if err != nil {
-		return nil, err
-	}
-	out := map[string]interface{}{
-		"input_type":         "record_share_url",
-		"resource_type":      "bitable",
-		"record_share_token": firstNonEmpty(common.GetString(data, "record_share_token"), shareToken),
-		"base_token":         common.GetString(data, "base_token"),
-		"table_id":           common.GetString(data, "table_id"),
-		"record_id":          common.GetString(data, "record_id"),
-	}
-	enrichRecordShareResolveHint(runtime, out)
-	return out, nil
-}
-
-func resolveFormShareURL(u *url.URL) map[string]interface{} {
-	return map[string]interface{}{
-		"input_type":    "form_share_url",
-		"resource_type": "bitable_form",
-		"share_token":   firstPathSegmentAfter(u.Path, "/share/base/form/"),
-		"hint": map[string]interface{}{
-			"next_step": "use +form-detail to inspect the form, or use +form-submit to submit a response",
-		},
-	}
-}
-
-func executeBaseTitleResolve(runtime *common.RuntimeContext) error {
-	query, err := readTitleResolveQuery(runtime)
-	if err != nil {
-		return err
-	}
-	data, err := runtime.CallAPITyped("POST", "/open-apis/search/v2/doc_wiki/search", nil, buildTitleResolveSearchBody(query))
-	if err != nil {
-		return err
-	}
-	candidates := normalizeTitleResolveCandidates(common.GetSlice(data, "res_units"))
-	switch len(candidates) {
-	case 0:
-		return resolveValidationError(
-			"No Base matched this title or keyword.",
-			"Try a more specific Base title, or provide a /base/ URL and use base +url-resolve.",
-		)
-	case 1:
-		out := map[string]interface{}{
-			"input_type":    "title_query",
-			"resource_type": "bitable",
-			"title":         candidates[0]["title"],
-			"base_token":    candidates[0]["base_token"],
-			"url":           candidates[0]["url"],
-			"owner_name":    candidates[0]["owner_name"],
-			"update_time":   candidates[0]["update_time"],
-			"hint":          resolveHint("", nil),
-		}
-		runtime.OutFormat(out, nil, nil)
-		return nil
-	default:
-		runtime.OutFormat(map[string]interface{}{
-			"input_type":    "title_query",
-			"resource_type": "bitable",
-			"candidates":    candidates,
-			"hint": map[string]interface{}{
-				"next_step": baseTitleResolveHint,
-			},
-		}, nil, nil)
-		return nil
-	}
-}
-
-func enrichBaseResolveHint(runtime *common.RuntimeContext, out map[string]interface{}) {
-	baseToken := strings.TrimSpace(common.GetString(out, "base_token"))
-	tableID := strings.TrimSpace(common.GetString(out, "table_id"))
-	if baseToken == "" || tableID == "" {
-		out["hint"] = resolveHint("", nil)
-		return
-	}
-	fields, total, err := listAllFields(runtime, baseToken, tableID, 0, 100)
-	if err != nil {
-		out["hint"] = resolveHint(tableID, nil)
-		return
-	}
-	out["hint"] = resolveHint(tableID, map[string]interface{}{"fields": map[string]interface{}{"fields": fields, "total": total}})
-}
-
-func enrichRecordShareResolveHint(runtime *common.RuntimeContext, out map[string]interface{}) {
-	baseToken := strings.TrimSpace(common.GetString(out, "base_token"))
-	tableID := strings.TrimSpace(common.GetString(out, "table_id"))
-	recordID := strings.TrimSpace(common.GetString(out, "record_id"))
-	hint := map[string]interface{}{}
-	if baseToken != "" && tableID != "" && recordID != "" {
-		if record, err := getResolveRecord(runtime, baseToken, tableID, recordID); err == nil {
-			hint["record_data"] = formatResolvedRecordData(record)
-		}
-	}
-	if baseToken != "" && tableID != "" {
-		if fields, total, err := listAllFields(runtime, baseToken, tableID, 0, 100); err == nil {
-			hint["fields"] = map[string]interface{}{"fields": fields, "total": total}
-		}
-	}
-	out["hint"] = resolveHint(tableID, hint)
-	common.GetMap(out, "hint")["next_step"] = recordShareNextStep(baseToken, tableID, recordID)
-}
-
-func getResolveRecord(runtime *common.RuntimeContext, baseToken, tableID, recordID string) (map[string]interface{}, error) {
-	body := map[string]interface{}{"record_id_list": []string{recordID}}
-	result, err := baseV3Raw(runtime, "POST", baseV3Path("bases", baseToken, "tables", tableID, "records", "batch_get"), nil, body)
-	return handleBaseAPIResult(result, err, "batch get records")
-}
-
-func formatResolvedRecordData(record map[string]interface{}) map[string]interface{} {
-	fieldIDs := common.GetSlice(record, "field_id_list")
-	fieldNames := common.GetSlice(record, "fields")
-	rows := common.GetSlice(record, "data")
-
-	data := map[string]interface{}{}
-	if len(rows) > 0 {
-		if values, ok := rows[0].([]interface{}); ok {
-			for i, value := range values {
-				data[resolvedRecordFieldKey(fieldIDs, fieldNames, i)] = value
-			}
-		}
-	}
-	return data
-}
-
-func resolvedRecordFieldKey(fieldIDs, fieldNames []interface{}, index int) string {
-	if index < len(fieldIDs) {
-		if fieldID := strings.TrimSpace(fmt.Sprintf("%v", fieldIDs[index])); fieldID != "" {
-			return fieldID
-		}
-	}
-	if index < len(fieldNames) {
-		if fieldName := strings.TrimSpace(fmt.Sprintf("%v", fieldNames[index])); fieldName != "" {
-			return fieldName
-		}
-	}
-	return fmt.Sprintf("field_%d", index+1)
-}
-
-func recordShareNextStep(baseToken, tableID, recordID string) string {
-	return fmt.Sprintf(`use +record-upsert --base-token %s --table-id %s --record-id %s --json '{"<field_id>":"<new_value>"}' to update this record`, baseToken, tableID, recordID)
-}
-
-func resolveHint(tableID string, extra map[string]interface{}) map[string]interface{} {
-	hint := map[string]interface{}{}
-	for key, value := range extra {
-		hint[key] = value
-	}
-	if strings.TrimSpace(tableID) != "" {
-		hint["next_step"] = nextStepRecordList
-	} else {
-		hint["next_step"] = nextStepBaseBlockList
-	}
-	return hint
-}
-
-func buildTitleResolveSearchBody(query string) map[string]interface{} {
-	filter := map[string]interface{}{"doc_types": []string{"BITABLE"}}
-	return map[string]interface{}{
-		"query":       query,
-		"page_size":   5,
-		"doc_filter":  filter,
-		"wiki_filter": filter,
-	}
-}
-
-func normalizeTitleResolveCandidates(items []interface{}) []map[string]interface{} {
-	candidates := make([]map[string]interface{}, 0, len(items))
-	for _, item := range items {
-		row, _ := item.(map[string]interface{})
-		meta, _ := row["result_meta"].(map[string]interface{})
-		if row == nil || meta == nil || strings.ToUpper(common.GetString(meta, "doc_types")) != "BITABLE" {
-			continue
-		}
-		token := strings.TrimSpace(common.GetString(meta, "token"))
-		if token == "" {
-			continue
-		}
-		title := stripSearchHighlight(common.GetString(row, "title_highlighted"))
-		if title == "" {
-			title = strings.TrimSpace(common.GetString(row, "title"))
-		}
-		candidates = append(candidates, map[string]interface{}{
-			"title":       title,
-			"base_token":  token,
-			"url":         common.GetString(meta, "url"),
-			"owner_name":  common.GetString(meta, "owner_name"),
-			"update_time": common.GetString(meta, "update_time_iso"),
-		})
-	}
-	return candidates
-}
-
-var searchHighlightTagRe = regexp.MustCompile(`</?h>`)
-
-func stripSearchHighlight(s string) string {
-	return strings.TrimSpace(searchHighlightTagRe.ReplaceAllString(s, ""))
-}
-
-func resolveValidationError(message, hint string) error {
-	return errs.NewValidationError(errs.SubtypeInvalidArgument, "%s", message).WithHint("%s", hint)
-}
-
-func normalizeResolvePath(path string) string {
-	if path == "" {
-		return "/"
-	}
-	if !strings.HasPrefix(path, "/") {
-		path = "/" + path
-	}
-	return path
-}
-
-func pathSegmentExists(path, prefix string) bool {
-	return firstPathSegmentAfter(path, prefix) != ""
-}
-
-func firstPathSegmentAfter(path, prefix string) string {
-	path = normalizeResolvePath(path)
-	if !strings.HasPrefix(path, prefix) {
-		return ""
-	}
-	rest := path[len(prefix):]
-	if idx := strings.IndexByte(rest, '/'); idx >= 0 {
-		rest = rest[:idx]
-	}
-	return strings.TrimSpace(rest)
-}
-
-func valueOrUnknown(s string) string {
-	if strings.TrimSpace(s) == "" {
-		return "an unknown resource type"
-	}
-	return s
-}
-
-func firstNonEmpty(values ...string) string {
-	for _, value := range values {
-		if strings.TrimSpace(value) != "" {
-			return strings.TrimSpace(value)
-		}
-	}
-	return ""
-}

shortcuts/base/base_resolve_test.go

@@ -1,454 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package base
-
-import (
-	"strings"
-	"testing"
-
-	"github.com/larksuite/cli/errs"
-	"github.com/larksuite/cli/internal/cmdutil"
-	"github.com/larksuite/cli/internal/httpmock"
-	"github.com/larksuite/cli/shortcuts/common"
-	"github.com/spf13/cobra"
-)
-
-func TestBaseURLResolveBaseURL(t *testing.T) {
-	t.Run("with coordinates", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(fieldListStub("bas123", "tbl123"))
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve",
-			"--url", "https://example.larkoffice.com/base/bas123?table=tbl123&view=vew123&record=rec123",
-			"--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-
-		data := decodeBaseEnvelope(t, stdout)
-		if data["input_type"] != "base_url" || data["base_token"] != "bas123" {
-			t.Fatalf("unexpected output: %#v", data)
-		}
-		if data["table_id"] != "tbl123" || data["view_id"] != "vew123" || data["record_id"] != "rec123" {
-			t.Fatalf("missing Base coordinates: %#v", data)
-		}
-		hint, _ := data["hint"].(map[string]interface{})
-		fields, _ := hint["fields"].(map[string]interface{})
-		if hint["next_step"] != nextStepRecordList || fields["total"] != float64(2) {
-			t.Fatalf("unexpected hint: %#v", hint)
-		}
-	})
-
-	t.Run("base only", func(t *testing.T) {
-		factory, stdout, _ := newExecuteFactory(t)
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve", "--url", "https://example.larkoffice.com/base/bas123", "--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-		data := decodeBaseEnvelope(t, stdout)
-		if data["input_type"] != "base_url" || data["base_token"] != "bas123" {
-			t.Fatalf("unexpected output: %#v", data)
-		}
-		if _, ok := data["table_id"]; ok {
-			t.Fatalf("table_id should be omitted for base-only URL: %#v", data)
-		}
-		hint, _ := data["hint"].(map[string]interface{})
-		if hint["next_step"] != nextStepBaseBlockList {
-			t.Fatalf("unexpected hint: %#v", hint)
-		}
-	})
-
-	t.Run("field list enrichment failure still returns coordinates", func(t *testing.T) {
-		factory, stdout, _ := newExecuteFactory(t)
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve", "--url", "https://example.larkoffice.com/base/bas123?table=tbl123", "--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-		data := decodeBaseEnvelope(t, stdout)
-		if data["base_token"] != "bas123" || data["table_id"] != "tbl123" {
-			t.Fatalf("unexpected output: %#v", data)
-		}
-		hint, _ := data["hint"].(map[string]interface{})
-		if hint["next_step"] != nextStepRecordList {
-			t.Fatalf("unexpected hint: %#v", hint)
-		}
-		if _, ok := hint["fields"]; ok {
-			t.Fatalf("fields should be omitted when enrichment fails: %#v", hint)
-		}
-	})
-}
-
-func TestBaseURLResolveWikiURL(t *testing.T) {
-	t.Run("bitable", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(&httpmock.Stub{
-			Method: "GET",
-			URL:    "/open-apis/wiki/v2/spaces/get_node?token=wik123",
-			Body: map[string]interface{}{
-				"code": 0,
-				"data": map[string]interface{}{
-					"node": map[string]interface{}{
-						"obj_type":  "bitable",
-						"obj_token": "bas123",
-						"title":     "Demo Base",
-					},
-				},
-			},
-		})
-
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve", "--url", "https://example.larkoffice.com/wiki/wik123", "--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-		data := decodeBaseEnvelope(t, stdout)
-		if data["input_type"] != "wiki_url" || data["base_token"] != "bas123" || data["title"] != "Demo Base" {
-			t.Fatalf("unexpected output: %#v", data)
-		}
-	})
-
-	t.Run("non bitable", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(&httpmock.Stub{
-			Method: "GET",
-			URL:    "/open-apis/wiki/v2/spaces/get_node?token=wikdoc",
-			Body: map[string]interface{}{
-				"code": 0,
-				"data": map[string]interface{}{
-					"node": map[string]interface{}{"obj_type": "docx", "obj_token": "docx123"},
-				},
-			},
-		})
-
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve", "--url", "https://example.larkoffice.com/wiki/wikdoc", "--as", "user",
-		}, factory, stdout)
-		if err == nil || !strings.Contains(err.Error(), "not Base") {
-			t.Fatalf("err=%v, want non-Base validation error", err)
-		}
-	})
-}
-
-func TestBaseURLResolveRecordShareURL(t *testing.T) {
-	t.Run("enriched", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(recordShareMetaStub("shr123", "bas123", "tbl123", "rec123"))
-		reg.Register(recordBatchGetStub("bas123", "tbl123", "rec123"))
-		reg.Register(fieldListStub("bas123", "tbl123"))
-
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve", "--url", "https://example.larkoffice.com/record/shr123", "--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-		data := decodeBaseEnvelope(t, stdout)
-		if data["input_type"] != "record_share_url" || data["base_token"] != "bas123" || data["record_id"] != "rec123" {
-			t.Fatalf("unexpected output: %#v", data)
-		}
-		hint, _ := data["hint"].(map[string]interface{})
-		recordData, _ := hint["record_data"].(map[string]interface{})
-		fields, _ := hint["fields"].(map[string]interface{})
-		nextStep, _ := hint["next_step"].(string)
-		if !strings.Contains(nextStep, "+record-upsert --base-token bas123 --table-id tbl123 --record-id rec123") || recordData["fld_name"] != "Alice" || fields["total"] != float64(2) {
-			t.Fatalf("unexpected hint: %#v", hint)
-		}
-	})
-
-	t.Run("enrichment failure still returns meta", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(recordShareMetaStub("shr123", "bas123", "tbl123", "rec123"))
-
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve", "--url", "https://example.larkoffice.com/record/shr123", "--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-		data := decodeBaseEnvelope(t, stdout)
-		if data["input_type"] != "record_share_url" || data["base_token"] != "bas123" || data["record_id"] != "rec123" {
-			t.Fatalf("unexpected output: %#v", data)
-		}
-		hint, _ := data["hint"].(map[string]interface{})
-		nextStep, _ := hint["next_step"].(string)
-		if !strings.Contains(nextStep, "+record-upsert --base-token bas123 --table-id tbl123 --record-id rec123") {
-			t.Fatalf("unexpected hint: %#v", hint)
-		}
-		if _, ok := hint["record_data"]; ok {
-			t.Fatalf("record_data should be omitted when enrichment fails: %#v", hint)
-		}
-		if _, ok := hint["fields"]; ok {
-			t.Fatalf("fields should be omitted when enrichment fails: %#v", hint)
-		}
-	})
-}
-
-func recordShareMetaStub(shareToken, baseToken, tableID, recordID string) *httpmock.Stub {
-	return &httpmock.Stub{
-		Method: "GET",
-		URL:    "/open-apis/base/v3/record_share/" + shareToken + "/meta",
-		Body: map[string]interface{}{
-			"code": 0,
-			"data": map[string]interface{}{
-				"record_share_token": shareToken,
-				"base_token":         baseToken,
-				"table_id":           tableID,
-				"record_id":          recordID,
-			},
-		},
-	}
-}
-
-func TestBaseURLResolveFormShareURL(t *testing.T) {
-	factory, stdout, _ := newExecuteFactory(t)
-	err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-		"+url-resolve", "--query", "https://example.larkoffice.com/share/base/form/shrform", "--as", "user",
-	}, factory, stdout)
-	if err != nil {
-		t.Fatalf("err=%v", err)
-	}
-	data := decodeBaseEnvelope(t, stdout)
-	if data["input_type"] != "form_share_url" || data["share_token"] != "shrform" {
-		t.Fatalf("unexpected output: %#v", data)
-	}
-}
-
-func TestBaseURLResolveValidationErrors(t *testing.T) {
-	tests := []struct {
-		name     string
-		rawURL   string
-		wantText string
-		wantHint string
-	}{
-		{"dashboard share", "https://example.larkoffice.com/share/base/dashboard/shr1", "CLI does not support resolving Base dashboard share URLs", "provide the URL of the Base itself"},
-		{"view share", "https://example.larkoffice.com/share/base/view/shr1", "CLI does not support resolving Base view share URLs", "provide the URL of the Base itself"},
-		{"workspace", "https://example.larkoffice.com/base/workspace/ws1", "CLI does not support resolving Base workspace URLs", "provide the URL of the Base itself"},
-		{"add record", "https://example.larkoffice.com/base/add/addtoken", "CLI does not support resolving Base add-record URLs", "provide the URL of the Base itself"},
-		{"unrelated", "https://example.larkoffice.com/docx/doc1", "not a supported Base URL pattern", ""},
-		{"not url", "bas123", "only accepts full URLs", ""},
-	}
-	for _, tc := range tests {
-		t.Run(tc.name, func(t *testing.T) {
-			factory, stdout, _ := newExecuteFactory(t)
-			err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-				"+url-resolve", "--url", tc.rawURL, "--as", "user",
-			}, factory, stdout)
-			if err == nil || !strings.Contains(err.Error(), tc.wantText) {
-				t.Fatalf("err=%v, want contains %q", err, tc.wantText)
-			}
-			p, ok := errs.ProblemOf(err)
-			if !ok || p.Hint == "" {
-				t.Fatalf("err=%v, want typed error with hint", err)
-			}
-			if tc.wantHint != "" && !strings.Contains(p.Hint, tc.wantHint) {
-				t.Fatalf("hint=%q, want contains %q", p.Hint, tc.wantHint)
-			}
-			if strings.Contains(p.Hint, "original /base/{base_token}") {
-				t.Fatalf("hint should not require original /base URL: %q", p.Hint)
-			}
-		})
-	}
-}
-
-func TestBaseResolveInputXOR(t *testing.T) {
-	t.Run("url resolve", func(t *testing.T) {
-		factory, stdout, _ := newExecuteFactory(t)
-		err := runShortcutWithAuthTypes(t, BaseURLResolve, authTypes(), []string{
-			"+url-resolve", "--url", "https://example.com/base/bas1", "--query", "https://example.com/base/bas2", "--as", "user",
-		}, factory, stdout)
-		if err == nil || !strings.Contains(err.Error(), "mutually exclusive") {
-			t.Fatalf("err=%v, want xor validation", err)
-		}
-	})
-
-	t.Run("title resolve", func(t *testing.T) {
-		factory, stdout, _ := newExecuteFactory(t)
-		err := runShortcutWithAuthTypes(t, BaseTitleResolve, nil, []string{
-			"+title-resolve", "--title", "Pipeline", "--query", "Sales", "--as", "user",
-		}, factory, stdout)
-		if err == nil || !strings.Contains(err.Error(), "mutually exclusive") {
-			t.Fatalf("err=%v, want xor validation", err)
-		}
-	})
-}
-
-func TestBaseResolveHelpFlags(t *testing.T) {
-	for _, tc := range []struct {
-		shortcut    string
-		definition  common.Shortcut
-		primaryFlag string
-		primaryDesc string
-		aliasFlags  []string
-	}{
-		{
-			shortcut:    "+url-resolve",
-			definition:  BaseURLResolve,
-			primaryFlag: "url",
-			primaryDesc: "Base/Wiki/record-share URL to resolve",
-			aliasFlags:  []string{"query"},
-		},
-		{
-			shortcut:    "+title-resolve",
-			definition:  BaseTitleResolve,
-			primaryFlag: "title",
-			primaryDesc: "Base title keyword",
-			aliasFlags:  []string{"query", "url"},
-		},
-	} {
-		t.Run(tc.shortcut, func(t *testing.T) {
-			parent := &cobra.Command{Use: "base"}
-			tc.definition.Mount(parent, &cmdutil.Factory{})
-			cmd := parent.Commands()[0]
-			primary := cmd.Flags().Lookup(tc.primaryFlag)
-			primaryUsage := ""
-			if primary != nil {
-				primaryUsage = primary.Usage
-			}
-			if primary == nil || !strings.Contains(primaryUsage, tc.primaryDesc) {
-				t.Fatalf("primary flag %q usage=%q", tc.primaryFlag, primaryUsage)
-			}
-			for _, aliasFlag := range tc.aliasFlags {
-				alias := cmd.Flags().Lookup(aliasFlag)
-				if alias == nil || !alias.Hidden {
-					t.Fatalf("alias flag %q should exist and be hidden: %#v", aliasFlag, alias)
-				}
-			}
-		})
-	}
-}
-
-func TestBaseTitleResolve(t *testing.T) {
-	t.Run("single result", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(titleResolveSearchStub([]interface{}{
-			map[string]interface{}{
-				"title_highlighted": "Sales <h>Pipeline</h>",
-				"result_meta": map[string]interface{}{
-					"doc_types":       "BITABLE",
-					"token":           "bas123",
-					"url":             "https://example.larkoffice.com/base/bas123",
-					"owner_name":      "Alice",
-					"update_time_iso": "2026-06-09T10:00:00+08:00",
-				},
-			},
-		}))
-
-		err := runShortcutWithAuthTypes(t, BaseTitleResolve, nil, []string{
-			"+title-resolve", "--title", "Pipeline", "--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-		data := decodeBaseEnvelope(t, stdout)
-		if data["title"] != "Sales Pipeline" || data["base_token"] != "bas123" || data["owner_name"] != "Alice" {
-			t.Fatalf("unexpected output: %#v", data)
-		}
-	})
-
-	t.Run("multiple results and filter non bitable", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(titleResolveSearchStub([]interface{}{
-			map[string]interface{}{
-				"title_highlighted": "Doc hit",
-				"result_meta":       map[string]interface{}{"doc_types": "DOCX", "token": "docx123"},
-			},
-			map[string]interface{}{
-				"title_highlighted": "Base <h>One</h>",
-				"result_meta":       map[string]interface{}{"doc_types": "BITABLE", "token": "bas1", "url": "https://example/base/bas1"},
-			},
-			map[string]interface{}{
-				"title_highlighted": "Base <h>Two</h>",
-				"result_meta":       map[string]interface{}{"doc_types": "BITABLE", "token": "bas2", "url": "https://example/base/bas2"},
-			},
-		}))
-
-		err := runShortcutWithAuthTypes(t, BaseTitleResolve, nil, []string{
-			"+title-resolve", "--url", "Base", "--as", "user",
-		}, factory, stdout)
-		if err != nil {
-			t.Fatalf("err=%v", err)
-		}
-		data := decodeBaseEnvelope(t, stdout)
-		candidates, _ := data["candidates"].([]interface{})
-		if len(candidates) != 2 {
-			t.Fatalf("candidates=%#v, want 2", data["candidates"])
-		}
-	})
-
-	t.Run("no results", func(t *testing.T) {
-		factory, stdout, reg := newExecuteFactory(t)
-		reg.Register(titleResolveSearchStub(nil))
-		err := runShortcutWithAuthTypes(t, BaseTitleResolve, nil, []string{
-			"+title-resolve", "--title", "missing", "--as", "user",
-		}, factory, stdout)
-		if err == nil || !strings.Contains(err.Error(), "No Base matched") {
-			t.Fatalf("err=%v, want no result validation", err)
-		}
-	})
-
-	t.Run("query too long", func(t *testing.T) {
-		factory, stdout, _ := newExecuteFactory(t)
-		err := runShortcutWithAuthTypes(t, BaseTitleResolve, nil, []string{
-			"+title-resolve", "--title", "codex record share resolve 20260616152113", "--as", "user",
-		}, factory, stdout)
-		if err == nil || !strings.Contains(err.Error(), "30 characters or fewer") {
-			t.Fatalf("err=%v, want query length validation", err)
-		}
-	})
-}
-
-func titleResolveSearchStub(items []interface{}) *httpmock.Stub {
-	if items == nil {
-		items = []interface{}{}
-	}
-	return &httpmock.Stub{
-		Method: "POST",
-		URL:    "/open-apis/search/v2/doc_wiki/search",
-		Body: map[string]interface{}{
-			"code": 0,
-			"data": map[string]interface{}{
-				"res_units": items,
-			},
-		},
-	}
-}
-
-func fieldListStub(baseToken, tableID string) *httpmock.Stub {
-	return &httpmock.Stub{
-		Method: "GET",
-		URL:    "/open-apis/base/v3/bases/" + baseToken + "/tables/" + tableID + "/fields",
-		Body: map[string]interface{}{
-			"code": 0,
-			"data": map[string]interface{}{
-				"total": 2,
-				"fields": []interface{}{
-					map[string]interface{}{"field_id": "fld_name", "field_name": "Name", "type": "text"},
-					map[string]interface{}{"field_id": "fld_status", "field_name": "Status", "type": "singleSelect"},
-				},
-			},
-		},
-	}
-}
-
-func recordBatchGetStub(baseToken, tableID, recordID string) *httpmock.Stub {
-	return &httpmock.Stub{
-		Method: "POST",
-		URL:    "/open-apis/base/v3/bases/" + baseToken + "/tables/" + tableID + "/records/batch_get",
-		Body: map[string]interface{}{
-			"code": 0,
-			"data": map[string]interface{}{
-				"record_id_list": []interface{}{recordID},
-				"field_id_list":  []interface{}{"fld_name", "fld_status"},
-				"fields":         []interface{}{"Name", "Status"},
-				"data":           []interface{}{[]interface{}{"Alice", "Done"}},
-			},
-		},
-	}
-}

shortcuts/base/base_shortcuts_test.go

@@ -155,7 +155,6 @@ func TestViewSetVisibleFieldsValidateHook(t *testing.T) {
 func TestShortcutsCatalog(t *testing.T) {
 	shortcuts := Shortcuts()
 	want := []string{
-		"+url-resolve", "+title-resolve",
 		"+base-block-list", "+base-block-create", "+base-block-move", "+base-block-rename", "+base-block-delete",
 		"+table-list", "+table-get", "+table-create", "+table-update", "+table-delete",
 		"+field-list", "+field-get", "+field-create", "+field-update", "+field-delete", "+field-search-options",

shortcuts/base/shortcuts.go

@@ -8,8 +8,6 @@ import "github.com/larksuite/cli/shortcuts/common"
 // Shortcuts returns all base shortcuts.
 func Shortcuts() []common.Shortcut {
 	return []common.Shortcut{
-		BaseURLResolve,
-		BaseTitleResolve,
 		BaseBaseBlockList,
 		BaseBaseBlockCreate,
 		BaseBaseBlockMove,

shortcuts/common/drive_media_upload_typed_test.go

@@ -199,7 +199,16 @@ func TestParseDriveMediaMultipartUploadSessionTypedValidatesResponseFields(t *te
 			t.Parallel()
 
 			_, err := parseDriveMediaMultipartUploadSessionTyped(tt.data)
-			requireProblem(t, err, errs.CategoryInternal, errs.SubtypeInvalidResponse, tt.wantText)
+			if err == nil || !strings.Contains(err.Error(), tt.wantText) {
+				t.Fatalf("err = %v, want substring %q", err, tt.wantText)
+			}
+			p, ok := errs.ProblemOf(err)
+			if !ok {
+				t.Fatalf("expected typed problem, got %T (%v)", err, err)
+			}
+			if p.Subtype != errs.SubtypeInvalidResponse {
+				t.Fatalf("subtype = %s, want invalid_response", p.Subtype)
+			}
 		})
 	}
 }

shortcuts/common/mcp_client_test.go

@@ -142,7 +142,9 @@ func TestNormalizeMCPToolResult(t *testing.T) {
 
 			got, err := normalizeMCPToolResult(tt.raw)
 			if tt.wantErr != "" {
-				requireProblem(t, err, errs.CategoryAPI, errs.SubtypeUnknown, tt.wantErr)
+				if err == nil || !strings.Contains(err.Error(), tt.wantErr) {
+					t.Fatalf("expected error containing %q, got %v", tt.wantErr, err)
+				}
 				return
 			}
 			if err != nil {

shortcuts/common/runner.go

@@ -49,7 +49,6 @@ type RuntimeContext struct {
 	apiClientFunc func() (*client.APIClient, error) // sync.OnceValues; initialized in newRuntimeContext
 	botInfoFunc   func() (*BotInfo, error)          // sync.OnceValues; lazy bot identity from /bot/v3/info
 	larkSDK       *lark.Client                      // eagerly initialized in mountDeclarative
-	stdinConsumed bool                              // set when an Input flag has consumed stdin (`-`); guards against a second flag also using `-` within the same call
 }
 
 // ── Identity ──
@@ -1030,6 +1029,7 @@ func stripUTF8BOM(s string) string {
 // resolveInputFlags resolves @file and - (stdin) for flags with Input sources.
 // Must be called before Validate/DryRun/Execute so that runtime.Str() returns resolved content.
 func resolveInputFlags(rctx *RuntimeContext, flags []Flag) error {
+	stdinUsed := false
 	for _, fl := range flags {
 		if len(fl.Input) == 0 {
 			continue
@@ -1049,14 +1049,11 @@ func resolveInputFlags(rctx *RuntimeContext, flags []Flag) error {
 				return ValidationErrorf("--%s does not support stdin (-)", fl.Name).
 					WithParam("--" + fl.Name)
 			}
-			// A process has a single stdin, so we reject a second Input flag
-			// trying to use `-` after the first one has already consumed it.
-			if rctx.stdinConsumed {
+			if stdinUsed {
 				return ValidationErrorf("--%s: stdin (-) can only be used by one flag", fl.Name).
-					WithParam("--"+fl.Name).
-					WithHint("a process has a single stdin, so only one flag per call may use '-'; pass the others as @file (e.g. --%s @/path/to/file)", fl.Name)
+					WithParam("--" + fl.Name)
 			}
-			rctx.stdinConsumed = true
+			stdinUsed = true
 			data, err := io.ReadAll(rctx.IO().In)
 			if err != nil {
 				return ValidationErrorf("--%s: failed to read from stdin: %v", fl.Name, err).
@@ -1169,13 +1166,7 @@ func registerShortcutFlagsWithContext(ctx context.Context, cmd *cobra.Command, f
 				hints = append(hints, "@file")
 			}
 			if slices.Contains(fl.Input, Stdin) {
-				// "- reads stdin" intentionally avoids implying each flag has
-				// its own stdin: a process has a single stdin, so at most one
-				// flag per call may use "-" (the rest must use @file). The old
-				// per-flag "- for stdin" wording led AI agents to write
-				// `--a - <x --b - <y`, where the second `<` silently clobbers
-				// the first and `--a` reads the wrong payload.
-				hints = append(hints, "- reads stdin (one flag per call; use @file for others)")
+				hints = append(hints, "- for stdin")
 			}
 			desc += " (supports " + strings.Join(hints, ", ") + ")"
 		}

shortcuts/common/runner_args_test.go

@@ -22,7 +22,6 @@ func TestRejectPositionalArgs_WithArgs(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for positional arg, got nil")
 	}
-	// rejectPositionalArgs returns a raw fmt.Errorf via cobra's PositionalArgs contract — not a typed envelope, message-substring assertion is intentional.
 	if !strings.Contains(err.Error(), "positional arguments are not supported") {
 		t.Errorf("expected positional args rejection message, got: %v", err)
 	}
@@ -40,7 +39,6 @@ func TestRejectPositionalArgs_MultipleArgs(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for multiple positional args, got nil")
 	}
-	// rejectPositionalArgs returns a raw fmt.Errorf via cobra's PositionalArgs contract — not a typed envelope, message-substring assertion is intentional.
 	if !strings.Contains(err.Error(), "positional arguments are not supported") {
 		t.Errorf("unexpected error message: %v", err)
 	}

shortcuts/common/runner_botinfo_test.go

@@ -171,7 +171,6 @@ func TestFetchBotInfo_APICodeNonZero(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for non-zero code")
 	}
-	// fetchBotInfo returns a raw fmt.Errorf, not a typed envelope — message-substring assertion is intentional.
 	if !strings.Contains(err.Error(), "[99991]") {
 		t.Errorf("error = %q, want substring [99991]", err.Error())
 	}
@@ -198,7 +197,6 @@ func TestFetchBotInfo_EmptyOpenID(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for empty open_id")
 	}
-	// fetchBotInfo returns a raw fmt.Errorf, not a typed envelope — message-substring assertion is intentional.
 	if !strings.Contains(err.Error(), "open_id is empty") {
 		t.Errorf("error = %q, want substring 'open_id is empty'", err.Error())
 	}
@@ -220,7 +218,6 @@ func TestFetchBotInfo_HTTP4xx(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for HTTP 403")
 	}
-	// fetchBotInfo returns a raw fmt.Errorf, not a typed envelope — message-substring assertion is intentional.
 	if !strings.Contains(err.Error(), "403") {
 		t.Errorf("error = %q, want substring '403'", err.Error())
 	}
@@ -241,7 +238,7 @@ func TestFetchBotInfo_InvalidJSON(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for invalid JSON")
 	}
-	// Error may come from SDK-level parse or our unmarshal wrapper — both are raw fmt.Errorf, not a typed envelope.
+	// Error may come from SDK-level parse or our unmarshal wrapper
 	if !strings.Contains(err.Error(), "unmarshal") && !strings.Contains(err.Error(), "invalid character") {
 		t.Errorf("error = %q, want JSON parse failure", err.Error())
 	}
@@ -282,7 +279,6 @@ func TestFetchBotInfo_CanBotFalse(t *testing.T) {
 	if info != nil {
 		t.Errorf("expected nil info, got %+v", info)
 	}
-	// fetchBotInfo returns a raw fmt.Errorf, not a typed envelope — message-substring assertion is intentional.
 	if !strings.Contains(err.Error(), "not available") {
 		t.Errorf("error = %q, want substring 'not available'", err.Error())
 	}
@@ -295,7 +291,6 @@ func TestBotInfo_NilFunc(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for nil botInfoFunc")
 	}
-	// BotInfo() returns a raw fmt.Errorf when botInfoFunc is nil, not a typed envelope — message-substring assertion is intentional.
 	if !strings.Contains(err.Error(), "not fully initialized") {
 		t.Errorf("unexpected error: %v", err)
 	}

shortcuts/common/runner_input_test.go

@@ -129,9 +129,9 @@ func TestResolveInputFlags_StdinNotSupported(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for stdin not supported")
 	}
-	vErr := assertValidationParam(t, err, "--data")
-	if !strings.Contains(vErr.Message, "does not support stdin") {
-		t.Errorf("unexpected error message: %q", vErr.Message)
+	assertValidationParam(t, err, "--data")
+	if !strings.Contains(err.Error(), "does not support stdin") {
+		t.Errorf("unexpected error: %v", err)
 	}
 }
 
@@ -143,9 +143,9 @@ func TestResolveInputFlags_FileNotSupported(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for file not supported")
 	}
-	vErr := assertValidationParam(t, err, "--data")
-	if !strings.Contains(vErr.Message, "does not support file input") {
-		t.Errorf("unexpected error message: %q", vErr.Message)
+	assertValidationParam(t, err, "--data")
+	if !strings.Contains(err.Error(), "does not support file input") {
+		t.Errorf("unexpected error: %v", err)
 	}
 }
 
@@ -160,9 +160,9 @@ func TestResolveInputFlags_FileNotFound(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for missing file")
 	}
-	vErr := assertValidationParam(t, err, "--markdown")
-	if !strings.Contains(vErr.Message, "cannot read file") {
-		t.Errorf("unexpected error message: %q", vErr.Message)
+	assertValidationParam(t, err, "--markdown")
+	if !strings.Contains(err.Error(), "cannot read file") {
+		t.Errorf("unexpected error: %v", err)
 	}
 }
 
@@ -174,9 +174,9 @@ func TestResolveInputFlags_EmptyFilePath(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for empty file path")
 	}
-	vErr := assertValidationParam(t, err, "--markdown")
-	if !strings.Contains(vErr.Message, "file path cannot be empty after @") {
-		t.Errorf("unexpected error message: %q", vErr.Message)
+	assertValidationParam(t, err, "--markdown")
+	if !strings.Contains(err.Error(), "file path cannot be empty after @") {
+		t.Errorf("unexpected error: %v", err)
 	}
 }
 
@@ -216,14 +216,9 @@ func TestResolveInputFlags_DuplicateStdin(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for duplicate stdin usage")
 	}
-	vErr := assertValidationParam(t, err, "--b")
-	if !strings.Contains(vErr.Message, "stdin (-) can only be used by one flag") {
-		t.Errorf("unexpected error message: %q", vErr.Message)
-	}
-	// The hint must steer an AI agent to the fix (@file for the extra flags),
-	// since `--a - <x --b - <y` is the exact misuse this guards against.
-	if !strings.Contains(vErr.Hint, "@file") {
-		t.Errorf("hint %q should mention @file as the fix", vErr.Hint)
+	assertValidationParam(t, err, "--b")
+	if !strings.Contains(err.Error(), "stdin (-) can only be used by one flag") {
+		t.Errorf("unexpected error: %v", err)
 	}
 }
 

shortcuts/common/runner_jq_test.go

@@ -186,7 +186,9 @@ func TestRunShortcut_JqAndFormatConflict(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for --jq + --format table conflict")
 	}
-	requireValidation(t, err, "mutually exclusive")
+	if !strings.Contains(err.Error(), "mutually exclusive") {
+		t.Errorf("expected 'mutually exclusive' error, got: %v", err)
+	}
 }
 
 func TestRunShortcut_JqInvalidExpression(t *testing.T) {
@@ -206,7 +208,9 @@ func TestRunShortcut_JqInvalidExpression(t *testing.T) {
 	if err == nil {
 		t.Fatal("expected error for invalid jq expression")
 	}
-	requireValidation(t, err, "invalid jq expression")
+	if !strings.Contains(err.Error(), "invalid jq expression") {
+		t.Errorf("expected 'invalid jq expression' error, got: %v", err)
+	}
 }
 
 func TestRunShortcut_JqRuntimeError_PropagatesError(t *testing.T) {

shortcuts/common/typed_error_assertions_test.go

@@ -1,50 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package common
-
-import (
-	"errors"
-	"strings"
-	"testing"
-
-	"github.com/larksuite/cli/errs"
-)
-
-// requireProblem asserts err carries a typed errs.Problem with the given
-// category and (optional) subtype, and that its message contains msgContains
-// (skip the message check by passing ""). Returns the Problem so callers can
-// drill into the typed envelope's category-specific fields (e.g. cast to
-// *errs.ValidationError to read .Param / .Params / .Cause).
-func requireProblem(t *testing.T, err error, wantCategory errs.Category, wantSubtype errs.Subtype, msgContains string) *errs.Problem {
-	t.Helper()
-	if err == nil {
-		t.Fatal("expected error, got nil")
-	}
-	p, ok := errs.ProblemOf(err)
-	if !ok {
-		t.Fatalf("expected typed error carrying errs.Problem, got %T: %v", err, err)
-	}
-	if p.Category != wantCategory {
-		t.Errorf("category = %q, want %q (err=%v)", p.Category, wantCategory, err)
-	}
-	if wantSubtype != "" && p.Subtype != wantSubtype {
-		t.Errorf("subtype = %q, want %q (err=%v)", p.Subtype, wantSubtype, err)
-	}
-	if msgContains != "" && !strings.Contains(p.Message, msgContains) {
-		t.Errorf("message = %q, want containing %q", p.Message, msgContains)
-	}
-	return p
-}
-
-// requireValidation is shorthand for CategoryValidation + SubtypeInvalidArgument.
-// Returns *errs.ValidationError so callers can also assert on .Param / .Params / .Cause.
-func requireValidation(t *testing.T, err error, msgContains string) *errs.ValidationError {
-	t.Helper()
-	requireProblem(t, err, errs.CategoryValidation, errs.SubtypeInvalidArgument, msgContains)
-	var ve *errs.ValidationError
-	if !errors.As(err, &ve) {
-		t.Fatalf("expected *errs.ValidationError, got %T: %v", err, err)
-	}
-	return ve
-}

shortcuts/doc/doc_resource_cover.go

@@ -55,7 +55,7 @@ var docCoverAllowedContentTypes = map[string]string{
 
 var DocResourceDownload = common.Shortcut{
 	Service:     "docs",
-	Command:     "+resource-download",
+	Command:     "resource-download",
 	Description: "Download a document resource (type=cover downloads the cover image content)",
 	Risk:        "read",
 	Scopes:      []string{"docx:document:readonly", "docs:document.media:download"},
@@ -154,7 +154,7 @@ var DocResourceDownload = common.Shortcut{
 
 var DocResourceUpdate = common.Shortcut{
 	Service:     "docs",
-	Command:     "+resource-update",
+	Command:     "resource-update",
 	Description: "Upload and update a document resource (type=cover)",
 	Risk:        "write",
 	Scopes:      []string{"docx:document:readonly", "docx:document:write_only", "docs:document.media:upload"},
@@ -256,7 +256,7 @@ var DocResourceUpdate = common.Shortcut{
 
 var DocResourceDelete = common.Shortcut{
 	Service:     "docs",
-	Command:     "+resource-delete",
+	Command:     "resource-delete",
 	Description: "Delete a document resource (type=cover is idempotent when empty)",
 	Risk:        "write",
 	Scopes:      []string{"docx:document:readonly", "docx:document:write_only"},

shortcuts/doc/doc_resource_cover_test.go

@@ -48,7 +48,7 @@ func TestDocResourceDownloadCoverDownloadsImageContent(t *testing.T) {
 	withDocsWorkingDir(t, tmpDir)
 
 	err := mountAndRunDocs(t, DocResourceDownload, []string{
-		"+resource-download",
+		"resource-download",
 		"--doc", documentID,
 		"--type", "cover",
 		"--output", "cover",
@@ -95,7 +95,7 @@ func TestDocResourceDownloadCoverEmptyReturnsErrorWithoutDownload(t *testing.T)
 	withDocsWorkingDir(t, tmpDir)
 
 	err := mountAndRunDocs(t, DocResourceDownload, []string{
-		"+resource-download",
+		"resource-download",
 		"--doc", documentID,
 		"--type", "cover",
 		"--output", "cover.png",
@@ -116,7 +116,7 @@ func TestDocResourceDeleteCoverEmptyIsIdempotent(t *testing.T) {
 	reg.Register(docCoverMetadataStub(documentID, map[string]interface{}{}))
 
 	err := mountAndRunDocs(t, DocResourceDelete, []string{
-		"+resource-delete",
+		"resource-delete",
 		"--doc", documentID,
 		"--type", "cover",
 		"--as", "bot",
@@ -146,7 +146,7 @@ func TestDocResourceDeleteCoverClearsExistingCover(t *testing.T) {
 	reg.Register(patchStub)
 
 	err := mountAndRunDocs(t, DocResourceDelete, []string{
-		"+resource-delete",
+		"resource-delete",
 		"--doc", documentID,
 		"--type", "cover",
 		"--as", "bot",
@@ -195,7 +195,7 @@ func TestDocResourceUpdateCoverUploadsFileAndReturnsFullTokenOnlyOnStdout(t *tes
 	reg.Register(patchStub)
 
 	err := mountAndRunDocs(t, DocResourceUpdate, []string{
-		"+resource-update",
+		"resource-update",
 		"--doc", documentID,
 		"--type", "cover",
 		"--file", "cover.png",
@@ -241,7 +241,7 @@ func TestDocResourceUpdateCoverRejectsMultipleSources(t *testing.T) {
 	f, _, _, _ := cmdutil.TestFactory(t, docsTestConfigWithAppID("docs-cover-source-validation-app"))
 
 	err := mountAndRunDocs(t, DocResourceUpdate, []string{
-		"+resource-update",
+		"resource-update",
 		"--doc", "doxcnCoverValidate1",
 		"--type", "cover",
 		"--file", "cover.png",
@@ -258,7 +258,7 @@ func TestDocResourceUpdateCoverRejectsMissingSource(t *testing.T) {
 	f, _, _, _ := cmdutil.TestFactory(t, docsTestConfigWithAppID("docs-cover-source-required-app"))
 
 	err := mountAndRunDocs(t, DocResourceUpdate, []string{
-		"+resource-update",
+		"resource-update",
 		"--doc", "doxcnCoverValidateRequired1",
 		"--type", "cover",
 		"--as", "bot",
@@ -273,7 +273,7 @@ func TestDocResourceUpdateCoverRejectsUnsafeURLSource(t *testing.T) {
 	f, _, _, _ := cmdutil.TestFactory(t, docsTestConfigWithAppID("docs-cover-url-validation-app"))
 
 	err := mountAndRunDocs(t, DocResourceUpdate, []string{
-		"+resource-update",
+		"resource-update",
 		"--doc", "doxcnCoverURLValidate1",
 		"--type", "cover",
 		"--url", "https://127.0.0.1/cover.png",
@@ -617,7 +617,7 @@ func TestDocShortcutsIncludeCoverResourceCommands(t *testing.T) {
 	for _, shortcut := range Shortcuts() {
 		got[shortcut.Command] = true
 	}
-	for _, want := range []string{"+resource-download", "+resource-update", "+resource-delete"} {
+	for _, want := range []string{"resource-download", "resource-update", "resource-delete"} {
 		if !got[want] {
 			t.Fatalf("Shortcuts() missing %s", want)
 		}

shortcuts/doc/docs_fetch_im_markdown.go

@@ -1,861 +0,0 @@
-// Copyright (c) 2026 Lark Technologies Pte. Ltd.
-// SPDX-License-Identifier: MIT
-
-package doc
-
-import (
-	"fmt"
-	"html"
-	"net/url"
-	"regexp"
-	"strings"
-	"unicode/utf8"
-)
-
-type imMarkdownContext struct {
-	baseURL         string
-	blockquoteDepth int
-}
-
-type imMarkdownHandleFunc func(segment, inner string, attrs map[string]string, imCtx imMarkdownContext) string
-
-type imMarkdownTagHandler struct {
-	closeRE *regexp.Regexp
-	handle  imMarkdownHandleFunc
-}
-
-func registerIMMarkdownHandler(tag string, handle imMarkdownHandleFunc) {
-	imMarkdownHandlers[tag] = imMarkdownTagHandler{
-		closeRE: regexp.MustCompile(`(?is)<(/?)` + regexp.QuoteMeta(tag) + `(?:\s[^<>]*?)?\s*/?>`),
-		handle:  handle,
-	}
-}
-
-var (
-	imMarkdownTagStartRE  = regexp.MustCompile(`(?s)<([A-Za-z][A-Za-z0-9:_-]*)(?:\s[^<>]*?)?\s*/?>`)
-	imMarkdownAttrRE      = regexp.MustCompile(`([A-Za-z_:][A-Za-z0-9_:.-]*)\s*=\s*(?:"([^"]*)"|'([^']*)')`)
-	imMarkdownRowTagRE    = regexp.MustCompile(`(?is)<(/?)tr\b[^>]*?\s*/?>`)
-	imMarkdownCellTagRE   = regexp.MustCompile(`(?is)<(/?)t[dh]\b[^>]*?\s*/?>`)
-	imMarkdownCellBreakRE = regexp.MustCompile(`(?i)<br\s*/?>`)
-	imMarkdownAnyTagRE    = regexp.MustCompile(`(?s)</?([A-Za-z][A-Za-z0-9:_-]*)(?:\s[^<>]*?)?>`)
-	imMarkdownLinkRE      = regexp.MustCompile(`(?is)<a\b[^>]*\bhref=(?:"([^"]*)"|'([^']*)')[^>]*>(.*?)</a>`)
-	imMarkdownCodeBlockRE = regexp.MustCompile(`(?is)^\s*<code(?:\s[^<>]*?)?>(.*?)</code>\s*$`)
-	imMarkdownLiOpenRE    = regexp.MustCompile(`(?is)<li(?:\s[^<>]*?)?>`)
-	imMarkdownLiCloseRE   = regexp.MustCompile(`(?is)<(/?)li(?:\s[^<>]*?)?\s*/?>`)
-)
-
-var imMarkdownHandlers = map[string]imMarkdownTagHandler{}
-
-func init() {
-	registerIMMarkdownHandler("title", handleIMMarkdownTitle)
-	for level := 1; level <= 9; level++ {
-		registerIMMarkdownHandler(fmt.Sprintf("h%d", level), handleIMMarkdownHeading(level))
-	}
-	registerIMMarkdownHandler("p", handleIMMarkdownParagraph)
-	registerIMMarkdownHandler("ul", handleIMMarkdownUnorderedList)
-	registerIMMarkdownHandler("ol", handleIMMarkdownOrderedList)
-	registerIMMarkdownHandler("li", handleIMMarkdownListItem)
-	registerIMMarkdownHandler("callout", handleIMMarkdownCallout)
-	registerIMMarkdownHandler("blockquote", handleIMMarkdownBlockquote)
-	registerIMMarkdownHandler("grid", handleIMMarkdownPassthroughContainer)
-	registerIMMarkdownHandler("column", handleIMMarkdownColumn)
-	registerIMMarkdownHandler("table", handleIMMarkdownTable)
-	registerIMMarkdownHandler("colgroup", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("col", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("pre", handleIMMarkdownPre)
-	registerIMMarkdownHandler("code", handleIMMarkdownCode)
-	registerIMMarkdownHandler("latex", handleIMMarkdownLatex)
-	registerIMMarkdownHandler("hr", handleIMMarkdownHorizontalRule)
-	registerIMMarkdownHandler("img", handleIMMarkdownImage)
-	registerIMMarkdownHandler("figure", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("source", handleIMMarkdownSource)
-	registerIMMarkdownHandler("button", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("time", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("whiteboard", handleIMMarkdownInlineCode)
-	registerIMMarkdownHandler("sheet", handleIMMarkdownSheet)
-	registerIMMarkdownHandler("task", handleIMMarkdownConditionalResourceLabel("任务", "task-id", "guid", "token", "id"))
-	registerIMMarkdownHandler("chat_card", handleIMMarkdownConditionalResourceLabel("群聊卡片", "chat-id", "chat_id", "id"))
-	registerIMMarkdownHandler("bitable", handleIMMarkdownResourceLabel("多维表格"))
-	registerIMMarkdownHandler("base_refer", handleIMMarkdownResourceLabel("多维表格"))
-	registerIMMarkdownHandler("okr", handleIMMarkdownResourceLabel("OKR"))
-	registerIMMarkdownHandler("poll", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("agenda", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("folder_manager", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("wiki_catalog", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("wiki_recent_update", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("chart_refer_host_perm", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("synced_reference", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("synced-source", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("mindnote", handleIMMarkdownDiscard)
-	registerIMMarkdownHandler("bookmark", handleIMMarkdownBookmark)
-	registerIMMarkdownHandler("cite", handleIMMarkdownCite)
-	registerIMMarkdownHandler("b", handleIMMarkdownStrong)
-	registerIMMarkdownHandler("em", handleIMMarkdownEmphasis)
-	registerIMMarkdownHandler("del", handleIMMarkdownDelete)
-	registerIMMarkdownHandler("u", handleIMMarkdownPlainInline)
-	registerIMMarkdownHandler("span", handleIMMarkdownPlainInline)
-	registerIMMarkdownHandler("a", handleIMMarkdownAnchor)
-}
-
-func isIMMarkdownFetch(runtime interface{ Str(string) string }) bool {
-	return strings.TrimSpace(runtime.Str("doc-format")) == "im-markdown"
-}
-
-func applyFetchIMMarkdown(data map[string]interface{}, docInput string) {
-	doc, ok := data["document"].(map[string]interface{})
-	if !ok {
-		return
-	}
-	content, ok := doc["content"].(string)
-	if !ok {
-		return
-	}
-	doc["content"] = convertToIMMarkdown(content, newIMMarkdownContext(docInput))
-}
-
-func newIMMarkdownContext(docInput string) imMarkdownContext {
-	base := "https://larkoffice.com"
-	raw := strings.TrimSpace(docInput)
-	if extracted, ok := imMarkdownBaseURLFromInput(raw); ok {
-		base = extracted
-	}
-	return imMarkdownContext{baseURL: base}
-}
-
-func (c imMarkdownContext) withBlockquote() imMarkdownContext {
-	c.blockquoteDepth++
-	return c
-}
-
-func (c imMarkdownContext) inBlockquote() bool {
-	return c.blockquoteDepth > 0
-}
-
-// imMarkdownBaseURLFromInput keeps the tenant host from --doc when it is a URL
-// so generated doc/sheet links point back to the same tenant. parseDocumentRef
-// intentionally strips host information, so it cannot serve this formatting path.
-func imMarkdownBaseURLFromInput(raw string) (string, bool) {
-	if raw == "" {
-		return "", false
-	}
-	if u, err := url.Parse(raw); err == nil && u.Scheme != "" && u.Host != "" {
-		return u.Scheme + "://" + u.Host, true
-	}
-	for _, marker := range []string{"/docx/", "/wiki/", "/doc/"} {
-		idx := strings.Index(raw, marker)
-		if idx <= 0 {
-			continue
-		}
-		candidate := strings.Trim(raw[:idx], "/")
-		if candidate == "" {
-			continue
-		}
-		if u, err := url.Parse(candidate); err == nil && u.Scheme != "" && u.Host != "" {
-			return u.Scheme + "://" + u.Host, true
-		}
-		if u, err := url.Parse("https://" + candidate); err == nil && u.Host != "" && strings.Contains(u.Host, ".") {
-			return "https://" + u.Host, true
-		}
-	}
-	return "", false
-}
-
-func convertToIMMarkdown(content string, imCtx imMarkdownContext) string {
-	var out strings.Builder
-	for offset := 0; offset < len(content); {
-		// Scan only to the next XML-like opening tag. Plain Markdown text between
-		// registered tags is copied unchanged, so ordinary Markdown is not re-parsed.
-		loc := imMarkdownTagStartRE.FindStringSubmatchIndex(content[offset:])
-		if loc == nil {
-			out.WriteString(content[offset:])
-			break
-		}
-		start := offset + loc[0]
-		openEnd := offset + loc[1]
-		tag := strings.ToLower(content[offset+loc[2] : offset+loc[3]])
-		handler, ok := imMarkdownHandlers[tag]
-		if !ok {
-			// Unknown tags are left intact. im-markdown only downgrades tags with
-			// explicit handlers so future server output does not get guessed at.
-			out.WriteString(content[offset:openEnd])
-			offset = openEnd
-			continue
-		}
-
-		out.WriteString(content[offset:start])
-		opening := content[start:openEnd]
-		attrs := parseIMMarkdownAttrs(opening)
-		if isSelfClosingIMMarkdownTag(opening) {
-			out.WriteString(handler.handle(opening, "", attrs, imCtx))
-			offset = openEnd
-			continue
-		}
-
-		// Use the handler's precompiled close regexp to find the matching end tag.
-		// Depth tracking keeps nested same-name containers paired correctly.
-		closeStart, closeEnd, found := findIMMarkdownClosingTag(content, openEnd, handler)
-		if !found {
-			// Malformed or truncated fragments are preserved as-is from the opening
-			// tag onward; do not drop content when the XML-ish structure is incomplete.
-			out.WriteString(content[start:])
-			break
-		}
-		segment := content[start:closeEnd]
-		inner := content[openEnd:closeStart]
-		out.WriteString(handler.handle(segment, inner, attrs, imCtx))
-		offset = closeEnd
-	}
-	return out.String()
-}
-
-func findIMMarkdownClosingTag(content string, from int, handler imMarkdownTagHandler) (int, int, bool) {
-	depth := 1
-	for _, loc := range handler.closeRE.FindAllStringSubmatchIndex(content[from:], -1) {
-		start := from + loc[0]
-		end := from + loc[1]
-		token := content[start:end]
-		if loc[2] >= 0 && content[from+loc[2]:from+loc[3]] == "/" {
-			depth--
-			if depth == 0 {
-				return start, end, true
-			}
-			continue
-		}
-		if !isSelfClosingIMMarkdownTag(token) {
-			depth++
-		}
-	}
-	return 0, 0, false
-}
-
-func parseIMMarkdownAttrs(opening string) map[string]string {
-	attrs := map[string]string{}
-	for _, match := range imMarkdownAttrRE.FindAllStringSubmatch(opening, -1) {
-		value := match[2]
-		if value == "" {
-			value = match[3]
-		}
-		attrs[strings.ToLower(match[1])] = html.UnescapeString(value)
-	}
-	return attrs
-}
-
-func isSelfClosingIMMarkdownTag(tag string) bool {
-	return strings.HasSuffix(strings.TrimSpace(tag), "/>")
-}
-
-func handleIMMarkdownTitle(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	text := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	if text == "" {
-		return ""
-	}
-	return "# " + text
-}
-
-func handleIMMarkdownHeading(level int) imMarkdownHandleFunc {
-	return func(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-		text := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-		if text == "" {
-			return ""
-		}
-		markdownLevel := level
-		if markdownLevel > 6 {
-			markdownLevel = 6
-		}
-		return strings.Repeat("#", markdownLevel) + " " + text
-	}
-}
-
-func handleIMMarkdownParagraph(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	if body == "" {
-		return ""
-	}
-	if imCtx.inBlockquote() {
-		return body + "\n"
-	}
-	return body
-}
-
-func handleIMMarkdownUnorderedList(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	return convertIMMarkdownListItems(inner, false, imCtx)
-}
-
-func handleIMMarkdownOrderedList(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	return convertIMMarkdownListItems(inner, true, imCtx)
-}
-
-func handleIMMarkdownListItem(_ string, inner string, attrs map[string]string, imCtx imMarkdownContext) string {
-	prefix := "-"
-	if seq := strings.TrimSpace(attrs["seq"]); seq != "" && seq != "auto" {
-		prefix = strings.TrimSuffix(seq, ".") + "."
-	}
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	if body == "" {
-		return ""
-	}
-	return prefix + " " + indentIMMarkdownListContinuation(body) + "\n"
-}
-
-func handleIMMarkdownCallout(_ string, inner string, attrs map[string]string, imCtx imMarkdownContext) string {
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	emoji := strings.TrimSpace(attrs["emoji"])
-	if emoji != "" {
-		if body == "" {
-			body = emoji
-		} else {
-			body = emoji + " " + body
-		}
-	}
-	if body == "" {
-		return "---\n---"
-	}
-	return fmt.Sprintf("---\n%s\n---", body)
-}
-
-func handleIMMarkdownBlockquote(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx.withBlockquote()))
-	if body == "" {
-		return ""
-	}
-	lines := strings.Split(body, "\n")
-	for i, line := range lines {
-		if strings.TrimSpace(line) == "" {
-			lines[i] = ">"
-			continue
-		}
-		lines[i] = "> " + line
-	}
-	return strings.Join(lines, "\n")
-}
-
-func handleIMMarkdownPassthroughContainer(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	return strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-}
-
-func handleIMMarkdownColumn(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	if body == "" {
-		return ""
-	}
-	return body + "\n"
-}
-
-func handleIMMarkdownDiscard(_ string, _ string, _ map[string]string, _ imMarkdownContext) string {
-	return ""
-}
-
-func handleIMMarkdownInlineCode(segment string, _ string, _ map[string]string, _ imMarkdownContext) string {
-	return imMarkdownInlineCode(segment)
-}
-
-func handleIMMarkdownPre(_ string, inner string, attrs map[string]string, _ imMarkdownContext) string {
-	lang := strings.TrimSpace(attrs["lang"])
-	code := strings.TrimSpace(inner)
-	if match := imMarkdownCodeBlockRE.FindStringSubmatch(code); match != nil {
-		code = match[1]
-	}
-	return imMarkdownFencedCode(html.UnescapeString(code), lang)
-}
-
-func handleIMMarkdownCode(_ string, inner string, _ map[string]string, _ imMarkdownContext) string {
-	return imMarkdownInlineCode(markdownPlainText(inner))
-}
-
-func handleIMMarkdownLatex(_ string, inner string, _ map[string]string, _ imMarkdownContext) string {
-	expr := strings.TrimSpace(markdownPlainText(inner))
-	if expr == "" {
-		return ""
-	}
-	return "$" + strings.ReplaceAll(expr, "$", `\$`) + "$"
-}
-
-func handleIMMarkdownHorizontalRule(_ string, _ string, _ map[string]string, _ imMarkdownContext) string {
-	return "---"
-}
-
-func handleIMMarkdownImage(_ string, _ string, attrs map[string]string, _ imMarkdownContext) string {
-	href := firstNonEmpty(attrs["href"], attrs["src"], attrs["url"])
-	if href == "" {
-		return ""
-	}
-	alt := firstNonEmpty(attrs["alt"], attrs["name"], attrs["title"])
-	return fmt.Sprintf("![%s](%s)", escapeMarkdownLinkText(alt), escapeMarkdownLinkDestination(href))
-}
-
-func handleIMMarkdownSource(_ string, _ string, attrs map[string]string, _ imMarkdownContext) string {
-	name := strings.TrimSpace(attrs["name"])
-	if name == "" {
-		return ""
-	}
-	return imMarkdownInlineCode(name)
-}
-
-func handleIMMarkdownResourceLabel(label string) imMarkdownHandleFunc {
-	return func(_ string, _ string, _ map[string]string, _ imMarkdownContext) string {
-		return imMarkdownInlineCode(label)
-	}
-}
-
-func handleIMMarkdownConditionalResourceLabel(label string, attrNames ...string) imMarkdownHandleFunc {
-	return func(_ string, _ string, attrs map[string]string, _ imMarkdownContext) string {
-		for _, attrName := range attrNames {
-			if strings.TrimSpace(attrs[attrName]) != "" {
-				return imMarkdownInlineCode(label)
-			}
-		}
-		return ""
-	}
-}
-
-func handleIMMarkdownSheet(segment string, _ string, attrs map[string]string, imCtx imMarkdownContext) string {
-	token := strings.TrimSpace(attrs["token"])
-	if token == "" {
-		return imMarkdownInlineCode(segment)
-	}
-	label := "sheet"
-	if sheetID := strings.TrimSpace(attrs["sheet-id"]); sheetID != "" {
-		label = "sheet " + sheetID
-	}
-	return markdownLink(label, strings.TrimRight(imCtx.baseURL, "/")+"/sheets/"+token)
-}
-
-func handleIMMarkdownBookmark(segment string, inner string, attrs map[string]string, imCtx imMarkdownContext) string {
-	href := strings.TrimSpace(attrs["href"])
-	name := firstNonEmpty(attrs["name"], attrs["title"], markdownLinkLabelText(convertToIMMarkdown(inner, imCtx)), href)
-	if href == "" {
-		return name
-	}
-	return markdownLink(name, href)
-}
-
-func handleIMMarkdownStrong(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	if body == "" {
-		return ""
-	}
-	return "**" + body + "**"
-}
-
-func handleIMMarkdownEmphasis(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	if body == "" {
-		return ""
-	}
-	return "*" + body + "*"
-}
-
-func handleIMMarkdownDelete(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	body := strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-	if body == "" {
-		return ""
-	}
-	return "~~" + body + "~~"
-}
-
-func handleIMMarkdownPlainInline(_ string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	return strings.TrimSpace(convertToIMMarkdown(inner, imCtx))
-}
-
-func handleIMMarkdownAnchor(_ string, inner string, attrs map[string]string, imCtx imMarkdownContext) string {
-	href := strings.TrimSpace(attrs["href"])
-	text := firstNonEmpty(markdownLinkLabelText(convertToIMMarkdown(inner, imCtx)), attrs["name"], attrs["title"], href)
-	if href == "" {
-		return text
-	}
-	return markdownLink(text, href)
-}
-
-func handleIMMarkdownCite(segment string, inner string, attrs map[string]string, imCtx imMarkdownContext) string {
-	switch strings.ToLower(strings.TrimSpace(attrs["type"])) {
-	case "user":
-		userID := firstNonEmpty(attrs["user-id"], attrs["open-id"], attrs["id"])
-		name := firstNonEmpty(attrs["user-name"], attrs["name"], markdownPlainText(inner), userID)
-		if userID == "" {
-			return name
-		}
-		return fmt.Sprintf(`<at user_id="%s">%s</at>`, html.EscapeString(userID), html.EscapeString(name))
-	case "doc":
-		title := firstNonEmpty(attrs["title"], attrs["name"], attrs["doc-id"], "document")
-		if href := firstNonEmpty(attrs["href"], attrs["url"]); href != "" {
-			return markdownLink(title, href)
-		}
-		docID := firstNonEmpty(attrs["doc-id"], attrs["token"])
-		if docID == "" {
-			return imMarkdownInlineCode(segment)
-		}
-		fileType := strings.Trim(strings.ToLower(firstNonEmpty(attrs["file-type"], "docx")), "/")
-		return markdownLink(title, strings.TrimRight(imCtx.baseURL, "/")+"/"+fileType+"/"+docID)
-	case "citation":
-		if text, href, ok := extractIMMarkdownInnerLink(inner); ok {
-			return markdownLink(text, href)
-		}
-		if href := firstNonEmpty(attrs["href"], attrs["url"]); href != "" {
-			return markdownLink(firstNonEmpty(attrs["title"], attrs["name"], href), href)
-		}
-		return markdownPlainText(convertToIMMarkdown(inner, imCtx))
-	default:
-		return imMarkdownInlineCode(segment)
-	}
-}
-
-func handleIMMarkdownTable(segment string, inner string, _ map[string]string, imCtx imMarkdownContext) string {
-	// Rows and cells are matched with tag-depth tracking instead of non-greedy
-	// regex captures. A table nested inside a cell can contain its own </tr> and
-	// </td>; treating those as the outer row/cell boundary corrupts the table.
-	rowBodies := extractIMMarkdownElementBodies(inner, imMarkdownRowTagRE)
-	if len(rowBodies) == 0 {
-		return imMarkdownInlineCode(segment)
-	}
-
-	rows := make([][]string, 0, len(rowBodies))
-	for _, rowBody := range rowBodies {
-		cellBodies := extractIMMarkdownElementBodies(rowBody, imMarkdownCellTagRE)
-		if len(cellBodies) == 0 {
-			continue
-		}
-		row := make([]string, 0, len(cellBodies))
-		for _, cellBody := range cellBodies {
-			row = append(row, normalizeIMMarkdownTableCell(convertToIMMarkdown(cellBody, imCtx)))
-		}
-		rows = append(rows, row)
-	}
-	if len(rows) == 0 {
-		return imMarkdownInlineCode(segment)
-	}
-
-	cols := 0
-	for _, row := range rows {
-		if len(row) > cols {
-			cols = len(row)
-		}
-	}
-	var out strings.Builder
-	writeIMMarkdownTableRow(&out, padIMMarkdownTableRow(rows[0], cols))
-	separator := make([]string, cols)
-	for i := range separator {
-		separator[i] = "-"
-	}
-	writeIMMarkdownTableRow(&out, separator)
-	for _, row := range rows[1:] {
-		writeIMMarkdownTableRow(&out, padIMMarkdownTableRow(row, cols))
-	}
-	return strings.TrimRight(out.String(), "\n")
-}
-
-// extractIMMarkdownElementBodies returns the inner content of each top-level
-// element matched by tagRE. tagRE must expose the optional closing slash as its
-// first capture group, matching the row/cell regexes above.
-func extractIMMarkdownElementBodies(content string, tagRE *regexp.Regexp) []string {
-	var bodies []string
-	for offset := 0; offset < len(content); {
-		loc := tagRE.FindStringSubmatchIndex(content[offset:])
-		if loc == nil {
-			break
-		}
-		openStart := offset + loc[0]
-		openEnd := offset + loc[1]
-		opening := content[openStart:openEnd]
-		if loc[2] >= 0 && content[offset+loc[2]:offset+loc[3]] == "/" {
-			offset = openEnd
-			continue
-		}
-		if isSelfClosingIMMarkdownTag(opening) {
-			bodies = append(bodies, "")
-			offset = openEnd
-			continue
-		}
-		closeStart, closeEnd, found := findIMMarkdownElementClosingTag(content, openEnd, tagRE)
-		if !found {
-			break
-		}
-		bodies = append(bodies, content[openEnd:closeStart])
-		offset = closeEnd
-	}
-	return bodies
-}
-
-func findIMMarkdownElementClosingTag(content string, from int, tagRE *regexp.Regexp) (int, int, bool) {
-	depth := 1
-	for _, loc := range tagRE.FindAllStringSubmatchIndex(content[from:], -1) {
-		start := from + loc[0]
-		end := from + loc[1]
-		token := content[start:end]
-		if loc[2] >= 0 && content[from+loc[2]:from+loc[3]] == "/" {
-			depth--
-			if depth == 0 {
-				return start, end, true
-			}
-			continue
-		}
-		if !isSelfClosingIMMarkdownTag(token) {
-			depth++
-		}
-	}
-	return 0, 0, false
-}
-
-func normalizeIMMarkdownTableCell(cell string) string {
-	const brPlaceholder = "\x00BR\x00"
-	cell = imMarkdownCellBreakRE.ReplaceAllString(cell, brPlaceholder)
-	cell = imMarkdownAnyTagRE.ReplaceAllStringFunc(cell, func(tag string) string {
-		name := strings.ToLower(strings.TrimPrefix(imMarkdownAnyTagRE.FindStringSubmatch(tag)[1], "/"))
-		if name == "at" {
-			return tag
-		}
-		return ""
-	})
-	cell = html.UnescapeString(cell)
-	cell = strings.ReplaceAll(cell, brPlaceholder, "<br>")
-	cell = strings.ReplaceAll(cell, "  \n", "<br>")
-	cell = strings.ReplaceAll(cell, "\n", "<br>")
-	cell = strings.ReplaceAll(cell, "|", `\|`)
-	lines := strings.Fields(cell)
-	if len(lines) == 0 {
-		return ""
-	}
-	return strings.Join(lines, " ")
-}
-
-func writeIMMarkdownTableRow(out *strings.Builder, row []string) {
-	out.WriteString("| ")
-	out.WriteString(strings.Join(row, " | "))
-	out.WriteString(" |\n")
-}
-
-func padIMMarkdownTableRow(row []string, cols int) []string {
-	if len(row) >= cols {
-		return row
-	}
-	padded := make([]string, cols)
-	copy(padded, row)
-	return padded
-}
-
-func convertIMMarkdownListItems(inner string, ordered bool, imCtx imMarkdownContext) string {
-	var out strings.Builder
-	for offset, index := 0, 1; offset < len(inner); {
-		loc := imMarkdownLiOpenRE.FindStringIndex(inner[offset:])
-		if loc == nil {
-			break
-		}
-		openStart := offset + loc[0]
-		openEnd := offset + loc[1]
-		opening := inner[openStart:openEnd]
-		closeStart, closeEnd, found := findIMMarkdownListItemClosingTag(inner, openEnd)
-		if !found {
-			break
-		}
-		body := strings.TrimSpace(convertToIMMarkdown(inner[openEnd:closeStart], imCtx))
-		if body != "" {
-			prefix := "-"
-			if ordered {
-				attrs := parseIMMarkdownAttrs(opening)
-				if seq := strings.TrimSpace(attrs["seq"]); seq != "" && seq != "auto" {
-					prefix = strings.TrimSuffix(seq, ".") + "."
-				} else {
-					prefix = fmt.Sprintf("%d.", index)
-				}
-				index++
-			}
-			out.WriteString(prefix)
-			out.WriteString(" ")
-			out.WriteString(indentIMMarkdownListContinuation(body))
-			out.WriteString("\n")
-		}
-		offset = closeEnd
-	}
-	return strings.TrimRight(out.String(), "\n")
-}
-
-func findIMMarkdownListItemClosingTag(content string, from int) (int, int, bool) {
-	depth := 1
-	for _, loc := range imMarkdownLiCloseRE.FindAllStringSubmatchIndex(content[from:], -1) {
-		start := from + loc[0]
-		end := from + loc[1]
-		token := content[start:end]
-		if loc[2] >= 0 && content[from+loc[2]:from+loc[3]] == "/" {
-			depth--
-			if depth == 0 {
-				return start, end, true
-			}
-			continue
-		}
-		if !isSelfClosingIMMarkdownTag(token) {
-			depth++
-		}
-	}
-	return 0, 0, false
-}
-
-func indentIMMarkdownListContinuation(body string) string {
-	return strings.ReplaceAll(body, "\n", "\n  ")
-}
-
-func extractIMMarkdownInnerLink(inner string) (string, string, bool) {
-	match := imMarkdownLinkRE.FindStringSubmatch(inner)
-	if match == nil {
-		return "", "", false
-	}
-	href := match[1]
-	if href == "" {
-		href = match[2]
-	}
-	text := strings.TrimSpace(markdownPlainText(match[3]))
-	if text == "" {
-		text = href
-	}
-	return text, html.UnescapeString(href), true
-}
-
-func markdownPlainText(s string) string {
-	s = imMarkdownCellBreakRE.ReplaceAllString(s, "\n")
-	s = imMarkdownAnyTagRE.ReplaceAllString(s, "")
-	return strings.TrimSpace(html.UnescapeString(s))
-}
-
-func markdownLinkLabelText(s string) string {
-	text := markdownPlainText(s)
-	if !strings.Contains(text, "---") {
-		return text
-	}
-	lines := strings.Split(text, "\n")
-	kept := lines[:0]
-	for _, line := range lines {
-		if strings.TrimSpace(line) == "---" {
-			continue
-		}
-		kept = append(kept, line)
-	}
-	return strings.TrimSpace(strings.Join(kept, "\n"))
-}
-
-func markdownLink(text, href string) string {
-	cleanHref := strings.TrimSpace(href)
-	return fmt.Sprintf("[%s](%s)", escapeMarkdownLinkText(firstNonEmpty(text, cleanHref)), escapeMarkdownLinkDestination(cleanHref))
-}
-
-func escapeMarkdownLinkText(text string) string {
-	text = strings.ReplaceAll(text, `\`, `\\`)
-	text = strings.ReplaceAll(text, `[`, `\[`)
-	text = strings.ReplaceAll(text, `]`, `\]`)
-	return text
-}
-
-func escapeMarkdownLinkDestination(href string) string {
-	// Lark/Feishu IM Markdown does not reliably parse raw spaces or parentheses
-	// inside (...). Keep URL delimiters like :/?#&= intact, but percent-encode
-	// characters that can terminate or split the Markdown link destination.
-	var out strings.Builder
-	out.Grow(len(href))
-	for i := 0; i < len(href); {
-		if href[i] == '%' {
-			if i+2 < len(href) && isHexDigit(href[i+1]) && isHexDigit(href[i+2]) {
-				out.WriteString(href[i : i+3])
-				i += 3
-			} else {
-				writePercentEncodedByte(&out, href[i])
-				i++
-			}
-			continue
-		}
-		if href[i] < utf8.RuneSelf {
-			if shouldPercentEncodeIMMarkdownURLByte(href[i]) {
-				writePercentEncodedByte(&out, href[i])
-			} else {
-				out.WriteByte(href[i])
-			}
-			i++
-			continue
-		}
-		r, size := utf8.DecodeRuneInString(href[i:])
-		if r == utf8.RuneError && size == 1 {
-			writePercentEncodedByte(&out, href[i])
-			i++
-			continue
-		}
-		for _, b := range []byte(href[i : i+size]) {
-			writePercentEncodedByte(&out, b)
-		}
-		i += size
-	}
-	return out.String()
-}
-
-func shouldPercentEncodeIMMarkdownURLByte(b byte) bool {
-	if b <= ' ' || b >= 0x7f {
-		return true
-	}
-	switch b {
-	case '(', ')', '<', '>', '"', '\\', '^', '`', '{', '|', '}':
-		return true
-	default:
-		return false
-	}
-}
-
-func writePercentEncodedByte(out *strings.Builder, b byte) {
-	const hex = "0123456789ABCDEF"
-	out.WriteByte('%')
-	out.WriteByte(hex[b>>4])
-	out.WriteByte(hex[b&0x0f])
-}
-
-func isHexDigit(b byte) bool {
-	return ('0' <= b && b <= '9') || ('a' <= b && b <= 'f') || ('A' <= b && b <= 'F')
-}
-
-func imMarkdownInlineCode(s string) string {
-	maxRun := 0
-	run := 0
-	for _, r := range s {
-		if r == '`' {
-			run++
-			if run > maxRun {
-				maxRun = run
-			}
-			continue
-		}
-		run = 0
-	}
-	fence := strings.Repeat("`", maxRun+1)
-	if strings.HasPrefix(s, "`") || strings.HasSuffix(s, "`") {
-		return fence + " " + s + " " + fence
-	}
-	return fence + s + fence
-}
-
-func imMarkdownFencedCode(code, lang string) string {
-	maxRun := 0
-	for _, line := range strings.Split(code, "\n") {
-		if run := leadingBacktickRun(line); run > maxRun {
-			maxRun = run
-		}
-	}
-	fenceLen := maxRun + 1
-	if fenceLen < 3 {
-		fenceLen = 3
-	}
-	fence := strings.Repeat("`", fenceLen)
-	return fence + strings.TrimSpace(lang) + "\n" + strings.Trim(code, "\n") + "\n" + fence
-}
-
-func leadingBacktickRun(s string) int {
-	run := 0
-	for _, r := range s {
-		if r != '`' {
-			break
-		}
-		run++
-	}
-	return run
-}
-
-func firstNonEmpty(values ...string) string {
-	for _, value := range values {
-		if strings.TrimSpace(value) != "" {
-			return strings.TrimSpace(value)
-		}
-	}
-	return ""
-}

shortcuts/doc/docs_fetch_v2.go

@@ -17,7 +17,7 @@ import (
 // v2FetchFlags returns the flag definitions for the v2 (OpenAPI) fetch path.
 func v2FetchFlags() []common.Flag {
 	return []common.Flag{
-		{Name: "doc-format", Desc: "output content format; xml keeps DocxXML structure and optional block ids, markdown is plain export, im-markdown downgrades residual DocxXML fragments for IM messages", Default: "xml", Enum: []string{"xml", "markdown", "im-markdown"}},
+		{Name: "doc-format", Desc: "output content format; xml keeps DocxXML structure and optional block ids, markdown is plain export", Default: "xml", Enum: []string{"xml", "markdown"}},
 		{Name: "detail", Desc: "detail level; simple for reading, with-ids for block references, full for styles and edit metadata", Default: "simple", Enum: []string{"simple", "with-ids", "full"}},
 		{Name: "lang", Desc: "user cite display language, e.g. en-US, zh-CN, ja-JP"},
 		{Name: "revision-id", Desc: "document revision id; -1 means latest", Type: "int", Default: "-1"},
@@ -72,9 +72,6 @@ func executeFetchV2(_ context.Context, runtime *common.RuntimeContext) error {
 	if warning := addFetchDetailDowngradeWarning(runtime, data); warning != "" && runtime.Format == "pretty" {
 		fmt.Fprintf(runtime.IO().ErrOut, "warning: %s\n", warning)
 	}
-	if isIMMarkdownFetch(runtime) {
-		applyFetchIMMarkdown(data, runtime.Str("doc"))
-	}
 
 	runtime.OutFormatRaw(data, nil, func(w io.Writer) {
 		if doc, ok := data["document"].(map[string]interface{}); ok {
@@ -88,7 +85,7 @@ func executeFetchV2(_ context.Context, runtime *common.RuntimeContext) error {
 
 func buildFetchBody(runtime *common.RuntimeContext) map[string]interface{} {
 	body := map[string]interface{}{
-		"format": effectiveFetchFormat(runtime),
+		"format": runtime.Str("doc-format"),
 	}
 	if v := runtime.Int("revision-id"); v > 0 {
 		body["revision_id"] = v
@@ -125,14 +122,6 @@ func buildFetchBody(runtime *common.RuntimeContext) map[string]interface{} {
 	return body
 }
 
-func effectiveFetchFormat(runtime *common.RuntimeContext) string {
-	format := strings.TrimSpace(runtime.Str("doc-format"))
-	if format == "im-markdown" {
-		return "markdown"
-	}
-	return format
-}
-
 func resolveFetchLang(runtime *common.RuntimeContext) string {
 	if runtime.Changed("lang") {
 		return strings.TrimSpace(runtime.Str("lang"))

shortcuts/doc/docs_fetch_v2_test.go

@@ -6,12 +6,9 @@ package doc
 import (
 	"context"
 	"encoding/json"
-	"errors"
-	"reflect"
 	"strings"
 	"testing"
 
-	"github.com/larksuite/cli/errs"
 	"github.com/larksuite/cli/internal/cmdutil"
 	"github.com/larksuite/cli/internal/core"
 	"github.com/larksuite/cli/internal/httpmock"
@@ -107,369 +104,6 @@ func TestBuildFetchBodyExplicitBlankLangOmitsLang(t *testing.T) {
 	}
 }
 
-func TestBuildFetchBodyIncludesRevisionAndFullDetail(t *testing.T) {
-	t.Parallel()
-
-	runtime := newFetchBodyTestRuntime(context.Background())
-	mustSetFetchFlag(t, runtime, "revision-id", "42")
-	mustSetFetchFlag(t, runtime, "detail", "full")
-
-	body := buildFetchBody(runtime)
-	if got := body["revision_id"]; got != 42 {
-		t.Fatalf("revision_id = %#v, want 42", got)
-	}
-	exportOption, _ := body["export_option"].(map[string]interface{})
-	want := map[string]interface{}{
-		"export_block_id":        true,
-		"export_style_attrs":     true,
-		"export_cite_extra_data": true,
-	}
-	if !reflect.DeepEqual(exportOption, want) {
-		t.Fatalf("export_option = %#v, want %#v", exportOption, want)
-	}
-}
-
-func TestBuildFetchBodyIncludesWithIDsDetail(t *testing.T) {
-	t.Parallel()
-
-	runtime := newFetchBodyTestRuntime(context.Background())
-	mustSetFetchFlag(t, runtime, "detail", "with-ids")
-
-	body := buildFetchBody(runtime)
-	exportOption, _ := body["export_option"].(map[string]interface{})
-	want := map[string]interface{}{
-		"export_block_id": true,
-	}
-	if !reflect.DeepEqual(exportOption, want) {
-		t.Fatalf("export_option = %#v, want %#v", exportOption, want)
-	}
-}
-
-func TestBuildFetchBodyIncludesReadOption(t *testing.T) {
-	t.Parallel()
-
-	runtime := newFetchBodyTestRuntime(context.Background())
-	mustSetFetchFlag(t, runtime, "scope", "section")
-	mustSetFetchFlag(t, runtime, "start-block-id", "blk_heading")
-
-	body := buildFetchBody(runtime)
-	want := map[string]interface{}{
-		"read_mode":      "section",
-		"start_block_id": "blk_heading",
-	}
-	if got := body["read_option"]; !reflect.DeepEqual(got, want) {
-		t.Fatalf("read_option = %#v, want %#v", got, want)
-	}
-}
-
-func TestBuildReadOptionModes(t *testing.T) {
-	t.Parallel()
-
-	tests := []struct {
-		name     string
-		setFlags map[string]string
-		want     map[string]interface{}
-	}{
-		{
-			name: "full omits read option",
-			setFlags: map[string]string{
-				"scope": "full",
-			},
-			want: nil,
-		},
-		{
-			name: "outline with max depth",
-			setFlags: map[string]string{
-				"scope":     "outline",
-				"max-depth": "3",
-			},
-			want: map[string]interface{}{
-				"read_mode": "outline",
-				"max_depth": "3",
-			},
-		},
-		{
-			name: "range with block ids and context",
-			setFlags: map[string]string{
-				"scope":          "range",
-				"start-block-id": "blk_start",
-				"end-block-id":   "blk_end",
-				"context-before": "2",
-				"context-after":  "1",
-				"max-depth":      "0",
-			},
-			want: map[string]interface{}{
-				"read_mode":      "range",
-				"start_block_id": "blk_start",
-				"end_block_id":   "blk_end",
-				"context_before": "2",
-				"context_after":  "1",
-				"max_depth":      "0",
-			},
-		},
-		{
-			name: "keyword with query",
-			setFlags: map[string]string{
-				"scope":          "keyword",
-				"keyword":        "foo|bar",
-				"context-before": "1",
-			},
-			want: map[string]interface{}{
-				"read_mode":      "keyword",
-				"keyword":        "foo|bar",
-				"context_before": "1",
-			},
-		},
-		{
-			name: "section keeps unlimited depth omitted",
-			setFlags: map[string]string{
-				"scope":          "section",
-				"start-block-id": "blk_heading",
-				"max-depth":      "-1",
-			},
-			want: map[string]interface{}{
-				"read_mode":      "section",
-				"start_block_id": "blk_heading",
-			},
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			t.Parallel()
-
-			runtime := newFetchBodyTestRuntime(context.Background())
-			for name, value := range tt.setFlags {
-				mustSetFetchFlag(t, runtime, name, value)
-			}
-
-			if got := buildReadOption(runtime); !reflect.DeepEqual(got, tt.want) {
-				t.Fatalf("buildReadOption() = %#v, want %#v", got, tt.want)
-			}
-		})
-	}
-}
-
-func TestValidateReadModeFlagsRejectsInvalidScopeOptions(t *testing.T) {
-	t.Parallel()
-
-	tests := []struct {
-		name       string
-		setFlags   map[string]string
-		wantParam  string
-		wantParams []string
-	}{
-		{
-			name: "negative context before",
-			setFlags: map[string]string{
-				"scope":          "range",
-				"start-block-id": "blk_start",
-				"context-before": "-1",
-			},
-			wantParam: "--context-before",
-		},
-		{
-			name: "negative context after",
-			setFlags: map[string]string{
-				"scope":          "range",
-				"start-block-id": "blk_start",
-				"context-after":  "-1",
-			},
-			wantParam: "--context-after",
-		},
-		{
-			name: "max depth below unlimited sentinel",
-			setFlags: map[string]string{
-				"scope":          "range",
-				"start-block-id": "blk_start",
-				"max-depth":      "-2",
-			},
-			wantParam: "--max-depth",
-		},
-		{
-			name: "range needs boundary",
-			setFlags: map[string]string{
-				"scope": "range",
-			},
-			wantParams: []string{
-				"--start-block-id",
-				"--end-block-id",
-			},
-		},
-		{
-			name: "keyword needs keyword",
-			setFlags: map[string]string{
-				"scope": "keyword",
-			},
-			wantParam: "--keyword",
-		},
-		{
-			name: "section needs start block",
-			setFlags: map[string]string{
-				"scope": "section",
-			},
-			wantParam: "--start-block-id",
-		},
-		{
-			name: "unknown scope",
-			setFlags: map[string]string{
-				"scope": "bad",
-			},
-			wantParam: "--scope",
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			t.Parallel()
-
-			runtime := newFetchBodyTestRuntime(context.Background())
-			for name, value := range tt.setFlags {
-				mustSetFetchFlag(t, runtime, name, value)
-			}
-
-			err := validateReadModeFlags(runtime)
-			if err == nil {
-				t.Fatal("validateReadModeFlags() succeeded, want error")
-			}
-			assertValidationContract(t, err, errs.SubtypeInvalidArgument, tt.wantParam, tt.wantParams...)
-		})
-	}
-}
-
-func TestValidateReadModeFlagsAcceptsValidScopeOptions(t *testing.T) {
-	t.Parallel()
-
-	tests := []struct {
-		name     string
-		setFlags map[string]string
-	}{
-		{
-			name: "outline",
-			setFlags: map[string]string{
-				"scope": "outline",
-			},
-		},
-		{
-			name: "range with end block",
-			setFlags: map[string]string{
-				"scope":        "range",
-				"end-block-id": "blk_end",
-			},
-		},
-		{
-			name: "keyword with keyword",
-			setFlags: map[string]string{
-				"scope":   "keyword",
-				"keyword": "bug|缺陷",
-			},
-		},
-		{
-			name: "section with start block",
-			setFlags: map[string]string{
-				"scope":          "section",
-				"start-block-id": "blk_heading",
-			},
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			t.Parallel()
-
-			runtime := newFetchBodyTestRuntime(context.Background())
-			for name, value := range tt.setFlags {
-				mustSetFetchFlag(t, runtime, name, value)
-			}
-
-			if err := validateReadModeFlags(runtime); err != nil {
-				t.Fatalf("validateReadModeFlags() error = %v", err)
-			}
-		})
-	}
-}
-
-func TestValidateFetchV2RejectsInvalidDocAndScope(t *testing.T) {
-	t.Parallel()
-
-	tests := []struct {
-		name      string
-		setFlags  map[string]string
-		wantParam string
-	}{
-		{
-			name: "invalid doc",
-			setFlags: map[string]string{
-				"doc": "https://example.com/sheets/sht_token",
-			},
-			wantParam: "--doc",
-		},
-		{
-			name: "invalid scope",
-			setFlags: map[string]string{
-				"scope": "bad",
-			},
-			wantParam: "--scope",
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			t.Parallel()
-
-			runtime := newFetchShortcutTestRuntime(t, "", tt.setFlags)
-			err := validateFetchV2(context.Background(), runtime)
-			if err == nil {
-				t.Fatal("validateFetchV2() succeeded, want error")
-			}
-			assertValidationContract(t, err, errs.SubtypeInvalidArgument, tt.wantParam)
-		})
-	}
-}
-
-func TestAddFetchDetailDowngradeWarningNoops(t *testing.T) {
-	t.Parallel()
-
-	tests := []struct {
-		name     string
-		setFlags map[string]string
-	}{
-		{
-			name: "xml format",
-			setFlags: map[string]string{
-				"doc-format": "xml",
-				"detail":     "full",
-			},
-		},
-		{
-			name: "markdown simple detail",
-			setFlags: map[string]string{
-				"doc-format": "markdown",
-				"detail":     "simple",
-			},
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			t.Parallel()
-
-			runtime := newFetchBodyTestRuntime(context.Background())
-			for name, value := range tt.setFlags {
-				mustSetFetchFlag(t, runtime, name, value)
-			}
-
-			data := map[string]interface{}{}
-			if got := addFetchDetailDowngradeWarning(runtime, data); got != "" {
-				t.Fatalf("warning = %q, want empty", got)
-			}
-			if _, ok := data["warnings"]; ok {
-				t.Fatalf("unexpected warnings: %#v", data["warnings"])
-			}
-		})
-	}
-}
-
 func TestDocsFetchDryRunDefaultsToV2Endpoint(t *testing.T) {
 	t.Parallel()
 
@@ -507,54 +141,36 @@ func TestDocsFetchAPIVersionV1StillUsesV2Endpoint(t *testing.T) {
 	}
 }
 
-func TestDocsFetchIMMarkdownRequestsMarkdownFromAPI(t *testing.T) {
-	t.Parallel()
-
-	runtime := newFetchShortcutTestRuntime(t, "", map[string]string{
-		"doc-format": "im-markdown",
-	})
-	if err := validateFetchV2(context.Background(), runtime); err != nil {
-		t.Fatalf("validateFetchV2() error = %v", err)
-	}
-
-	dry := decodeDocDryRun(t, DocsFetch.DryRun(context.Background(), runtime))
-	if got, want := dry.API[0].Body["format"], "markdown"; got != want {
-		t.Fatalf("dry-run format = %#v, want %q", got, want)
-	}
-}
-
 func TestDocsFetchMarkdownDetailDowngradesToSimple(t *testing.T) {
 	t.Parallel()
 
-	for _, format := range []string{"markdown", "im-markdown"} {
-		for _, detail := range []string{"with-ids", "full"} {
-			t.Run(format+"/"+detail, func(t *testing.T) {
-				t.Parallel()
+	for _, detail := range []string{"with-ids", "full"} {
+		t.Run(detail, func(t *testing.T) {
+			t.Parallel()
 
-				runtime := newFetchShortcutTestRuntime(t, "", map[string]string{
-					"doc-format": format,
-					"detail":     detail,
-				})
-				if err := validateFetchV2(context.Background(), runtime); err != nil {
-					t.Fatalf("validateFetchV2() error = %v", err)
-				}
-
-				dry := decodeDocDryRun(t, DocsFetch.DryRun(context.Background(), runtime))
-				exportOption, _ := dry.API[0].Body["export_option"].(map[string]interface{})
-				if exportOption == nil {
-					t.Fatalf("missing export_option: %#v", dry.API[0].Body)
-				}
-				if got := exportOption["export_block_id"]; got != false {
-					t.Fatalf("export_block_id = %#v, want false after markdown detail downgrade", got)
-				}
-				if got := exportOption["export_style_attrs"]; got != false {
-					t.Fatalf("export_style_attrs = %#v, want false after markdown detail downgrade", got)
-				}
-				if got := exportOption["export_cite_extra_data"]; got != false {
-					t.Fatalf("export_cite_extra_data = %#v, want false after markdown detail downgrade", got)
-				}
+			runtime := newFetchShortcutTestRuntime(t, "", map[string]string{
+				"doc-format": "markdown",
+				"detail":     detail,
 			})
-		}
+			if err := validateFetchV2(context.Background(), runtime); err != nil {
+				t.Fatalf("validateFetchV2() error = %v", err)
+			}
+
+			dry := decodeDocDryRun(t, DocsFetch.DryRun(context.Background(), runtime))
+			exportOption, _ := dry.API[0].Body["export_option"].(map[string]interface{})
+			if exportOption == nil {
+				t.Fatalf("missing export_option: %#v", dry.API[0].Body)
+			}
+			if got := exportOption["export_block_id"]; got != false {
+				t.Fatalf("export_block_id = %#v, want false after markdown detail downgrade", got)
+			}
+			if got := exportOption["export_style_attrs"]; got != false {
+				t.Fatalf("export_style_attrs = %#v, want false after markdown detail downgrade", got)
+			}
+			if got := exportOption["export_cite_extra_data"]; got != false {
+				t.Fatalf("export_cite_extra_data = %#v, want false after markdown detail downgrade", got)
+			}
+		})
 	}
 }
 
@@ -645,107 +261,6 @@ func TestDocsFetchMarkdownDetailDowngradeWarnsInPrettyOutput(t *testing.T) {
 	}
 }
 
-func TestDocsFetchV2ReturnsAPIError(t *testing.T) {
-	t.Setenv("LARKSUITE_CLI_CONFIG_DIR", t.TempDir())
-
-	f, stdout, _, reg := cmdutil.TestFactory(t, docsTestConfigWithAppID("docs-fetch-api-error"))
-	reg.Register(&httpmock.Stub{
-		Method: "POST",
-		URL:    "/open-apis/docs_ai/v1/documents/doxcnFetchAPIError/fetch",
-		Body: map[string]interface{}{
-			"code": 999999,
-			"msg":  "fetch failed",
-		},
-	})
-
-	err := mountAndRunDocs(t, DocsFetch, []string{
-		"+fetch",
-		"--doc", "doxcnFetchAPIError",
-		"--as", "bot",
-	}, f, stdout)
-	if err == nil {
-		t.Fatal("mountAndRunDocs() succeeded, want API error")
-	}
-	var apiErr *errs.APIError
-	if !errors.As(err, &apiErr) {
-		t.Fatalf("error type = %T, want *errs.APIError (%v)", err, err)
-	}
-	p, ok := errs.ProblemOf(err)
-	if !ok {
-		t.Fatalf("ProblemOf() ok = false for %T (%v)", err, err)
-	}
-	if p.Category != errs.CategoryAPI {
-		t.Errorf("category = %q, want %q", p.Category, errs.CategoryAPI)
-	}
-	if p.Subtype != errs.SubtypeUnknown {
-		t.Errorf("subtype = %q, want %q", p.Subtype, errs.SubtypeUnknown)
-	}
-	if p.Code != 999999 {
-		t.Errorf("code = %d, want 999999", p.Code)
-	}
-	if p.Message != "fetch failed" {
-		t.Errorf("message = %q, want %q", p.Message, "fetch failed")
-	}
-	if cause := errors.Unwrap(err); cause != nil {
-		t.Fatalf("unexpected wrapped cause for API response error: %T %v", cause, cause)
-	}
-}
-
-func TestDocsFetchIMMarkdownConvertsContentInJSONOutput(t *testing.T) {
-	t.Setenv("LARKSUITE_CLI_CONFIG_DIR", t.TempDir())
-
-	f, stdout, _, reg := cmdutil.TestFactory(t, docsTestConfigWithAppID("docs-fetch-im-markdown"))
-	reg.Register(&httpmock.Stub{
-		Method: "POST",
-		URL:    "/open-apis/docs_ai/v1/documents/doxcnFetchIMMarkdown/fetch",
-		Body: map[string]interface{}{
-			"code": 0,
-			"msg":  "ok",
-			"data": map[string]interface{}{
-				"document": map[string]interface{}{
-					"document_id": "doxcnFetchIMMarkdown",
-					"revision_id": float64(1),
-					"content": strings.Join([]string{
-						`<title>Doc Title</title>`,
-						`<callout emoji="💡">Read **this**.</callout>`,
-						`<bookmark name="Example" href="https://example.com"></bookmark>`,
-					}, "\n\n"),
-				},
-			},
-		},
-	})
-
-	err := mountAndRunDocs(t, DocsFetch, []string{
-		"+fetch",
-		"--doc", "doxcnFetchIMMarkdown",
-		"--doc-format", "im-markdown",
-		"--as", "bot",
-	}, f, stdout)
-	if err != nil {
-		t.Fatalf("unexpected error: %v", err)
-	}
-
-	var envelope map[string]interface{}
-	if err := json.Unmarshal(stdout.Bytes(), &envelope); err != nil {
-		t.Fatalf("decode output: %v\nraw=%s", err, stdout.String())
-	}
-	data, _ := envelope["data"].(map[string]interface{})
-	doc, _ := data["document"].(map[string]interface{})
-	content, _ := doc["content"].(string)
-	for _, want := range []string{
-		"# Doc Title",
-		"---\n💡 Read **this**.\n---",
-		"[Example](https://example.com)",
-	} {
-		if !strings.Contains(content, want) {
-			t.Fatalf("converted content missing %q:\n%s", want, content)
-		}
-	}
-	if strings.Contains(content, "<title>") || strings.Contains(content, "<callout") || strings.Contains(content, "<bookmark") {
-		t.Fatalf("converted content still contains downgraded XML tags:\n%s", content)
-	}
-}
-
 func TestDocsFetchRejectsLegacyFlags(t *testing.T) {
 	tests := []struct {
 		name     string
@@ -776,7 +291,6 @@ func TestDocsFetchRejectsLegacyFlags(t *testing.T) {
 			if err == nil {
 				t.Fatal("expected v2-only validation error")
 			}
-			assertValidationContract(t, err, errs.SubtypeInvalidArgument, "--offset")
 			for _, want := range tt.want {
 				if !strings.Contains(err.Error(), want) {
 					t.Fatalf("error missing %q: %v", want, err)
@@ -802,14 +316,6 @@ func newFetchBodyTestRuntime(ctx context.Context) *common.RuntimeContext {
 	return common.TestNewRuntimeContextWithCtx(ctx, cmd, nil)
 }
 
-func mustSetFetchFlag(t *testing.T, runtime *common.RuntimeContext, name, value string) {
-	t.Helper()
-
-	if err := runtime.Cmd.Flags().Set(name, value); err != nil {
-		t.Fatalf("set %s: %v", name, err)
-	}
-}
-
 func newFetchShortcutTestRuntime(t *testing.T, apiVersion string, setFlags map[string]string) *common.RuntimeContext {
 	t.Helper()
 

shortcuts/drive/drive_export.go

@@ -5,7 +5,6 @@ package drive
 
 import (
 	"context"
-	"errors"
 	"fmt"
 	"path/filepath"
 	"strings"
@@ -16,24 +15,6 @@ import (
 	"github.com/larksuite/cli/shortcuts/common"
 )
 
-// wrapExportContextErr converts a context cancellation / deadline error into a
-// typed errs.NetworkError so the cobra layer sees a typed envelope (with cause
-// preserved for errors.Is) instead of an untyped context.Canceled /
-// context.DeadlineExceeded escaping as a plain string. CR-flagged hole on the
-// poll loop: returning ctx.Err() directly bypassed the typed-error contract.
-func wrapExportContextErr(err error) error {
-	if err == nil {
-		return nil
-	}
-	subtype := errs.SubtypeNetworkTransport
-	msg := "drive +export polling cancelled: %s"
-	if errors.Is(err, context.DeadlineExceeded) {
-		subtype = errs.SubtypeNetworkTimeout
-		msg = "drive +export polling deadline exceeded: %s"
-	}
-	return errs.NewNetworkError(subtype, msg, err).WithCause(err)
-}
-
 // DriveExport exports Drive-native documents to local files and falls back to
 // a follow-up command when the async export task does not finish in time.
 var DriveExport = common.Shortcut{
@@ -59,305 +40,236 @@ var DriveExport = common.Shortcut{
 		{Name: "overwrite", Type: "bool", Desc: "overwrite existing output file"},
 	},
 	Validate: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		return validateExport(exportParamsFromFlags(runtime))
+		return validateDriveExportSpec(driveExportSpec{
+			Token:         runtime.Str("token"),
+			DocType:       runtime.Str("doc-type"),
+			FileExtension: runtime.Str("file-extension"),
+			SubID:         runtime.Str("sub-id"),
+			OnlySchema:    runtime.Bool("only-schema"),
+		})
 	},
 	DryRun: func(ctx context.Context, runtime *common.RuntimeContext) *common.DryRunAPI {
-		return PlanExportDryRun(runtime, exportParamsFromFlags(runtime))
-	},
-	Execute: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		return RunExport(ctx, runtime, exportParamsFromFlags(runtime))
-	},
-}
+		spec := driveExportSpec{
+			Token:         runtime.Str("token"),
+			DocType:       runtime.Str("doc-type"),
+			FileExtension: runtime.Str("file-extension"),
+			SubID:         runtime.Str("sub-id"),
+			OnlySchema:    runtime.Bool("only-schema"),
+		}
+		// Markdown export is a special case: docx markdown comes from the V2
+		// docs_ai fetch API directly instead of the Drive export task API.
+		if spec.FileExtension == "markdown" {
+			apiPath := fmt.Sprintf("/open-apis/docs_ai/v1/documents/%s/fetch", validate.EncodePathSegment(spec.Token))
+			dr := common.NewDryRunAPI().
+				Desc("2-step orchestration: fetch docx markdown -> write local file").
+				POST(apiPath).
+				Body(map[string]interface{}{
+					"format": "markdown",
+				}).
+				Set("output_dir", runtime.Str("output-dir"))
+			if name := strings.TrimSpace(runtime.Str("file-name")); name != "" {
+				dr.Set("file_name", ensureExportFileExtension(sanitizeExportFileName(name, spec.Token), spec.FileExtension))
+			}
+			return dr
+		}
 
-// ExportParams holds the user-facing inputs for an export flow, decoupled from
-// cobra flags so other command groups (e.g. sheets +workbook-export) can reuse
-// the drive export implementation. An empty OutputDir means "create the export
-// task and poll, but do not download" — callers that only need the ready file
-// token / status get it back without writing a local file.
-type ExportParams struct {
-	Token         string
-	DocType       string
-	FileExtension string
-	SubID         string
-	OnlySchema    bool
-	OutputDir     string
-	FileName      string
-	Overwrite     bool
-}
+		body := map[string]interface{}{
+			"token":          spec.Token,
+			"type":           spec.DocType,
+			"file_extension": spec.FileExtension,
+		}
+		if strings.TrimSpace(spec.SubID) != "" {
+			body["sub_id"] = spec.SubID
+		}
+		if spec.OnlySchema {
+			body["only_schema"] = true
+		}
 
-func (p ExportParams) spec() driveExportSpec {
-	return driveExportSpec{
-		Token:         p.Token,
-		DocType:       p.DocType,
-		FileExtension: p.FileExtension,
-		SubID:         p.SubID,
-		OnlySchema:    p.OnlySchema,
-	}
-}
-
-// exportParamsFromFlags reads the standard drive +export flag set.
-func exportParamsFromFlags(runtime *common.RuntimeContext) ExportParams {
-	// drive +export always downloads; an empty --output-dir historically means
-	// the current directory (saveContentToOutputDir maps "" -> "."), so normalize
-	// it here to keep behavior identical and stay off the export-only ("" => skip
-	// download) path that only sheets +workbook-export uses.
-	outputDir := runtime.Str("output-dir")
-	if outputDir == "" {
-		outputDir = "."
-	}
-	return ExportParams{
-		Token:         runtime.Str("token"),
-		DocType:       runtime.Str("doc-type"),
-		FileExtension: runtime.Str("file-extension"),
-		SubID:         runtime.Str("sub-id"),
-		OnlySchema:    runtime.Bool("only-schema"),
-		OutputDir:     outputDir,
-		FileName:      strings.TrimSpace(runtime.Str("file-name")),
-		Overwrite:     runtime.Bool("overwrite"),
-	}
-}
-
-// validateExport runs the CLI-level export constraint checks. Unexported because
-// only drive +export's Validate consumes it directly; sheets +workbook-export
-// reuses RunExport / PlanExportDryRun but inlines its own (sheet-specific)
-// validation, so there is no cross-package call site to keep exported.
-func validateExport(p ExportParams) error {
-	return validateDriveExportSpec(p.spec())
-}
-
-// PlanExportDryRun builds the dry-run plan for an export without performing I/O.
-func PlanExportDryRun(runtime *common.RuntimeContext, p ExportParams) *common.DryRunAPI {
-	spec := p.spec()
-	// Markdown export is a special case: docx markdown comes from the V2
-	// docs_ai fetch API directly instead of the Drive export task API.
-	if spec.FileExtension == "markdown" {
-		apiPath := fmt.Sprintf("/open-apis/docs_ai/v1/documents/%s/fetch", validate.EncodePathSegment(spec.Token))
 		dr := common.NewDryRunAPI().
-			Desc("2-step orchestration: fetch docx markdown -> write local file").
-			POST(apiPath).
-			Body(map[string]interface{}{
-				"format": "markdown",
-			}).
-			Set("output_dir", p.OutputDir)
-		if name := strings.TrimSpace(p.FileName); name != "" {
+			Desc("3-step orchestration: create export task -> limited polling -> download file").
+			POST("/open-apis/drive/v1/export_tasks").
+			Body(body).
+			Set("output_dir", runtime.Str("output-dir"))
+		if name := strings.TrimSpace(runtime.Str("file-name")); name != "" {
 			dr.Set("file_name", ensureExportFileExtension(sanitizeExportFileName(name, spec.Token), spec.FileExtension))
 		}
 		return dr
-	}
-
-	body := map[string]interface{}{
-		"token":          spec.Token,
-		"type":           spec.DocType,
-		"file_extension": spec.FileExtension,
-	}
-	if strings.TrimSpace(spec.SubID) != "" {
-		body["sub_id"] = spec.SubID
-	}
-	if spec.OnlySchema {
-		body["only_schema"] = true
-	}
-
-	dr := common.NewDryRunAPI().
-		Desc("3-step orchestration: create export task -> limited polling -> download file").
-		POST("/open-apis/drive/v1/export_tasks").
-		Body(body).
-		Set("output_dir", p.OutputDir)
-	if name := strings.TrimSpace(p.FileName); name != "" {
-		dr.Set("file_name", ensureExportFileExtension(sanitizeExportFileName(name, spec.Token), spec.FileExtension))
-	}
-	return dr
-}
-
-// RunExport drives create export task -> bounded poll -> optional download. It
-// is the shared core behind both drive +export and sheets +workbook-export. An
-// empty p.OutputDir skips the download step and returns the ready file token.
-func RunExport(ctx context.Context, runtime *common.RuntimeContext, p ExportParams) error {
-	spec := p.spec()
-	outputDir := p.OutputDir
-	preferredFileName := strings.TrimSpace(p.FileName)
-	overwrite := p.Overwrite
-
-	// Markdown export bypasses the async export task and writes the fetched
-	// markdown content directly to disk. Uses the V2 docs_ai fetch API for
-	// higher-quality Lark-flavored Markdown output.
-	if spec.FileExtension == "markdown" {
-		fmt.Fprintf(runtime.IO().ErrOut, "Exporting docx as markdown: %s\n", common.MaskToken(spec.Token))
-		apiPath := fmt.Sprintf("/open-apis/docs_ai/v1/documents/%s/fetch", validate.EncodePathSegment(spec.Token))
-		data, err := runtime.CallAPITyped(
-			"POST",
-			apiPath,
-			nil,
-			map[string]interface{}{
-				"format": "markdown",
-			},
-		)
-		if err != nil {
-			return err
+	},
+	Execute: func(ctx context.Context, runtime *common.RuntimeContext) error {
+		spec := driveExportSpec{
+			Token:         runtime.Str("token"),
+			DocType:       runtime.Str("doc-type"),
+			FileExtension: runtime.Str("file-extension"),
+			SubID:         runtime.Str("sub-id"),
+			OnlySchema:    runtime.Bool("only-schema"),
 		}
+		outputDir := runtime.Str("output-dir")
+		preferredFileName := strings.TrimSpace(runtime.Str("file-name"))
+		overwrite := runtime.Bool("overwrite")
 
-		// Extract content from the V2 response: data.document.content
-		doc, ok := data["document"].(map[string]interface{})
-		if !ok {
-			return errs.NewInternalError(errs.SubtypeInvalidResponse, "invalid markdown fetch response: missing document object")
-		}
-		content, ok := doc["content"].(string)
-		if !ok {
-			return errs.NewInternalError(errs.SubtypeInvalidResponse, "invalid markdown fetch response: missing document.content")
-		}
-
-		fileName := preferredFileName
-		if fileName == "" {
-			// Prefer the remote title for the exported file name, but still fall
-			// back to the token if metadata is empty.
-			title, err := common.FetchDriveMetaTitle(runtime, spec.Token, spec.DocType)
+		// Markdown export bypasses the async export task and writes the fetched
+		// markdown content directly to disk. Uses the V2 docs_ai fetch API for
+		// higher-quality Lark-flavored Markdown output.
+		if spec.FileExtension == "markdown" {
+			fmt.Fprintf(runtime.IO().ErrOut, "Exporting docx as markdown: %s\n", common.MaskToken(spec.Token))
+			apiPath := fmt.Sprintf("/open-apis/docs_ai/v1/documents/%s/fetch", validate.EncodePathSegment(spec.Token))
+			data, err := runtime.CallAPITyped(
+				"POST",
+				apiPath,
+				nil,
+				map[string]interface{}{
+					"format": "markdown",
+				},
+			)
 			if err != nil {
-				fmt.Fprintf(runtime.IO().ErrOut, "Title lookup failed, using token as filename: %v\n", err)
-				title = spec.Token
+				return err
 			}
-			fileName = title
-		}
-		fileName = ensureExportFileExtension(sanitizeExportFileName(fileName, spec.Token), spec.FileExtension)
-		savedPath, err := saveContentToOutputDir(runtime.FileIO(), outputDir, fileName, []byte(content), overwrite)
-		if err != nil {
-			return err
-		}
 
-		runtime.Out(map[string]interface{}{
-			"token":          spec.Token,
-			"doc_type":       spec.DocType,
-			"file_extension": spec.FileExtension,
-			"file_name":      filepath.Base(savedPath),
-			"saved_path":     savedPath,
-			"size_bytes":     len(content),
-		}, nil)
-		return nil
-	}
-
-	ticket, err := createDriveExportTask(runtime, spec)
-	if err != nil {
-		return err
-	}
-	fmt.Fprintf(runtime.IO().ErrOut, "Created export task: %s\n", ticket)
-
-	var lastStatus driveExportStatus
-	var lastPollErr error
-	hasObservedStatus := false
-	// Keep the command responsive by polling for a bounded window. If the task
-	// is still running after that, return a resume command instead of blocking.
-	for attempt := 1; attempt <= driveExportPollAttempts; attempt++ {
-		if attempt > 1 {
-			select {
-			case <-ctx.Done():
-				return wrapExportContextErr(ctx.Err())
-			case <-time.After(driveExportPollInterval):
+			// Extract content from the V2 response: data.document.content
+			doc, ok := data["document"].(map[string]interface{})
+			if !ok {
+				return errs.NewInternalError(errs.SubtypeInvalidResponse, "invalid markdown fetch response: missing document object")
 			}
-		}
-		if err := ctx.Err(); err != nil {
-			return wrapExportContextErr(err)
-		}
-
-		status, err := getDriveExportStatus(runtime, spec.Token, ticket)
-		if err != nil {
-			// Treat polling failures as transient so short-lived backend hiccups
-			// do not immediately fail an otherwise healthy export task.
-			lastPollErr = err
-			fmt.Fprintf(runtime.IO().ErrOut, "Export status attempt %d/%d failed: %v\n", attempt, driveExportPollAttempts, err)
-			continue
-		}
-		lastStatus = status
-		hasObservedStatus = true
-
-		if status.Ready() {
-			fmt.Fprintf(runtime.IO().ErrOut, "Export task completed: %s\n", common.MaskToken(status.FileToken))
-
-			// Export-only mode: caller wants the ready file token / metadata but
-			// no local download (e.g. sheets +workbook-export without an output
-			// path). Skip the download and return the status envelope.
-			if strings.TrimSpace(outputDir) == "" {
-				runtime.Out(map[string]interface{}{
-					"ticket":         ticket,
-					"token":          spec.Token,
-					"doc_type":       spec.DocType,
-					"file_extension": spec.FileExtension,
-					"file_token":     status.FileToken,
-					"file_name":      status.FileName,
-					"file_size":      status.FileSize,
-					"ready":          true,
-					"downloaded":     false,
-				}, nil)
-				return nil
+			content, ok := doc["content"].(string)
+			if !ok {
+				return errs.NewInternalError(errs.SubtypeInvalidResponse, "invalid markdown fetch response: missing document.content")
 			}
 
 			fileName := preferredFileName
 			if fileName == "" {
-				fileName = status.FileName
+				// Prefer the remote title for the exported file name, but still fall
+				// back to the token if metadata is empty.
+				title, err := common.FetchDriveMetaTitle(runtime, spec.Token, spec.DocType)
+				if err != nil {
+					fmt.Fprintf(runtime.IO().ErrOut, "Title lookup failed, using token as filename: %v\n", err)
+					title = spec.Token
+				}
+				fileName = title
 			}
 			fileName = ensureExportFileExtension(sanitizeExportFileName(fileName, spec.Token), spec.FileExtension)
-			out, err := downloadDriveExportFile(ctx, runtime, status.FileToken, outputDir, fileName, overwrite)
+			savedPath, err := saveContentToOutputDir(runtime.FileIO(), outputDir, fileName, []byte(content), overwrite)
 			if err != nil {
-				recoveryCommand := driveExportDownloadCommand(status.FileToken, fileName, outputDir, overwrite)
-				hint := fmt.Sprintf(
-					"the export artifact is already ready (ticket=%s, file_token=%s)\nretry download with: %s",
-					ticket,
-					status.FileToken,
-					recoveryCommand,
-				)
-				return appendDriveExportRecoveryHint(err, hint)
+				return err
 			}
-			out["ticket"] = ticket
-			out["doc_type"] = spec.DocType
-			out["file_extension"] = spec.FileExtension
-			runtime.Out(out, nil)
+
+			runtime.Out(map[string]interface{}{
+				"token":          spec.Token,
+				"doc_type":       spec.DocType,
+				"file_extension": spec.FileExtension,
+				"file_name":      filepath.Base(savedPath),
+				"saved_path":     savedPath,
+				"size_bytes":     len(content),
+			}, nil)
 			return nil
 		}
 
-		if status.Failed() {
-			msg := strings.TrimSpace(status.JobErrorMsg)
-			if msg == "" {
-				msg = status.StatusLabel()
+		ticket, err := createDriveExportTask(runtime, spec)
+		if err != nil {
+			return err
+		}
+		fmt.Fprintf(runtime.IO().ErrOut, "Created export task: %s\n", ticket)
+
+		var lastStatus driveExportStatus
+		var lastPollErr error
+		hasObservedStatus := false
+		// Keep the command responsive by polling for a bounded window. If the task
+		// is still running after that, return a resume command instead of blocking.
+		for attempt := 1; attempt <= driveExportPollAttempts; attempt++ {
+			if attempt > 1 {
+				select {
+				case <-ctx.Done():
+					return ctx.Err()
+				case <-time.After(driveExportPollInterval):
+				}
 			}
-			return errs.NewAPIError(errs.SubtypeServerError, "export task failed: %s (ticket=%s)", msg, ticket)
+			if err := ctx.Err(); err != nil {
+				return err
+			}
+
+			status, err := getDriveExportStatus(runtime, spec.Token, ticket)
+			if err != nil {
+				// Treat polling failures as transient so short-lived backend hiccups
+				// do not immediately fail an otherwise healthy export task.
+				lastPollErr = err
+				fmt.Fprintf(runtime.IO().ErrOut, "Export status attempt %d/%d failed: %v\n", attempt, driveExportPollAttempts, err)
+				continue
+			}
+			lastStatus = status
+			hasObservedStatus = true
+
+			if status.Ready() {
+				fmt.Fprintf(runtime.IO().ErrOut, "Export task completed: %s\n", common.MaskToken(status.FileToken))
+				fileName := preferredFileName
+				if fileName == "" {
+					fileName = status.FileName
+				}
+				fileName = ensureExportFileExtension(sanitizeExportFileName(fileName, spec.Token), spec.FileExtension)
+				out, err := downloadDriveExportFile(ctx, runtime, status.FileToken, outputDir, fileName, overwrite)
+				if err != nil {
+					recoveryCommand := driveExportDownloadCommand(status.FileToken, fileName, outputDir, overwrite)
+					hint := fmt.Sprintf(
+						"the export artifact is already ready (ticket=%s, file_token=%s)\nretry download with: %s",
+						ticket,
+						status.FileToken,
+						recoveryCommand,
+					)
+					return appendDriveExportRecoveryHint(err, hint)
+				}
+				out["ticket"] = ticket
+				out["doc_type"] = spec.DocType
+				out["file_extension"] = spec.FileExtension
+				runtime.Out(out, nil)
+				return nil
+			}
+
+			if status.Failed() {
+				msg := strings.TrimSpace(status.JobErrorMsg)
+				if msg == "" {
+					msg = status.StatusLabel()
+				}
+				return errs.NewAPIError(errs.SubtypeServerError, "export task failed: %s (ticket=%s)", msg, ticket)
+			}
+
+			fmt.Fprintf(runtime.IO().ErrOut, "Export status %d/%d: %s\n", attempt, driveExportPollAttempts, status.StatusLabel())
 		}
 
-		fmt.Fprintf(runtime.IO().ErrOut, "Export status %d/%d: %s\n", attempt, driveExportPollAttempts, status.StatusLabel())
-	}
+		nextCommand := driveExportTaskResultCommand(ticket, spec.Token)
+		if !hasObservedStatus && lastPollErr != nil {
+			hint := fmt.Sprintf(
+				"the export task was created but every status poll failed (ticket=%s)\nretry status lookup with: %s",
+				ticket,
+				nextCommand,
+			)
+			return appendDriveExportRecoveryHint(lastPollErr, hint)
+		}
 
-	nextCommand := driveExportTaskResultCommand(ticket, spec.Token)
-	if !hasObservedStatus && lastPollErr != nil {
-		hint := fmt.Sprintf(
-			"the export task was created but every status poll failed (ticket=%s)\nretry status lookup with: %s",
-			ticket,
-			nextCommand,
-		)
-		return appendDriveExportRecoveryHint(lastPollErr, hint)
-	}
-
-	failed := false
-	var jobStatus interface{}
-	jobStatusLabel := "unknown"
-	if hasObservedStatus {
-		failed = lastStatus.Failed()
-		jobStatus = lastStatus.JobStatus
-		jobStatusLabel = lastStatus.StatusLabel()
-	}
-	// Return the last observed status so callers can resume from a known task
-	// state instead of losing all progress information on timeout.
-	result := map[string]interface{}{
-		"ticket":           ticket,
-		"token":            spec.Token,
-		"doc_type":         spec.DocType,
-		"file_extension":   spec.FileExtension,
-		"ready":            false,
-		"failed":           failed,
-		"job_status":       jobStatus,
-		"job_status_label": jobStatusLabel,
-		"timed_out":        true,
-		"next_command":     nextCommand,
-	}
-	if preferredFileName != "" {
-		result["file_name"] = ensureExportFileExtension(sanitizeExportFileName(preferredFileName, spec.Token), spec.FileExtension)
-	}
-	runtime.Out(result, nil)
-	fmt.Fprintf(runtime.IO().ErrOut, "Export task is still in progress. Continue with: %s\n", nextCommand)
-	return nil
+		failed := false
+		var jobStatus interface{}
+		jobStatusLabel := "unknown"
+		if hasObservedStatus {
+			failed = lastStatus.Failed()
+			jobStatus = lastStatus.JobStatus
+			jobStatusLabel = lastStatus.StatusLabel()
+		}
+		// Return the last observed status so callers can resume from a known task
+		// state instead of losing all progress information on timeout.
+		result := map[string]interface{}{
+			"ticket":           ticket,
+			"token":            spec.Token,
+			"doc_type":         spec.DocType,
+			"file_extension":   spec.FileExtension,
+			"ready":            false,
+			"failed":           failed,
+			"job_status":       jobStatus,
+			"job_status_label": jobStatusLabel,
+			"timed_out":        true,
+			"next_command":     nextCommand,
+		}
+		if preferredFileName != "" {
+			result["file_name"] = ensureExportFileExtension(sanitizeExportFileName(preferredFileName, spec.Token), spec.FileExtension)
+		}
+		runtime.Out(result, nil)
+		fmt.Fprintf(runtime.IO().ErrOut, "Export task is still in progress. Continue with: %s\n", nextCommand)
+		return nil
+	},
 }

shortcuts/drive/drive_export_test.go

@@ -5,7 +5,6 @@ package drive
 
 import (
 	"bytes"
-	"context"
 	"encoding/json"
 	"errors"
 	"net/http"
@@ -498,72 +497,6 @@ func TestDriveExportAsyncSuccess(t *testing.T) {
 	}
 }
 
-// TestDriveExportEmptyOutputDirDownloadsToCwd guards the export refactor: an
-// explicit empty --output-dir must still download to the current directory
-// (normalized to "."), not trigger the export-only no-download path that the
-// shared RunExport core uses for sheets +workbook-export.
-func TestDriveExportEmptyOutputDirDownloadsToCwd(t *testing.T) {
-	f, stdout, _, reg := cmdutil.TestFactory(t, driveTestConfig())
-	reg.Register(&httpmock.Stub{
-		Method: "POST",
-		URL:    "/open-apis/drive/v1/export_tasks",
-		Body:   map[string]interface{}{"code": 0, "data": map[string]interface{}{"ticket": "tk_e"}},
-	})
-	reg.Register(&httpmock.Stub{
-		Method: "GET",
-		URL:    "/open-apis/drive/v1/export_tasks/tk_e",
-		Body: map[string]interface{}{"code": 0, "data": map[string]interface{}{
-			"result": map[string]interface{}{
-				"job_status": 0, "file_token": "box_e", "file_name": "report",
-				"file_extension": "pdf", "type": "docx", "file_size": 3,
-			},
-		}},
-	})
-	reg.Register(&httpmock.Stub{
-		Method:  "GET",
-		URL:     "/open-apis/drive/v1/export_tasks/file/box_e/download",
-		Status:  200,
-		RawBody: []byte("pdf"),
-		Headers: http.Header{
-			"Content-Type":        []string{"application/pdf"},
-			"Content-Disposition": []string{`attachment; filename="report.pdf"`},
-		},
-	})
-
-	tmpDir := t.TempDir()
-	withDriveWorkingDir(t, tmpDir)
-
-	prevAttempts, prevInterval := driveExportPollAttempts, driveExportPollInterval
-	driveExportPollAttempts, driveExportPollInterval = 1, 0
-	t.Cleanup(func() {
-		driveExportPollAttempts, driveExportPollInterval = prevAttempts, prevInterval
-	})
-
-	err := mountAndRunDrive(t, DriveExport, []string{
-		"+export",
-		"--token", "docx123",
-		"--doc-type", "docx",
-		"--file-extension", "pdf",
-		"--output-dir", "",
-		"--as", "bot",
-	}, f, stdout)
-	if err != nil {
-		t.Fatalf("unexpected error: %v", err)
-	}
-
-	// Empty --output-dir must still write to cwd, not skip the download.
-	data, err := os.ReadFile(filepath.Join(tmpDir, "report.pdf"))
-	if err != nil {
-		t.Fatalf("empty --output-dir should still download to cwd: %v", err)
-	}
-	if string(data) != "pdf" {
-		t.Fatalf("downloaded content = %q", string(data))
-	}
-	if strings.Contains(stdout.String(), `"downloaded": false`) {
-		t.Fatalf("export-only path must not trigger for drive +export: %s", stdout.String())
-	}
-}
-
 func TestDriveExportAsyncUsesProvidedFileName(t *testing.T) {
 	f, stdout, _, reg := cmdutil.TestFactory(t, driveTestConfig())
 	reg.Register(&httpmock.Stub{
@@ -1101,37 +1034,3 @@ func TestDriveTaskResultExportIncludesReadyFlags(t *testing.T) {
 		t.Fatalf("stdout missing job_status_label: %s", stdout.String())
 	}
 }
-
-// TestWrapExportContextErr verifies the export poll loop's typed wrapping for
-// context cancellation / deadline. Previously the poll loop returned ctx.Err()
-// directly so an untyped context.Canceled would escape as a plain string at
-// the command layer, bypassing the typed-error contract.
-func TestWrapExportContextErr(t *testing.T) {
-	if err := wrapExportContextErr(nil); err != nil {
-		t.Errorf("wrapExportContextErr(nil) = %v, want nil", err)
-	}
-
-	cancelled := wrapExportContextErr(context.Canceled)
-	var netErrCancel *errs.NetworkError
-	if !errors.As(cancelled, &netErrCancel) {
-		t.Fatalf("wrapExportContextErr(Canceled) = %T, want *errs.NetworkError", cancelled)
-	}
-	if netErrCancel.Subtype != errs.SubtypeNetworkTransport {
-		t.Errorf("Canceled subtype = %q, want %q", netErrCancel.Subtype, errs.SubtypeNetworkTransport)
-	}
-	if !errors.Is(cancelled, context.Canceled) {
-		t.Error("wrapExportContextErr should preserve context.Canceled via errors.Is")
-	}
-
-	deadline := wrapExportContextErr(context.DeadlineExceeded)
-	var netErrDeadline *errs.NetworkError
-	if !errors.As(deadline, &netErrDeadline) {
-		t.Fatalf("wrapExportContextErr(DeadlineExceeded) = %T, want *errs.NetworkError", deadline)
-	}
-	if netErrDeadline.Subtype != errs.SubtypeNetworkTimeout {
-		t.Errorf("DeadlineExceeded subtype = %q, want %q", netErrDeadline.Subtype, errs.SubtypeNetworkTimeout)
-	}
-	if !errors.Is(deadline, context.DeadlineExceeded) {
-		t.Error("wrapExportContextErr should preserve context.DeadlineExceeded via errors.Is")
-	}
-}

shortcuts/drive/drive_import.go

@@ -35,164 +35,132 @@ var DriveImport = common.Shortcut{
 		{Name: "target-token", Desc: "existing token to import data into (only for type=bitable); when set, data is mounted into this bitable instead of creating a new one"},
 	},
 	Validate: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		return ValidateImport(importParamsFromFlags(runtime))
+		return validateDriveImportSpec(driveImportSpec{
+			FilePath:    runtime.Str("file"),
+			DocType:     strings.ToLower(runtime.Str("type")),
+			FolderToken: runtime.Str("folder-token"),
+			Name:        runtime.Str("name"),
+			TargetToken: runtime.Str("target-token"),
+		})
 	},
 	DryRun: func(ctx context.Context, runtime *common.RuntimeContext) *common.DryRunAPI {
-		return PlanImportDryRun(runtime, importParamsFromFlags(runtime))
+		spec := driveImportSpec{
+			FilePath:    runtime.Str("file"),
+			DocType:     strings.ToLower(runtime.Str("type")),
+			FolderToken: runtime.Str("folder-token"),
+			Name:        runtime.Str("name"),
+			TargetToken: runtime.Str("target-token"),
+		}
+		fileSize, err := preflightDriveImportFile(runtime.FileIO(), &spec)
+		if err != nil {
+			return common.NewDryRunAPI().Set("error", err.Error())
+		}
+		if valErr := validateDriveImportSpec(spec); valErr != nil {
+			return common.NewDryRunAPI().Set("error", valErr.Error())
+		}
+
+		dry := common.NewDryRunAPI()
+		dry.Desc("Upload file (single-part or multipart) -> create import task -> poll status")
+
+		appendDriveImportFolderTokenWikiCheckDryRun(dry, spec)
+		appendDriveImportUploadDryRun(dry, spec, fileSize)
+
+		dry.POST("/open-apis/drive/v1/import_tasks").
+			Desc("[2] Create import task").
+			Body(spec.CreateTaskBody("<file_token>"))
+
+		dry.GET("/open-apis/drive/v1/import_tasks/:ticket").
+			Desc("[3] Poll import task result").
+			Set("ticket", "<ticket>")
+		if runtime.IsBot() {
+			dry.Desc("After the import result returns the final cloud document target in bot mode, the CLI will also try to grant the current CLI user full_access (可管理权限) on it.")
+		}
+
+		return dry
 	},
 	Execute: func(ctx context.Context, runtime *common.RuntimeContext) error {
-		return RunImport(ctx, runtime, importParamsFromFlags(runtime))
+		spec := driveImportSpec{
+			FilePath:    runtime.Str("file"),
+			DocType:     strings.ToLower(runtime.Str("type")),
+			FolderToken: runtime.Str("folder-token"),
+			Name:        runtime.Str("name"),
+			TargetToken: runtime.Str("target-token"),
+		}
+		if _, err := preflightDriveImportFile(runtime.FileIO(), &spec); err != nil {
+			return err
+		}
+		if err := rejectDriveImportWikiFolderToken(runtime, spec.FolderToken); err != nil {
+			return err
+		}
+
+		// Step 1: Upload file as media
+		fileToken, uploadErr := uploadMediaForImport(ctx, runtime, spec.FilePath, spec.SourceFileName(), spec.DocType)
+		if uploadErr != nil {
+			return uploadErr
+		}
+
+		fmt.Fprintf(runtime.IO().ErrOut, "Creating import task for %s as %s...\n", spec.TargetFileName(), spec.DocType)
+
+		// Step 2: Create import task
+		ticket, err := createDriveImportTask(runtime, spec, fileToken)
+		if err != nil {
+			return err
+		}
+
+		// Step 3: Poll task
+		fmt.Fprintf(runtime.IO().ErrOut, "Polling import task %s...\n", ticket)
+
+		status, ready, err := pollDriveImportTask(runtime, ticket)
+		if err != nil {
+			return err
+		}
+
+		// Some intermediate responses omit the final type, so fall back to the
+		// requested type to keep the output shape stable.
+		resultType := status.DocType
+		if resultType == "" {
+			resultType = spec.DocType
+		}
+		out := map[string]interface{}{
+			"ticket":           ticket,
+			"type":             resultType,
+			"ready":            ready,
+			"job_status":       status.JobStatus,
+			"job_status_label": status.StatusLabel(),
+		}
+		if status.Token != "" {
+			out["token"] = status.Token
+		}
+		if statusURL := strings.TrimSpace(status.URL); statusURL != "" {
+			out["url"] = statusURL
+		} else if status.Token != "" {
+			if u := common.BuildResourceURL(runtime.Config.Brand, normalizeDriveImportKindForURL(resultType, spec.DocType), status.Token); u != "" {
+				out["url"] = u
+			}
+		}
+		if status.JobErrorMsg != "" {
+			out["job_error_msg"] = status.JobErrorMsg
+		}
+		if status.Extra != nil {
+			out["extra"] = status.Extra
+		}
+		if !ready {
+			nextCommand := driveImportTaskResultCommand(ticket)
+			fmt.Fprintf(runtime.IO().ErrOut, "Import task is still in progress. Continue with: %s\n", nextCommand)
+			out["timed_out"] = true
+			out["next_command"] = nextCommand
+		}
+		if ready {
+			if grant := common.AutoGrantCurrentUserDrivePermission(runtime, common.GetString(out, "token"), resultType); grant != nil {
+				out["permission_grant"] = grant
+			}
+		}
+
+		runtime.Out(out, nil)
+		return nil
 	},
 }
 
-// ImportParams holds the user-facing inputs for an import flow, decoupled from
-// cobra flags so other command groups (e.g. sheets +workbook-import) can reuse
-// the drive import implementation without taking a dependency on a --type flag.
-type ImportParams struct {
-	File        string
-	DocType     string
-	FolderToken string
-	Name        string
-	TargetToken string
-}
-
-func (p ImportParams) spec() driveImportSpec {
-	return driveImportSpec{
-		FilePath:    p.File,
-		DocType:     strings.ToLower(p.DocType),
-		FolderToken: p.FolderToken,
-		Name:        p.Name,
-		TargetToken: p.TargetToken,
-	}
-}
-
-// importParamsFromFlags reads the standard drive +import flag set.
-func importParamsFromFlags(runtime *common.RuntimeContext) ImportParams {
-	return ImportParams{
-		File:        runtime.Str("file"),
-		DocType:     runtime.Str("type"),
-		FolderToken: runtime.Str("folder-token"),
-		Name:        runtime.Str("name"),
-		TargetToken: runtime.Str("target-token"),
-	}
-}
-
-// ValidateImport runs the CLI-level compatibility checks for an import.
-func ValidateImport(p ImportParams) error {
-	return validateDriveImportSpec(p.spec())
-}
-
-// PlanImportDryRun builds the dry-run plan (upload -> create task -> poll) for
-// an import without performing any network or file I/O beyond a local stat.
-func PlanImportDryRun(runtime *common.RuntimeContext, p ImportParams) *common.DryRunAPI {
-	spec := p.spec()
-	fileSize, err := preflightDriveImportFile(runtime.FileIO(), &spec)
-	if err != nil {
-		return common.NewDryRunAPI().Set("error", err.Error())
-	}
-	if valErr := validateDriveImportSpec(spec); valErr != nil {
-		return common.NewDryRunAPI().Set("error", valErr.Error())
-	}
-
-	dry := common.NewDryRunAPI()
-	dry.Desc("Upload file (single-part or multipart) -> create import task -> poll status")
-
-	appendDriveImportFolderTokenWikiCheckDryRun(dry, spec)
-	appendDriveImportUploadDryRun(dry, spec, fileSize)
-
-	dry.POST("/open-apis/drive/v1/import_tasks").
-		Desc("[2] Create import task").
-		Body(spec.CreateTaskBody("<file_token>"))
-
-	dry.GET("/open-apis/drive/v1/import_tasks/:ticket").
-		Desc("[3] Poll import task result").
-		Set("ticket", "<ticket>")
-	if runtime.IsBot() {
-		dry.Desc("After the import result returns the final cloud document target in bot mode, the CLI will also try to grant the current CLI user full_access (可管理权限) on it.")
-	}
-
-	return dry
-}
-
-// RunImport executes the full import flow: upload media -> create import task ->
-// bounded poll, then writes the result envelope to the runtime output. It is
-// the shared core behind both drive +import and sheets +workbook-import.
-func RunImport(ctx context.Context, runtime *common.RuntimeContext, p ImportParams) error {
-	spec := p.spec()
-	if _, err := preflightDriveImportFile(runtime.FileIO(), &spec); err != nil {
-		return err
-	}
-	if err := rejectDriveImportWikiFolderToken(runtime, spec.FolderToken); err != nil {
-		return err
-	}
-
-	// Step 1: Upload file as media
-	fileToken, uploadErr := uploadMediaForImport(ctx, runtime, spec.FilePath, spec.SourceFileName(), spec.DocType)
-	if uploadErr != nil {
-		return uploadErr
-	}
-
-	fmt.Fprintf(runtime.IO().ErrOut, "Creating import task for %s as %s...\n", spec.TargetFileName(), spec.DocType)
-
-	// Step 2: Create import task
-	ticket, err := createDriveImportTask(runtime, spec, fileToken)
-	if err != nil {
-		return err
-	}
-
-	// Step 3: Poll task
-	fmt.Fprintf(runtime.IO().ErrOut, "Polling import task %s...\n", ticket)
-
-	status, ready, err := pollDriveImportTask(runtime, ticket)
-	if err != nil {
-		return err
-	}
-
-	// Some intermediate responses omit the final type, so fall back to the
-	// requested type to keep the output shape stable.
-	resultType := status.DocType
-	if resultType == "" {
-		resultType = spec.DocType
-	}
-	out := map[string]interface{}{
-		"ticket":           ticket,
-		"type":             resultType,
-		"ready":            ready,
-		"job_status":       status.JobStatus,
-		"job_status_label": status.StatusLabel(),
-	}
-	if status.Token != "" {
-		out["token"] = status.Token
-	}
-	if statusURL := strings.TrimSpace(status.URL); statusURL != "" {
-		out["url"] = statusURL
-	} else if status.Token != "" {
-		if u := common.BuildResourceURL(runtime.Config.Brand, normalizeDriveImportKindForURL(resultType, spec.DocType), status.Token); u != "" {
-			out["url"] = u
-		}
-	}
-	if status.JobErrorMsg != "" {
-		out["job_error_msg"] = status.JobErrorMsg
-	}
-	if status.Extra != nil {
-		out["extra"] = status.Extra
-	}
-	if !ready {
-		nextCommand := driveImportTaskResultCommand(ticket)
-		fmt.Fprintf(runtime.IO().ErrOut, "Import task is still in progress. Continue with: %s\n", nextCommand)
-		out["timed_out"] = true
-		out["next_command"] = nextCommand
-	}
-	if ready {
-		if grant := common.AutoGrantCurrentUserDrivePermission(runtime, common.GetString(out, "token"), resultType); grant != nil {
-			out["permission_grant"] = grant
-		}
-	}
-
-	runtime.Out(out, nil)
-	return nil
-}
-
 func preflightDriveImportFile(fio fileio.FileIO, spec *driveImportSpec) (int64, error) {
 	// Keep dry-run and execution aligned on path normalization, file existence,
 	// and format-specific size limits before planning the upload path.

shortcuts/im/builders_test.go

@@ -469,29 +469,6 @@ func TestShortcutValidateBranches(t *testing.T) {
 		}
 	})
 
-	t.Run("ImMessagesSend audio rejects non-opus local file", func(t *testing.T) {
-		runtime := newTestRuntimeContext(t, map[string]string{
-			"chat-id": "oc_123",
-			"audio":   "./voice.mp3",
-		}, nil)
-		err := ImMessagesSend.Validate(context.Background(), runtime)
-		if err == nil || !strings.Contains(err.Error(), "--audio supports only Opus audio files") {
-			t.Fatalf("ImMessagesSend.Validate() error = %v", err)
-		}
-	})
-
-	t.Run("ImMessagesSend audio accepts opus and ogg local files", func(t *testing.T) {
-		for _, audio := range []string{"./voice.opus", "./voice.ogg"} {
-			runtime := newTestRuntimeContext(t, map[string]string{
-				"chat-id": "oc_123",
-				"audio":   audio,
-			}, nil)
-			if err := ImMessagesSend.Validate(context.Background(), runtime); err != nil {
-				t.Fatalf("ImMessagesSend.Validate(%q) unexpected error = %v", audio, err)
-			}
-		}
-	})
-
 	t.Run("ImMessagesSend conflicting explicit msg-type", func(t *testing.T) {
 		runtime := newTestRuntimeContext(t, map[string]string{
 			"chat-id":  "oc_123",
@@ -515,17 +492,6 @@ func TestShortcutValidateBranches(t *testing.T) {
 		}
 	})
 
-	t.Run("ImMessagesReply audio rejects non-opus local file", func(t *testing.T) {
-		runtime := newTestRuntimeContext(t, map[string]string{
-			"message-id": "om_123",
-			"audio":      "./voice.mp3",
-		}, nil)
-		err := ImMessagesReply.Validate(context.Background(), runtime)
-		if err == nil || !strings.Contains(err.Error(), "--audio supports only Opus audio files") {
-			t.Fatalf("ImMessagesReply.Validate() error = %v", err)
-		}
-	})
-
 	t.Run("ImThreadsMessagesList invalid thread", func(t *testing.T) {
 		runtime := newTestRuntimeContext(t, map[string]string{
 			"thread": "bad_thread",