emlbuilder: add LMSReplyType method for reply/forward type header

Add LMSReplyType method to the EML builder that sets the
X-LMS-Reply-Type header. This allows reply, reply-all, and forward
shortcuts to tag the EML with REPLY or FORWARD so that downstream
services can label the original message accordingly.

Call LMSReplyType("REPLY") in mail_reply.go and mail_reply_all.go,
and LMSReplyType("FORWARD") in mail_forward.go, immediately after
the existing LMSReplyToMessageID call.

.github/workflows/ci.yml

@@ -5,6 +5,7 @@ on:
     branches: [main]
   pull_request:
     branches: [main]
+    types: [opened, synchronize, reopened, edited]
   workflow_dispatch:
 
 permissions:
@@ -70,6 +71,7 @@ jobs:
       - uses: actions/checkout@93cb6efe18208431cddfb8368fd83d5badbf9bfd # v5
         with:
           fetch-depth: 0
+          persist-credentials: false
       - uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6
         with:
           go-version-file: go.mod
@@ -87,6 +89,23 @@ jobs:
       - name: Run errs/ lint guards (lintcheck)
         run: go run -C lint . --changed-from "$QUALITY_GATE_CHANGED_FROM" ..
 
+  script-test:
+    needs: fast-gate
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@93cb6efe18208431cddfb8368fd83d5badbf9bfd # v5
+        with:
+          fetch-depth: 0
+          persist-credentials: false
+      - uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6
+        with:
+          go-version-file: go.mod
+      - uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4
+        with:
+          node-version: '22'
+      - name: Run script tests
+        run: make script-test
+
   deterministic-gate:
     needs: fast-gate
     runs-on: ubuntu-latest
@@ -109,8 +128,28 @@ jobs:
         env:
           QUALITY_GATE_CHANGED_FROM: ${{ github.event.pull_request.base.sha || github.event.before || 'origin/main' }}
         run: echo "QUALITY_GATE_CHANGED_FROM=$(bash scripts/resolve-changed-from.sh)" >> "$GITHUB_ENV"
+      - name: Write public content metadata
+        if: ${{ github.event_name == 'pull_request' }}
+        env:
+          PR_TITLE: ${{ github.event.pull_request.title }}
+          PR_BODY: ${{ github.event.pull_request.body }}
+          PR_BRANCH: ${{ github.head_ref }}
+        run: |
+          mkdir -p .tmp/quality-gate
+          python3 - <<'PY'
+          import json
+          import os
+
+          with open(".tmp/quality-gate/public-content-metadata.json", "w", encoding="utf-8") as f:
+              json.dump({
+                  "title": os.environ.get("PR_TITLE", ""),
+                  "body": os.environ.get("PR_BODY", ""),
+                  "branch": os.environ.get("PR_BRANCH", ""),
+              }, f)
+              f.write("\n")
+          PY
       - name: Run CLI deterministic gate
-        run: make quality-gate
+        run: PUBLIC_CONTENT_METADATA=.tmp/quality-gate/public-content-metadata.json make quality-gate
       - name: Upload quality gate facts
         if: ${{ always() && github.event_name == 'pull_request' }}
         uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4
@@ -220,7 +259,7 @@ jobs:
 
   # ── Layer 3: E2E Gate ──────────────────────────────────────────────
   e2e-dry-run:
-    needs: [unit-test, lint, deterministic-gate]
+    needs: [unit-test, lint, script-test, deterministic-gate]
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@93cb6efe18208431cddfb8368fd83d5badbf9bfd # v5
@@ -241,7 +280,7 @@ jobs:
         run: go test -v -count=1 -timeout=5m ./tests/cli_e2e/... -run 'DryRun|Regression'
 
   e2e-live:
-    needs: [unit-test, lint, deterministic-gate]
+    needs: [unit-test, lint, script-test, deterministic-gate]
     if: ${{ github.event_name != 'pull_request' || !github.event.pull_request.head.repo.fork }}
     runs-on: ubuntu-latest
     permissions:
@@ -333,7 +372,7 @@ jobs:
   # ── Results Gate (single required check for branch protection) ─────
   results:
     if: ${{ always() }}
-    needs: [fast-gate, unit-test, lint, deterministic-gate, coverage, deadcode, e2e-dry-run, e2e-live, security, license-header]
+    needs: [fast-gate, unit-test, lint, script-test, deterministic-gate, coverage, deadcode, e2e-dry-run, e2e-live, security, license-header]
     runs-on: ubuntu-latest
     steps:
       - name: Evaluate results
@@ -345,6 +384,7 @@ jobs:
           echo "| L1 | fast-gate | ${{ needs.fast-gate.result }} |" >> $GITHUB_STEP_SUMMARY
           echo "| L2 | unit-test | ${{ needs.unit-test.result }} |" >> $GITHUB_STEP_SUMMARY
           echo "| L2 | lint | ${{ needs.lint.result }} |" >> $GITHUB_STEP_SUMMARY
+          echo "| L2 | script-test | ${{ needs.script-test.result }} |" >> $GITHUB_STEP_SUMMARY
           echo "| L2 | deterministic-gate | ${{ needs.deterministic-gate.result }} |" >> $GITHUB_STEP_SUMMARY
           echo "| L2 | coverage | ${{ needs.coverage.result }} |" >> $GITHUB_STEP_SUMMARY
           echo "| L2 | deadcode | ${{ needs.deadcode.result }} |" >> $GITHUB_STEP_SUMMARY
@@ -361,6 +401,7 @@ jobs:
             "${{ needs.fast-gate.result }}" \
             "${{ needs.unit-test.result }}" \
             "${{ needs.lint.result }}" \
+            "${{ needs.script-test.result }}" \
             "${{ needs.deterministic-gate.result }}" \
             "${{ needs.coverage.result }}" \
             "${{ needs.deadcode.result }}" \

.github/workflows/comment-audit.yml

@@ -0,0 +1,28 @@
+name: Comment Audit
+
+on:
+  issue_comment:
+    types: [created, edited]
+  pull_request_review:
+    types: [submitted, edited]
+  pull_request_review_comment:
+    types: [created, edited]
+
+permissions:
+  contents: read
+
+jobs:
+  public-content-comment-audit:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@93cb6efe18208431cddfb8368fd83d5badbf9bfd # v5
+        with:
+          persist-credentials: false
+      - uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6
+        with:
+          go-version-file: go.mod
+      - name: Post-publication comment audit
+        run: |
+          mkdir -p .tmp/comment-audit
+          cp "$GITHUB_EVENT_PATH" .tmp/comment-audit/event.json
+          go run ./internal/qualitygate/cmd/comment-audit --event .tmp/comment-audit/event.json --kind "$GITHUB_EVENT_NAME"

.github/workflows/semantic-review.yml

@@ -47,10 +47,13 @@ jobs:
               throw new Error(`ambiguous workflow_run pull request bindings: ${runPRs.length}`);
             }
             let prNumber = Number(runPRs[0]?.number || 0);
-            let eventBaseSha = runPRs[0]?.base?.sha || "";
+            const eventBaseSha = runPRs[0]?.base?.sha || "";
             const eventHeadSha = runPRs[0]?.head?.sha || "";
-            const targetHeadSha = eventHeadSha || run.head_sha;
+            const targetHeadSha = run.head_sha;
             if (!/^[a-f0-9]{40}$/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({
@@ -71,11 +74,11 @@ 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) {
@@ -85,31 +88,44 @@ jobs:
                 commit_sha: targetHeadSha,
               });
               const candidatePRs = associatedPRs.filter((candidate) =>
-                candidate.state === "open" &&
                 candidate.base?.repo?.id === context.payload.repository.id &&
                 candidate.head?.sha === targetHeadSha
               );
-              if (candidatePRs.length > 1) {
-                throw new Error(`ambiguous open PRs for workflow_run head ${targetHeadSha}: ${candidatePRs.length}`);
+              const openCandidatePRs = candidatePRs.filter((candidate) => candidate.state === "open");
+              if (openCandidatePRs.length > 1) {
+                throw new Error(`ambiguous open PRs for workflow_run head ${targetHeadSha}: ${openCandidatePRs.length}`);
               }
-              if (candidatePRs.length === 1) {
-                prNumber = candidatePRs[0].number;
+              if (openCandidatePRs.length === 1) {
+                prNumber = openCandidatePRs[0].number;
+              } else if (candidatePRs.length > 0) {
+                core.notice("PR quality summary skipped: workflow_run target PR is no longer open");
+                core.setOutput("stale", "true");
+                return;
               }
             }
             if (!prNumber) {
               const candidatePRs = await github.paginate(github.rest.pulls.list, {
                 owner: context.repo.owner,
                 repo: context.repo.repo,
-                state: "open",
+                state: "all",
                 per_page: 100,
               }).then((prs) => prs.filter((candidate) =>
                 candidate.base?.repo?.id === context.payload.repository.id &&
                 candidate.head?.sha === targetHeadSha
               ));
-              if (candidatePRs.length !== 1) {
+              const openCandidatePRs = candidatePRs.filter((candidate) => candidate.state === "open");
+              if (openCandidatePRs.length > 1) {
+                throw new Error(`ambiguous open PRs from pull list fallback for workflow_run head ${targetHeadSha}: ${openCandidatePRs.length}`);
+              }
+              if (openCandidatePRs.length === 1) {
+                prNumber = openCandidatePRs[0].number;
+              } else if (candidatePRs.length > 0) {
+                core.notice("PR quality summary skipped: workflow_run target PR is no longer open");
+                core.setOutput("stale", "true");
+                return;
+              } else {
                 throw new Error(`expected one open PR from pull list fallback for workflow_run head ${targetHeadSha}, got ${candidatePRs.length}`);
               }
-              prNumber = candidatePRs[0].number;
             }
             if (!Number.isInteger(prNumber) || prNumber <= 0) throw new Error("missing pull request binding");
             const { data: pr } = await github.rest.pulls.get({
@@ -118,12 +134,17 @@ jobs:
               pull_number: prNumber,
             });
             if (pr.base.repo.id !== context.payload.repository.id) throw new Error("PR base repo mismatch");
+            if (pr.state !== "open") {
+              core.notice("PR quality summary skipped: workflow_run target PR is no longer open");
+              core.setOutput("stale", "true");
+              return;
+            }
             if (pr.head.sha !== targetHeadSha) {
               core.notice("PR quality summary skipped: workflow_run is stale for this PR head");
               core.setOutput("stale", "true");
               return;
             }
-            const baseSha = eventBaseSha || artifactBaseSha || pr.base.sha;
+            const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha;
             if (!/^[a-f0-9]{40}$/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");
@@ -255,10 +276,13 @@ jobs:
               throw new Error(`ambiguous workflow_run pull request bindings: ${runPRs.length}`);
             }
             let prNumber = Number(runPRs[0]?.number || 0);
-            let eventBaseSha = runPRs[0]?.base?.sha || "";
+            const eventBaseSha = runPRs[0]?.base?.sha || "";
             const eventHeadSha = runPRs[0]?.head?.sha || "";
-            const targetHeadSha = eventHeadSha || run.head_sha;
+            const targetHeadSha = run.head_sha;
             if (!/^[a-f0-9]{40}$/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({
@@ -279,11 +303,11 @@ 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) {
@@ -293,31 +317,44 @@ jobs:
                 commit_sha: targetHeadSha,
               });
               const candidatePRs = associatedPRs.filter((candidate) =>
-                candidate.state === "open" &&
                 candidate.base?.repo?.id === context.payload.repository.id &&
                 candidate.head?.sha === targetHeadSha
               );
-              if (candidatePRs.length > 1) {
-                throw new Error(`ambiguous open PRs for workflow_run head ${targetHeadSha}: ${candidatePRs.length}`);
+              const openCandidatePRs = candidatePRs.filter((candidate) => candidate.state === "open");
+              if (openCandidatePRs.length > 1) {
+                throw new Error(`ambiguous open PRs for workflow_run head ${targetHeadSha}: ${openCandidatePRs.length}`);
               }
-              if (candidatePRs.length === 1) {
-                prNumber = candidatePRs[0].number;
+              if (openCandidatePRs.length === 1) {
+                prNumber = openCandidatePRs[0].number;
+              } else if (candidatePRs.length > 0) {
+                core.notice("semantic review skipped: workflow_run target PR is no longer open");
+                core.setOutput("stale", "true");
+                return;
               }
             }
             if (!prNumber) {
               const candidatePRs = await github.paginate(github.rest.pulls.list, {
                 owner: context.repo.owner,
                 repo: context.repo.repo,
-                state: "open",
+                state: "all",
                 per_page: 100,
               }).then((prs) => prs.filter((candidate) =>
                 candidate.base?.repo?.id === context.payload.repository.id &&
                 candidate.head?.sha === targetHeadSha
               ));
-              if (candidatePRs.length !== 1) {
+              const openCandidatePRs = candidatePRs.filter((candidate) => candidate.state === "open");
+              if (openCandidatePRs.length > 1) {
+                throw new Error(`ambiguous open PRs from pull list fallback for workflow_run head ${targetHeadSha}: ${openCandidatePRs.length}`);
+              }
+              if (openCandidatePRs.length === 1) {
+                prNumber = openCandidatePRs[0].number;
+              } else if (candidatePRs.length > 0) {
+                core.notice("semantic review skipped: workflow_run target PR is no longer open");
+                core.setOutput("stale", "true");
+                return;
+              } else {
                 throw new Error(`expected one open PR from pull list fallback for workflow_run head ${targetHeadSha}, got ${candidatePRs.length}`);
               }
-              prNumber = candidatePRs[0].number;
             }
             if (!Number.isInteger(prNumber) || prNumber <= 0) throw new Error("missing pull request binding");
             const { data: pr } = await github.rest.pulls.get({
@@ -326,12 +363,22 @@ jobs:
               pull_number: prNumber,
             });
             if (pr.base.repo.id !== context.payload.repository.id) throw new Error("PR base repo mismatch");
+            if (pr.state !== "open") {
+              core.notice("semantic review skipped: workflow_run target PR is no longer open");
+              core.setOutput("stale", "true");
+              return;
+            }
+            if (!pr.head.repo) {
+              core.notice("semantic review skipped: workflow_run target PR head repository is unavailable");
+              core.setOutput("stale", "true");
+              return;
+            }
             if (pr.head.sha !== targetHeadSha) {
               core.notice("semantic review skipped: workflow_run is stale for this PR head");
               core.setOutput("stale", "true");
               return;
             }
-            const baseSha = eventBaseSha || artifactBaseSha || pr.base.sha;
+            const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha;
             if (!/^[a-f0-9]{40}$/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");
@@ -383,6 +430,10 @@ jobs:
               repo: context.repo.repo,
               pull_number: pr,
             });
+            if (pull.state !== "open") {
+              core.notice("semantic review skipped infrastructure failure check: PR is no longer open");
+              return;
+            }
             if (pull.head.sha !== headSha) {
               core.notice("semantic review skipped infrastructure failure check: PR head changed");
               return;

.gitignore

@@ -7,6 +7,11 @@ bin/
 # Node
 node_modules/
 
+# Python (skill-bundled helper scripts)
+__pycache__/
+*.py[cod]
+*$py.class
+
 
 # OS
 .DS_Store
@@ -46,3 +51,4 @@ app.log
 cover*.out
 
 lark-env.sh
+/automations/

CHANGELOG.md

@@ -2,6 +2,71 @@
 
 All notable changes to this project will be documented in this file.
 
+## [v1.0.59] - 2026-06-26
+
+### Features
+
+- **slides**: Add `+replace-pages` and `xml get` shortcuts, and expose the presentation URL (#1585)
+- **minutes**: Support speaker list and no-Lark speaker replace (#1594)
+- **calendar/vc/minutes**: Optimize and extend calendar, vc, minutes, and note shortcuts and skills (#1571)
+
+### Bug Fixes
+
+- **docs**: Hide docs `api-version` compat flag (#1580)
+
+## [v1.0.58] - 2026-06-25
+
+### Features
+
+- **sheets**: Typed table I/O and error contract, workbook import/export, and skill refresh (#1355)
+- **base**: Add Base URL and title resolve shortcuts (#1338)
+- **drive**: Add `+member-add` shortcut with wiki space member collection collaborator support (#1204)
+- **doc**: Support `create` title option (#1536)
+- **doc**: Add `im-markdown` output format for doc fetch (#1550)
+- **whiteboard**: Export whiteboard as SVG and update whiteboard via SVG (#1559)
+- **card**: Support `card.action.trigger` event with auto-fetched card content (#1528)
+- **task**: Add task event consumer (#1510)
+
+### Bug Fixes
+
+- **doc**: Prefix docs resource shortcuts (#1564)
+- **binding**: Skip unix mode audit on Windows (#1525)
+
+### Documentation
+
+- **approval**: Sync approval skill for meta API commands (#1499)
+- **doc**: Restore lark-doc style requirements (#1579)
+- **im**: Document `chat.nickname` get/update/delete (#1378)
+- **im**: Clarify audio message opus requirement (#1271)
+
+### Build
+
+- **ci**: Add public content safeguards and reduce false positives
+
+## [v1.0.57] - 2026-06-23
+
+### Features
+
+- **slides**: Add `+screenshot` to capture slide page images (or render a single `<slide>` XML snippet), returning the local file path instead of Base64 (#1358)
+- **base**: Support record comments (#1043)
+- **search**: Surface search API notices (#1413)
+
+### Bug Fixes
+
+- **mail**: Resolve folder/label filter once per `+triage list` call (#1512)
+- **meta**: Backfill enum value descriptions from options (#1541)
+- **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
@@ -1212,6 +1277,9 @@ Bundled AI agent skills for intelligent assistance:
 - Bilingual documentation (English & Chinese).
 - CI/CD pipelines: linting, testing, coverage reporting, and automated releases.
 
+[v1.0.59]: https://github.com/larksuite/cli/releases/tag/v1.0.59
+[v1.0.58]: https://github.com/larksuite/cli/releases/tag/v1.0.58
+[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

Makefile

@@ -12,6 +12,7 @@ QUALITY_GATE_DIR ?= .tmp/quality-gate
 QUALITY_GATE_MANIFEST_OUT ?= $(QUALITY_GATE_DIR)/command-manifest.json
 QUALITY_GATE_COMMAND_INDEX_OUT ?= $(QUALITY_GATE_DIR)/command-index.json
 QUALITY_GATE_FACTS_OUT ?= $(QUALITY_GATE_DIR)/facts.json
+PUBLIC_CONTENT_METADATA ?= $(QUALITY_GATE_DIR)/public-content-metadata.json
 LDFLAGS  := -s -w -X $(MODULE)/internal/build.Version=$(VERSION) -X $(MODULE)/internal/build.Date=$(DATE)
 PREFIX   ?= /usr/local
 
@@ -69,7 +70,8 @@ integration-test: build
 test: vet fmt-check script-test unit-test examples-build integration-test
 
 quality-gate: build
-	mkdir -p $(QUALITY_GATE_DIR) $(dir $(QUALITY_GATE_FACTS_OUT))
+	mkdir -p $(QUALITY_GATE_DIR) $(dir $(QUALITY_GATE_FACTS_OUT)) $(dir $(PUBLIC_CONTENT_METADATA))
+	test -f $(PUBLIC_CONTENT_METADATA) || printf '{}\n' > $(PUBLIC_CONTENT_METADATA)
 	LARKSUITE_CLI_REMOTE_META=off \
 	LARKSUITE_CLI_NO_UPDATE_NOTIFIER=1 \
 	LARKSUITE_CLI_NO_SKILLS_NOTIFIER=1 \
@@ -89,6 +91,7 @@ quality-gate: build
 		--changed-from $(QUALITY_GATE_CHANGED_FROM_RESOLVED) \
 		--manifest $(QUALITY_GATE_MANIFEST_OUT) \
 		--command-index $(QUALITY_GATE_COMMAND_INDEX_OUT) \
+		--public-content-metadata $(PUBLIC_CONTENT_METADATA) \
 		--facts-out $(QUALITY_GATE_FACTS_OUT)
 
 install: build

README.md

@@ -198,7 +198,7 @@ Prefixed with `+`, designed to be friendly for both humans and AI, with smart de
 ```bash
 lark-cli calendar +agenda
 lark-cli im +messages-send --chat-id "oc_xxx" --text "Hello"
-lark-cli docs +create --api-version v2 --doc-format markdown --content $'<title>Weekly Report</title>\n# Progress\n- Completed feature X'
+lark-cli docs +create --doc-format markdown --content $'<title>Weekly Report</title>\n# Progress\n- Completed feature X'
 ```
 
 Run `lark-cli <service> --help` to see all shortcut commands.

README.zh.md

@@ -199,7 +199,7 @@ CLI 提供三种粒度的调用方式，覆盖从快速操作到完全自定义
 ```bash
 lark-cli calendar +agenda
 lark-cli im +messages-send --chat-id "oc_xxx" --text "Hello"
-lark-cli docs +create --api-version v2 --doc-format markdown --content $'<title>周报</title>\n# 本周进展\n- 完成了 X 功能'
+lark-cli docs +create --doc-format markdown --content $'<title>周报</title>\n# 本周进展\n- 完成了 X 功能'
 ```
 
 运行 `lark-cli <service> --help` 查看所有快捷命令。

cmd/event/list_test.go

@@ -26,6 +26,7 @@ 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)
@@ -55,4 +56,17 @@ 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

@@ -96,6 +96,34 @@ 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

@@ -0,0 +1,132 @@
+// 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

@@ -0,0 +1,432 @@
+// 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

@@ -27,6 +27,21 @@ 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

@@ -7,6 +7,7 @@ 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"
@@ -17,6 +18,7 @@ func init() {
 	all := [][]event.KeyDefinition{
 		im.Keys(),
 		minutes.Keys(),
+		task.Keys(),
 		vc.Keys(),
 		whiteboard.Keys(),
 	}

events/task/native.go

@@ -0,0 +1,23 @@
+// 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

@@ -0,0 +1,32 @@
+// 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

@@ -0,0 +1,119 @@
+// 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

@@ -0,0 +1,33 @@
+// 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

@@ -0,0 +1,95 @@
+// 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

@@ -1,5 +0,0 @@
-# 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

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

harness-opt/baseline/noise-floor.json

@@ -1,50 +0,0 @@
-{
-  "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

@@ -1,33 +0,0 @@
-{
-  "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

@@ -1,869 +0,0 @@
-{
-  "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

@@ -1,48 +0,0 @@
-{
-  "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

@@ -1,60 +0,0 @@
-{
-  "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

@@ -1,13 +0,0 @@
-# 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

@@ -1,12 +0,0 @@
-{
-  "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

@@ -1,35 +0,0 @@
-{
-  "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

@@ -1,35 +0,0 @@
-{
-  "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

@@ -1,35 +0,0 @@
-{
-  "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

@@ -1,35 +0,0 @@
-{
-  "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

@@ -1,67 +0,0 @@
-[
-  {
-    "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

@@ -1,24 +0,0 @@
-{
-  "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

@@ -1,29 +0,0 @@
-{
-  "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

@@ -1,97 +0,0 @@
-# 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

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

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

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

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

@@ -1,222 +0,0 @@
-{
-  "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

@@ -1,15 +0,0 @@
-{
-  "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

@@ -1,404 +0,0 @@
-{
-  "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

@@ -1,44 +0,0 @@
-# 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

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

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

@@ -1,35 +0,0 @@
-# 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

@@ -1,65 +0,0 @@
-[
-  {
-    "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

@@ -1,27 +0,0 @@
-{
-  "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

@@ -1,11 +0,0 @@
-{
-  "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

@@ -1,113 +0,0 @@
-# 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

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

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

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

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

@@ -1,220 +0,0 @@
-{
-  "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

@@ -1,15 +0,0 @@
-{
-  "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

@@ -1,380 +0,0 @@
-{
-  "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

@@ -1,48 +0,0 @@
-# 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

@@ -1,11 +0,0 @@
-[
-  {
-    "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

@@ -1,43 +0,0 @@
-# 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

@@ -1,59 +0,0 @@
-[
-  {
-    "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

@@ -1,27 +0,0 @@
-{
-  "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

@@ -1,20 +0,0 @@
-{
-  "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

@@ -1,119 +0,0 @@
-# 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

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

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

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

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

@@ -1,220 +0,0 @@
-{
-  "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

@@ -1,18 +0,0 @@
-{
-  "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

@@ -1,394 +0,0 @@
-{
-  "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

@@ -1,43 +0,0 @@
-# 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

@@ -1,20 +0,0 @@
-[
-  {
-    "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

@@ -1,44 +0,0 @@
-# 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

@@ -1,20 +0,0 @@
-[
-  {
-    "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

@@ -1,152 +0,0 @@
-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

@@ -1,334 +0,0 @@
-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

@@ -1,135 +0,0 @@
-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

@@ -131,31 +131,3 @@ 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

@@ -29,3 +29,31 @@ 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

@@ -5,7 +5,22 @@
 
 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

@@ -0,0 +1,33 @@
+// 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)
+	}
+}

internal/qualitygate/cmd/comment-audit/main.go

@@ -0,0 +1,92 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package main
+
+import (
+	"encoding/json"
+	"flag"
+	"fmt"
+	"os"
+
+	"github.com/larksuite/cli/errs"
+	"github.com/larksuite/cli/internal/qualitygate/publiccontent"
+	"github.com/larksuite/cli/internal/qualitygate/report"
+	"github.com/larksuite/cli/internal/validate"
+	"github.com/larksuite/cli/internal/vfs"
+)
+
+type eventPayload struct {
+	Comment *struct {
+		Body string `json:"body"`
+	} `json:"comment"`
+	Review *struct {
+		Body string `json:"body"`
+	} `json:"review"`
+}
+
+func main() {
+	eventPath := flag.String("event", os.Getenv("GITHUB_EVENT_PATH"), "GitHub event payload path")
+	kind := flag.String("kind", os.Getenv("GITHUB_EVENT_NAME"), "GitHub event kind")
+	flag.Parse()
+
+	if *eventPath == "" {
+		fmt.Fprintln(os.Stderr, "comment-audit: --event or GITHUB_EVENT_PATH is required")
+		os.Exit(2)
+	}
+	body, err := commentBody(*eventPath)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "comment-audit: %v\n", err)
+		os.Exit(2)
+	}
+	diags := diagnostics(publiccontent.ScanComment(*kind, body))
+	if len(diags) > 0 {
+		fmt.Fprintln(os.Stderr, auditFailureSummary(len(diags)))
+	}
+	report.Print(os.Stderr, diags)
+	os.Exit(report.ExitCode(diags))
+}
+
+func auditFailureSummary(count int) string {
+	return fmt.Sprintf("post-publication audit found public content findings: %d", count)
+}
+
+func commentBody(path string) (string, error) {
+	safePath, err := validate.SafeInputPath(path)
+	if err != nil {
+		return "", errs.NewValidationError(errs.SubtypeInvalidArgument, "invalid --event: %v", err).
+			WithParam("--event").
+			WithCause(err)
+	}
+	data, err := vfs.ReadFile(safePath)
+	if err != nil {
+		return "", err
+	}
+	var payload eventPayload
+	if err := json.Unmarshal(data, &payload); err != nil {
+		return "", err
+	}
+	switch {
+	case payload.Comment != nil:
+		return payload.Comment.Body, nil
+	case payload.Review != nil:
+		return payload.Review.Body, nil
+	default:
+		return "", nil
+	}
+}
+
+func diagnostics(items []publiccontent.Finding) []report.Diagnostic {
+	out := make([]report.Diagnostic, 0, len(items))
+	for _, item := range items {
+		out = append(out, report.Diagnostic{
+			Rule:       item.Rule,
+			Action:     item.Action,
+			File:       item.File,
+			Line:       item.Line,
+			Message:    item.Message,
+			Suggestion: item.Suggestion,
+		})
+	}
+	return out
+}

internal/qualitygate/cmd/comment-audit/main_test.go

@@ -0,0 +1,70 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package main
+
+import (
+	"errors"
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/larksuite/cli/errs"
+)
+
+func TestCommentBodyReadsSafeRelativeEventPath(t *testing.T) {
+	dir := t.TempDir()
+	if err := writeTestFile(filepath.Join(dir, "event.json"), `{"comment":{"body":"clean comment"}}`); err != nil {
+		t.Fatal(err)
+	}
+	origDir, err := os.Getwd()
+	if err != nil {
+		t.Fatal(err)
+	}
+	if err := os.Chdir(dir); err != nil {
+		t.Fatal(err)
+	}
+	t.Cleanup(func() {
+		_ = os.Chdir(origDir)
+	})
+
+	got, err := commentBody("event.json")
+	if err != nil {
+		t.Fatalf("commentBody() error = %v", err)
+	}
+	if got != "clean comment" {
+		t.Fatalf("comment body = %q", got)
+	}
+}
+
+func TestCommentBodyRejectsUnsafeEventPath(t *testing.T) {
+	path := filepath.Join(t.TempDir(), "event.json")
+	if err := writeTestFile(path, `{"comment":{"body":"clean"}}`); err != nil {
+		t.Fatal(err)
+	}
+
+	_, err := commentBody(path)
+	problem, ok := errs.ProblemOf(err)
+	if err == nil || !ok {
+		t.Fatalf("commentBody(%q) error = %v, want unsafe path validation error", path, err)
+	}
+	if problem.Category != errs.CategoryValidation || problem.Subtype != errs.SubtypeInvalidArgument {
+		t.Fatalf("commentBody(%q) problem = %#v, want invalid argument validation", path, problem)
+	}
+	var validationErr *errs.ValidationError
+	if !errors.As(err, &validationErr) || validationErr.Param != "--event" {
+		t.Fatalf("commentBody(%q) error = %v, want --event validation param", path, err)
+	}
+}
+
+func TestAuditFailureSummaryStatesPostPublicationAudit(t *testing.T) {
+	got := auditFailureSummary(2)
+	want := "post-publication audit found public content findings: 2"
+	if got != want {
+		t.Fatalf("auditFailureSummary() = %q, want %q", got, want)
+	}
+}
+
+func writeTestFile(path, data string) error {
+	return os.WriteFile(path, []byte(data), 0o644)
+}

internal/qualitygate/cmd/quality-gate/main.go

@@ -13,6 +13,7 @@ import (
 	"github.com/larksuite/cli/internal/qualitygate/manifest"
 	"github.com/larksuite/cli/internal/qualitygate/report"
 	"github.com/larksuite/cli/internal/qualitygate/rules"
+	"github.com/larksuite/cli/internal/validate"
 )
 
 func main() {
@@ -41,6 +42,7 @@ func runCheck(args []string) int {
 	fs.StringVar(&opts.FactsOut, "facts-out", "", "write facts JSON to this path")
 	fs.StringVar(&opts.ManifestPath, "manifest", "", "hand-authored command manifest JSON")
 	fs.StringVar(&opts.CommandIndexPath, "command-index", "", "full command index JSON")
+	fs.StringVar(&opts.PublicContentMetadataPath, "public-content-metadata", "", "PR title/body metadata JSON for public content checks")
 	fs.BoolVar(&printLegacyCommandCandidates, "print-legacy-command-candidates", false, "print current non-kebab-case hand-authored command candidates")
 	fs.BoolVar(&printLegacyFlagCandidates, "print-legacy-flag-candidates", false, "print current non-kebab-case flag candidates")
 	if err := fs.Parse(args); err != nil {
@@ -48,6 +50,15 @@ func runCheck(args []string) int {
 		return 2
 	}
 
+	if opts.PublicContentMetadataPath != "" {
+		safePath, err := validate.SafeInputPath(opts.PublicContentMetadataPath)
+		if err != nil {
+			fmt.Fprintf(os.Stderr, "quality-gate check: --public-content-metadata: %v\n", err)
+			return 2
+		}
+		opts.PublicContentMetadataPath = safePath
+	}
+
 	if opts.ManifestPath == "" || opts.CommandIndexPath == "" {
 		fmt.Fprintln(os.Stderr, "quality-gate check: --manifest and --command-index are required")
 		return 2

internal/qualitygate/cmd/quality-gate/main_test.go

@@ -37,6 +37,37 @@ func TestCheckRequiresManifestInputs(t *testing.T) {
 	}
 }
 
+func TestCheckAcceptsPublicContentMetadataFlag(t *testing.T) {
+	code, stderr := runCheckCaptureStderr(t, []string{
+		"--repo", t.TempDir(),
+		"--cli-bin", "./lark-cli",
+		"--public-content-metadata", ".tmp/quality-gate/pr.json",
+	})
+	if code != 2 {
+		t.Fatalf("exit code = %d, stderr=%s", code, stderr)
+	}
+	if strings.Contains(stderr, "flag provided but not defined") {
+		t.Fatalf("public content metadata flag was not registered: %s", stderr)
+	}
+	if !strings.Contains(stderr, "--manifest and --command-index are required") {
+		t.Fatalf("stderr = %s", stderr)
+	}
+}
+
+func TestCheckRejectsUnsafePublicContentMetadataPath(t *testing.T) {
+	code, stderr := runCheckCaptureStderr(t, []string{
+		"--repo", t.TempDir(),
+		"--cli-bin", "./lark-cli",
+		"--public-content-metadata", filepath.Join(t.TempDir(), "pr.json"),
+	})
+	if code != 2 {
+		t.Fatalf("exit code = %d, stderr=%s", code, stderr)
+	}
+	if !strings.Contains(stderr, "--public-content-metadata") || !strings.Contains(stderr, "--file") {
+		t.Fatalf("stderr = %s, want unsafe public content metadata path error", stderr)
+	}
+}
+
 func TestCheckReportsManifestReadErrorsWithFlagName(t *testing.T) {
 	dir := t.TempDir()
 	manifestPath := filepath.Join(dir, "command-manifest.json")

internal/qualitygate/cmd/semantic-review/main.go

@@ -56,6 +56,14 @@ func run(args []string) int {
 		_ = semantic.WriteMarkdown(markdownOut, decision)
 		return 0
 	}
+	if reviewPath == "" && !semantic.BuildInputView(f).HasReviewableFacts() {
+		decision := finalizeDecision(block, waiverDiags, semantic.Decision{})
+		if err := writeSemanticOutputs(decisionOut, markdownOut, decision); err != nil {
+			fmt.Fprintf(os.Stderr, "semantic-review: %v\n", err)
+			return 2
+		}
+		return decisionExitCode(decision)
+	}
 	review, err := semantic.LoadOrReviewWithConfig(context.Background(), f, reviewPath, modelConfig)
 	if err != nil {
 		fmt.Fprintf(os.Stderr, "semantic-review: %v\n", err)
@@ -72,6 +80,15 @@ func run(args []string) int {
 		return 0
 	}
 	decision := semantic.DecideWithWaivers(f, review, policy, waivers)
+	decision = finalizeDecision(block, waiverDiags, decision)
+	if err := writeSemanticOutputs(decisionOut, markdownOut, decision); err != nil {
+		fmt.Fprintf(os.Stderr, "semantic-review: %v\n", err)
+		return 2
+	}
+	return decisionExitCode(decision)
+}
+
+func finalizeDecision(block bool, waiverDiags []report.Diagnostic, decision semantic.Decision) semantic.Decision {
 	decision.BlockMode = block
 	if !block && len(decision.Blockers) > 0 {
 		for i := range decision.Blockers {
@@ -81,15 +98,21 @@ func run(args []string) int {
 		decision.Blockers = nil
 	}
 	decision.SystemWarnings = append(diagnosticSystemWarnings(waiverDiags), decision.SystemWarnings...)
+	return decision
+}
+
+func writeSemanticOutputs(decisionOut, markdownOut string, decision semantic.Decision) error {
 	if err := semantic.WriteDecision(decisionOut, decision); err != nil {
-		fmt.Fprintf(os.Stderr, "semantic-review: write decision: %v\n", err)
-		return 2
+		return fmt.Errorf("write decision: %w", err)
 	}
 	if err := semantic.WriteMarkdown(markdownOut, decision); err != nil {
-		fmt.Fprintf(os.Stderr, "semantic-review: write markdown: %v\n", err)
-		return 2
+		return fmt.Errorf("write markdown: %w", err)
 	}
-	if block && len(decision.Blockers) > 0 {
+	return nil
+}
+
+func decisionExitCode(decision semantic.Decision) int {
+	if decision.BlockMode && len(decision.Blockers) > 0 {
 		return 1
 	}
 	return 0

internal/qualitygate/cmd/semantic-review/main_test.go

@@ -7,6 +7,7 @@ import (
 	"encoding/json"
 	"os"
 	"path/filepath"
+	"strings"
 	"testing"
 
 	"github.com/larksuite/cli/internal/qualitygate/facts"
@@ -211,7 +212,19 @@ func TestRunWritesSkippedDecisionForUnavailableReviewer(t *testing.T) {
 	  "allowed_base_urls": ["https://ark.ap-southeast.bytepluses.com/api/v3"]
 	}`, "")
 	factsPath := filepath.Join(t.TempDir(), "facts.json")
-	if err := (facts.Facts{SchemaVersion: 1}).WriteFile(factsPath); err != nil {
+	f := facts.Facts{
+		SchemaVersion: 1,
+		Skills: []facts.SkillFact{{
+			SourceFile:               "skills/lark-wiki/SKILL.md",
+			Line:                     30,
+			Changed:                  true,
+			ReferencesInvalidCommand: true,
+		}},
+	}
+	if !semantic.BuildInputView(f).HasReviewableFacts() {
+		t.Fatal("test setup must contain reviewable facts")
+	}
+	if err := f.WriteFile(factsPath); err != nil {
 		t.Fatalf("write facts: %v", err)
 	}
 	decisionPath := filepath.Join(t.TempDir(), "decision.json")
@@ -228,6 +241,71 @@ func TestRunWritesSkippedDecisionForUnavailableReviewer(t *testing.T) {
 	}
 }
 
+func TestRunShortCircuitsEmptySemanticInputWithoutReviewer(t *testing.T) {
+	t.Setenv("ARK_API_KEY", "")
+	t.Setenv("ARK_BASE_URL", "")
+	t.Setenv("ARK_MODEL", "")
+
+	repo := t.TempDir()
+	writeSemanticConfig(t, repo, `{
+	  "schema_version": 1,
+	  "default_enforcement": "observe",
+	  "block_categories": ["skill_quality"]
+	}`, `{
+	  "allowed": ["semantic-review-v1"],
+	  "allowed_base_urls": ["https://ark.ap-southeast.bytepluses.com/api/v3"]
+	}`, "")
+	factsPath := filepath.Join(t.TempDir(), "facts.json")
+	f := facts.Facts{
+		SchemaVersion: 1,
+		Commands: []facts.CommandFact{{
+			Path:    "service command 1",
+			Domain:  "service",
+			Changed: true,
+			Source:  "service",
+		}},
+		Outputs: []facts.OutputFact{{
+			Command:          "service command 1",
+			Domain:           "service",
+			Changed:          true,
+			Source:           "service",
+			IsList:           true,
+			HasDefaultLimit:  true,
+			HasDecisionField: true,
+		}},
+	}
+	if semantic.BuildInputView(f).HasReviewableFacts() {
+		t.Fatal("test setup must not contain reviewable facts")
+	}
+	if err := f.WriteFile(factsPath); err != nil {
+		t.Fatalf("write facts: %v", err)
+	}
+	decisionPath := filepath.Join(t.TempDir(), "decision.json")
+	markdownPath := filepath.Join(t.TempDir(), "semantic.md")
+	code := run([]string{"--repo", repo, "--facts", factsPath, "--decision-out", decisionPath, "--markdown-out", markdownPath, "--block"})
+	if code != 0 {
+		t.Fatalf("run() = %d, want clean pass", code)
+	}
+	decision := readDecision(t, decisionPath)
+	if decision.Skipped || decision.Degraded || decision.InfrastructureFailure || !decision.BlockMode {
+		t.Fatalf("expected non-degraded pass decision: %#v", decision)
+	}
+	if len(decision.SystemWarnings) != 0 || len(decision.Warnings) != 0 || len(decision.Blockers) != 0 {
+		t.Fatalf("empty semantic view should not produce findings: %#v", decision)
+	}
+	data, err := os.ReadFile(markdownPath)
+	if err != nil {
+		t.Fatalf("read markdown: %v", err)
+	}
+	markdown := string(data)
+	if !strings.Contains(markdown, "No semantic blockers.") {
+		t.Fatalf("markdown missing pass summary: %s", markdown)
+	}
+	if strings.Contains(strings.ToLower(markdown), "skipped") || strings.Contains(strings.ToLower(markdown), "degraded") {
+		t.Fatalf("markdown should not report semantic review as skipped/degraded: %s", markdown)
+	}
+}
+
 func TestRunWritesInfrastructureFailureDecisionForInvalidReviewerConfig(t *testing.T) {
 	t.Setenv("ARK_API_KEY", "test-key")
 	t.Setenv("ARK_BASE_URL", "")
@@ -243,7 +321,19 @@ func TestRunWritesInfrastructureFailureDecisionForInvalidReviewerConfig(t *testi
 	  "allowed_base_urls": ["https://ark.ap-southeast.bytepluses.com/api/v3"]
 	}`, "")
 	factsPath := filepath.Join(t.TempDir(), "facts.json")
-	if err := (facts.Facts{SchemaVersion: 1}).WriteFile(factsPath); err != nil {
+	f := facts.Facts{
+		SchemaVersion: 1,
+		Skills: []facts.SkillFact{{
+			SourceFile:               "skills/lark-wiki/SKILL.md",
+			Line:                     30,
+			Changed:                  true,
+			ReferencesInvalidCommand: true,
+		}},
+	}
+	if !semantic.BuildInputView(f).HasReviewableFacts() {
+		t.Fatal("test setup must contain reviewable facts")
+	}
+	if err := f.WriteFile(factsPath); err != nil {
 		t.Fatalf("write facts: %v", err)
 	}
 	decisionPath := filepath.Join(t.TempDir(), "decision.json")

internal/qualitygate/config/semantic/policy.json

@@ -5,7 +5,8 @@
     "error_hint",
     "default_output",
     "naming",
-    "skill_quality"
+    "skill_quality",
+    "public_content_leakage"
   ],
   "rollout_groups": [
     {
@@ -16,7 +17,8 @@
       },
       "categories": [
         "error_hint",
-        "skill_quality"
+        "skill_quality",
+        "public_content_leakage"
       ],
       "owner": "cli-owner",
       "reason": "first semantic blocking rollout only affects changed facts"

internal/qualitygate/facts/schema.go

@@ -13,14 +13,15 @@ import (
 )
 
 type Facts struct {
-	SchemaVersion int                `json:"schema_version"`
-	Commands      []CommandFact      `json:"commands,omitempty"`
-	Skills        []SkillFact        `json:"skills,omitempty"`
-	SkillQuality  []SkillQualityFact `json:"skill_quality,omitempty"`
-	Errors        []ErrorFact        `json:"errors,omitempty"`
-	Outputs       []OutputFact       `json:"outputs,omitempty"`
-	Examples      []CommandExample   `json:"examples,omitempty"`
-	Diagnostics   []DiagnosticFact   `json:"diagnostics,omitempty"`
+	SchemaVersion int                 `json:"schema_version"`
+	Commands      []CommandFact       `json:"commands,omitempty"`
+	Skills        []SkillFact         `json:"skills,omitempty"`
+	SkillQuality  []SkillQualityFact  `json:"skill_quality,omitempty"`
+	Errors        []ErrorFact         `json:"errors,omitempty"`
+	Outputs       []OutputFact        `json:"outputs,omitempty"`
+	Examples      []CommandExample    `json:"examples,omitempty"`
+	PublicContent []PublicContentFact `json:"public_content,omitempty"`
+	Diagnostics   []DiagnosticFact    `json:"diagnostics,omitempty"`
 }
 
 type CommandFact struct {
@@ -109,6 +110,17 @@ type OutputFact struct {
 	HasDecisionField bool     `json:"has_decision_field,omitempty"`
 }
 
+type PublicContentFact struct {
+	Rule       string        `json:"rule"`
+	Action     report.Action `json:"action"`
+	File       string        `json:"file"`
+	Line       int           `json:"line"`
+	Source     string        `json:"source,omitempty"`
+	Excerpt    string        `json:"excerpt,omitempty"`
+	Message    string        `json:"message,omitempty"`
+	Suggestion string        `json:"suggestion,omitempty"`
+}
+
 type DryRunRequest struct {
 	Method string              `json:"method"`
 	URL    string              `json:"url"`
@@ -206,6 +218,11 @@ func BuildWithCommandLookup(m manifest.Manifest, commandLookup manifest.Manifest
 	}
 }
 
+func WithPublicContent(f Facts, publicContent []PublicContentFact) Facts {
+	f.PublicContent = publicContent
+	return f
+}
+
 type commandScope struct {
 	Domain string
 	Source string

internal/qualitygate/facts/schema_test.go

@@ -34,6 +34,7 @@ func TestFactsSchemaCarriesGatekeeperFields(t *testing.T) {
 		Errors:        []ErrorFact{{Code: "invalid_input", Message: "bad path", Hint: "pass --file", Retryable: false, HintActionCount: 1, RequiredHint: true}},
 		Outputs:       []OutputFact{{Command: "im messages list", Fields: []string{"message_id", "sender", "create_time"}, IsList: true, HasDefaultLimit: true, HasDecisionField: true}},
 		Skills:        []SkillFact{{SourceFile: "skills/lark-doc/SKILL.md", Line: 1, DestructiveWithoutGuard: true, ScopeConflict: true}},
+		PublicContent: []PublicContentFact{{Rule: "public_content_generic_credential", Action: report.ActionReject, File: "docs/public.md", Line: 4, Excerpt: "api_key = <redacted>"}},
 	}
 	data, err := json.Marshal(f)
 	if err != nil {
@@ -43,7 +44,10 @@ func TestFactsSchemaCarriesGatekeeperFields(t *testing.T) {
 	if err := json.Unmarshal(data, &got); err != nil {
 		t.Fatalf("unmarshal facts: %v", err)
 	}
-	if !got.Errors[0].RequiredHint || got.Outputs[0].Fields[0] != "message_id" || !got.Skills[0].ScopeConflict {
+	if !got.Errors[0].RequiredHint ||
+		got.Outputs[0].Fields[0] != "message_id" ||
+		!got.Skills[0].ScopeConflict ||
+		got.PublicContent[0].Rule != "public_content_generic_credential" {
 		t.Fatalf("facts lost gatekeeper fields: %#v", got)
 	}
 }

internal/qualitygate/publiccontent/collect.go

@@ -0,0 +1,343 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"sort"
+	"strconv"
+	"strings"
+)
+
+func Collect(ctx context.Context, opts Options) ([]Finding, error) {
+	metadata, err := LoadMetadata(opts.MetadataPath)
+	if err != nil {
+		return nil, err
+	}
+
+	var out []Finding
+	changedFiles, base, err := changedFiles(ctx, opts.Repo, opts.ChangedFrom)
+	if err != nil {
+		return nil, err
+	}
+	patches := map[string][]changedChunk{}
+	if base != "" {
+		patches, err = changedPatches(ctx, opts.Repo, base)
+		if err != nil {
+			return nil, err
+		}
+	}
+	for _, file := range changedFiles {
+		if !scanChangedFile(file) {
+			continue
+		}
+		for _, chunk := range patches[file] {
+			findings := scanText(file, "file", chunk.Text, isDetectorRuleFile(file))
+			for i := range findings {
+				findings[i].Line += chunk.StartLine - 1
+			}
+			out = append(out, findings...)
+			out = append(out, semanticCandidate(file, "file", chunk.Text, chunk.StartLine)...)
+		}
+		privateKeyFindings, err := scanTouchedPrivateKeyBlocks(ctx, opts.Repo, file, patches[file])
+		if err != nil {
+			return nil, err
+		}
+		out = appendUniqueFindings(out, privateKeyFindings...)
+	}
+	if base != "" {
+		commitFindings, err := scanCommitMessages(ctx, opts.Repo, base)
+		if err != nil {
+			return nil, err
+		}
+		out = append(out, commitFindings...)
+	}
+	branchName := opts.BranchName
+	if branchName == "" {
+		branchName = metadata.Branch
+	}
+	if branchName == "" {
+		branchName = branchFromEnv()
+	}
+	if branchName == "" {
+		branchName = currentBranch(ctx, opts.Repo)
+	}
+	if branchName != "" {
+		out = append(out, scanText("branch", "branch", branchName, false)...)
+	}
+	out = append(out, scanMetadata(metadata)...)
+	sort.SliceStable(out, func(i, j int) bool {
+		if out[i].File != out[j].File {
+			return out[i].File < out[j].File
+		}
+		if out[i].Line != out[j].Line {
+			return out[i].Line < out[j].Line
+		}
+		return out[i].Rule < out[j].Rule
+	})
+	return out, nil
+}
+
+func currentBranch(ctx context.Context, repo string) string {
+	data, err := gitOutput(ctx, repo, "branch", "--show-current")
+	if err != nil {
+		return ""
+	}
+	return strings.TrimSpace(string(data))
+}
+
+func branchFromEnv() string {
+	for _, key := range []string{"PR_BRANCH", "GITHUB_HEAD_REF", "GITHUB_REF_NAME"} {
+		if value := strings.TrimSpace(os.Getenv(key)); value != "" {
+			return value
+		}
+	}
+	return ""
+}
+
+func changedFiles(ctx context.Context, repo, changedFrom string) ([]string, string, error) {
+	if changedFrom == "" {
+		return nil, "", nil
+	}
+	baseBytes, err := gitOutput(ctx, repo, "merge-base", changedFrom, "HEAD")
+	if err != nil {
+		return nil, "", err
+	}
+	base := strings.TrimSpace(string(baseBytes))
+	files, err := diffFileNames(ctx, repo, base)
+	if err != nil {
+		return nil, "", err
+	}
+	sort.Strings(files)
+	return files, base, nil
+}
+
+func diffFileNames(ctx context.Context, repo, base string) ([]string, error) {
+	data, err := gitOutput(ctx, repo, "diff", "--name-only", "-z", "--diff-filter=ACMR", base+"..HEAD")
+	if err != nil {
+		return nil, err
+	}
+	var files []string
+	for _, file := range bytes.Split(data, []byte{0}) {
+		if len(file) == 0 {
+			continue
+		}
+		files = append(files, filepath.ToSlash(string(file)))
+	}
+	return files, nil
+}
+
+var detectorFixtureExclusions = map[string]bool{
+	"internal/qualitygate/publiccontent/collect_test.go": true,
+	"internal/qualitygate/publiccontent/rules.go":        true,
+	"internal/qualitygate/publiccontent/scan.go":         true,
+	"internal/qualitygate/publiccontent/scan_test.go":    true,
+}
+
+func scanChangedFile(file string) bool {
+	normalized := strings.TrimPrefix(strings.ReplaceAll(file, "\\", "/"), "./")
+	return !detectorFixtureExclusions[normalized]
+}
+
+type changedChunk struct {
+	StartLine int
+	Text      string
+}
+
+func (c changedChunk) endLine() int {
+	lines := strings.Count(strings.TrimRight(c.Text, "\n"), "\n") + 1
+	if lines < 1 {
+		lines = 1
+	}
+	return c.StartLine + lines - 1
+}
+
+func changedPatches(ctx context.Context, repo, base string) (map[string][]changedChunk, error) {
+	files, err := diffFileNames(ctx, repo, base)
+	if err != nil {
+		return nil, err
+	}
+	data, err := gitOutput(ctx, repo, "diff", "--no-ext-diff", "--unified=0", "--diff-filter=ACMR", base+"..HEAD")
+	if err != nil {
+		return nil, err
+	}
+	out := map[string][]changedChunk{}
+	var file string
+	var chunk *changedChunk
+	nextLine := 0
+	nextFile := 0
+	flush := func() {
+		if file == "" || chunk == nil || chunk.Text == "" {
+			chunk = nil
+			return
+		}
+		out[file] = append(out[file], *chunk)
+		chunk = nil
+	}
+	for _, raw := range strings.Split(string(data), "\n") {
+		switch {
+		case strings.HasPrefix(raw, "diff --git "):
+			flush()
+			file = ""
+			if nextFile < len(files) {
+				file = files[nextFile]
+				nextFile++
+			}
+		case strings.HasPrefix(raw, "@@ "):
+			flush()
+			start, ok := parseNewHunkStart(raw)
+			if !ok {
+				nextLine = 0
+				continue
+			}
+			nextLine = start
+			chunk = &changedChunk{StartLine: start}
+		case strings.HasPrefix(raw, "+") && !strings.HasPrefix(raw, "+++"):
+			if chunk == nil {
+				chunk = &changedChunk{StartLine: max(nextLine, 1)}
+			}
+			chunk.Text += strings.TrimPrefix(raw, "+") + "\n"
+			nextLine++
+		case strings.HasPrefix(raw, "-"):
+			continue
+		default:
+			if chunk != nil && strings.HasPrefix(raw, `\ No newline at end of file`) {
+				continue
+			}
+			flush()
+		}
+	}
+	flush()
+	return out, nil
+}
+
+func parseNewHunkStart(header string) (int, bool) {
+	parts := strings.Split(header, " ")
+	for _, part := range parts {
+		if !strings.HasPrefix(part, "+") {
+			continue
+		}
+		raw := strings.TrimPrefix(part, "+")
+		if before, _, ok := strings.Cut(raw, ","); ok {
+			raw = before
+		}
+		start, err := strconv.Atoi(raw)
+		return start, err == nil && start > 0
+	}
+	return 0, false
+}
+
+func scanCommitMessages(ctx context.Context, repo, base string) ([]Finding, error) {
+	data, err := gitOutput(ctx, repo, "log", "--format=%H%x00%B%x00", base+"..HEAD")
+	if err != nil {
+		return nil, err
+	}
+	parts := bytes.Split(data, []byte{0})
+	var out []Finding
+	for i := 0; i+1 < len(parts); i += 2 {
+		sha := strings.TrimSpace(string(parts[i]))
+		body := string(parts[i+1])
+		if sha == "" || body == "" {
+			continue
+		}
+		short := sha
+		if len(short) > 12 {
+			short = short[:12]
+		}
+		out = append(out, scanText("commit:"+short, "commit", body, false)...)
+		out = append(out, semanticCandidate("commit:"+short, "commit", body, 1)...)
+	}
+	return out, nil
+}
+
+type lineRange struct {
+	Start int
+	End   int
+}
+
+func scanTouchedPrivateKeyBlocks(ctx context.Context, repo, file string, chunks []changedChunk) ([]Finding, error) {
+	if len(chunks) == 0 {
+		return nil, nil
+	}
+	data, err := gitOutput(ctx, repo, "show", "HEAD:"+file)
+	if err != nil {
+		return nil, err
+	}
+	var added []lineRange
+	for _, chunk := range chunks {
+		added = append(added, lineRange{Start: chunk.StartLine, End: chunk.endLine()})
+	}
+	var out []Finding
+	for _, block := range privateKeyBlocks(string(data)) {
+		if !rangesIntersectAny(block, added) {
+			continue
+		}
+		out = append(out, newFinding("public_content_private_key_block", file, block.Start, "file", "private key block"))
+	}
+	return out, nil
+}
+
+func privateKeyBlocks(text string) []lineRange {
+	lines := strings.Split(text, "\n")
+	var out []lineRange
+	inPrivateKey := false
+	start := 0
+	for i, line := range lines {
+		lineNo := i + 1
+		if !inPrivateKey && strings.Contains(line, privateKeyBeginPrefix) && strings.Contains(line, privateKeyMarker) {
+			inPrivateKey = true
+			start = lineNo
+		}
+		if inPrivateKey && strings.Contains(line, privateKeyEndPrefix) && strings.Contains(line, privateKeyMarker) {
+			out = append(out, lineRange{Start: start, End: lineNo})
+			inPrivateKey = false
+		}
+	}
+	return out
+}
+
+func rangesIntersectAny(block lineRange, ranges []lineRange) bool {
+	for _, r := range ranges {
+		if block.Start <= r.End && r.Start <= block.End {
+			return true
+		}
+	}
+	return false
+}
+
+func appendUniqueFindings(items []Finding, additions ...Finding) []Finding {
+	for _, addition := range additions {
+		duplicate := false
+		for _, item := range items {
+			if item.Rule == addition.Rule &&
+				item.File == addition.File &&
+				item.Line == addition.Line &&
+				item.Source == addition.Source {
+				duplicate = true
+				break
+			}
+		}
+		if !duplicate {
+			items = append(items, addition)
+		}
+	}
+	return items
+}
+
+func gitOutput(ctx context.Context, repo string, args ...string) ([]byte, error) {
+	cmd := exec.CommandContext(ctx, "git", args...)
+	cmd.Dir = repo
+	var stdout, stderr bytes.Buffer
+	cmd.Stdout = &stdout
+	cmd.Stderr = &stderr
+	if err := cmd.Run(); err != nil {
+		return nil, fmt.Errorf("git %s: %w\n%s", strings.Join(args, " "), err, stderr.Bytes())
+	}
+	return stdout.Bytes(), nil
+}

internal/qualitygate/publiccontent/collect_test.go

@@ -0,0 +1,885 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import (
+	"context"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+func TestCollectScansOnlyCurrentContributionAndMetadata(t *testing.T) {
+	repo := t.TempDir()
+	runGit(t, repo, "init")
+	runGit(t, repo, "config", "user.email", "test@example.com")
+	runGit(t, repo, "config", "user.name", "Test User")
+
+	writeFile(t, filepath.Join(repo, "baseline.md"), `BASE_`+`TOKEN="baseline-only"
+`)
+	runGit(t, repo, "add", "baseline.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), `# Public change
+
+api_`+`key = "example-public-key"
+`)
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "add public doc", "-m", "Change"+"-Id: I0123456789abcdef0123456789abcdef01234567")
+
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{"title":"publish public docs","body":"Reviewed`+`-on: https://review.example.test/c/project/+/123"}`)
+
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		ChangedFrom:  "HEAD~1",
+		MetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+
+	rules := findingRules(got)
+	for _, want := range []string{
+		"public_content_generic_credential",
+		"public_content_change_id_trailer",
+		"public_content_reviewed_on_trailer",
+	} {
+		if !rules[want] {
+			t.Fatalf("missing rule %s in findings %#v", want, got)
+		}
+	}
+	for _, item := range got {
+		if item.File == "baseline.md" {
+			t.Fatalf("collector scanned unchanged baseline file: %#v", got)
+		}
+	}
+}
+
+func TestCollectScansOnlyChangedLinesInChangedFiles(t *testing.T) {
+	repo := t.TempDir()
+	runGit(t, repo, "init")
+	runGit(t, repo, "config", "user.email", "test@example.com")
+	runGit(t, repo, "config", "user.name", "Test User")
+
+	writeFile(t, filepath.Join(repo, "docs", "workflow.md"), "SECRET_TOKEN=legacy-example\npublic baseline\n")
+	runGit(t, repo, "add", "docs/workflow.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "workflow.md"), "SECRET_TOKEN=legacy-example\npublic baseline\nnew public line\n")
+	runGit(t, repo, "add", "docs/workflow.md")
+	runGit(t, repo, "commit", "-m", "add public line")
+
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{}`)
+
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		ChangedFrom:  "HEAD~1",
+		MetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+	for _, item := range got {
+		if item.Rule == "public_content_generic_credential" && item.File == "docs/workflow.md" {
+			t.Fatalf("collector scanned unchanged legacy line in changed file: %#v", got)
+		}
+	}
+}
+
+func TestCollectSemanticCandidatesStoreSanitizedReviewText(t *testing.T) {
+	repo := t.TempDir()
+	runGit(t, repo, "init")
+	runGit(t, repo, "config", "user.email", "test@example.com")
+	runGit(t, repo, "config", "user.name", "Test User")
+
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), "base\n")
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	raw := "private launch plan for alpha-service rollout on Friday with SERVICE_" + "TOKEN=real-" + "secret-value"
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), "base\n"+raw+"\n")
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "add semantic candidate")
+
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{}`)
+
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		ChangedFrom:  "HEAD~1",
+		MetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+	var found bool
+	for _, item := range got {
+		if item.Rule != "public_content_semantic_candidate" || item.File != "docs/public.md" {
+			continue
+		}
+		found = true
+		if !strings.Contains(item.Excerpt, "alpha-service rollout on Friday") {
+			t.Fatalf("semantic candidate should include sanitized review text, got %#v", item)
+		}
+		if strings.Contains(item.Excerpt, "real-"+"secret-value") {
+			t.Fatalf("semantic candidate leaked credential value: %#v", item)
+		}
+		if !strings.Contains(item.Excerpt, "SERVICE_TOKEN=<redacted>") {
+			t.Fatalf("semantic candidate should redact credentials in review text, got %#v", item)
+		}
+		if !strings.Contains(item.Excerpt, "semantic signals") || !strings.Contains(item.Excerpt, "roadmap_timing") {
+			t.Fatalf("semantic candidate excerpt should preserve semantic signals, got %#v", item)
+		}
+	}
+	if !found {
+		t.Fatalf("missing semantic candidate in findings %#v", got)
+	}
+}
+
+func TestCollectSemanticCandidatesDoNotLeakWhitespaceCredentialTail(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), "base\n")
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	raw := "private launch plan for internal rollout on Friday with SERVICE_" + "TOKEN=\"real " + "secret value\""
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), "base\n"+raw+"\n")
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "add semantic candidate")
+
+	got := collectFromPreviousCommit(t, repo)
+	for _, item := range got {
+		if item.Rule != "public_content_semantic_candidate" || item.File != "docs/public.md" {
+			continue
+		}
+		if strings.Contains(item.Excerpt, "secret value") || strings.Contains(item.Excerpt, "real "+"secret value") {
+			t.Fatalf("semantic candidate leaked credential tail: %#v", item)
+		}
+		if !strings.Contains(item.Excerpt, "SERVICE_TOKEN=<redacted>") {
+			t.Fatalf("semantic candidate should redact full credential assignment, got %#v", item)
+		}
+		return
+	}
+	t.Fatalf("missing semantic candidate in findings %#v", got)
+}
+
+func TestCollectJSONBearerHeadersDoNotLeakIntoSemanticCandidates(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), "base\n")
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	token := "abcdefghijklmnopqrstuvwxyz"
+	raw := "private launch plan for internal rollout on Friday with " +
+		`{"headers":{"Authorization":"Bearer ` + token + `"}}`
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), "base\n"+raw+"\n")
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "add json bearer")
+
+	got := collectFromPreviousCommit(t, repo)
+	requireFinding(t, got, "docs/public.md", "public_content_bearer_header")
+	for _, item := range got {
+		if item.File != "docs/public.md" {
+			continue
+		}
+		if strings.Contains(item.Excerpt, token) {
+			t.Fatalf("finding leaked JSON bearer token: %#v", item)
+		}
+	}
+}
+
+func TestCollectDetectsQuotedJSONCredentialAssignments(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "public.json"), "{}\n")
+	runGit(t, repo, "add", "docs/public.json")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "public.json"), strings.Join([]string{
+		`{"access_` + `token":"real-json-token"}`,
+		`{"client_` + `secret": "real ` + `secret value"}`,
+		`{"tenantAccess` + `Token":"real-tenant-camel-token"}`,
+		`{"github` + `Token":"real-github-token"}`,
+		`{"vendorApi` + `Key":"real-vendor-key"}`,
+		`{"slackBot` + `Token":"xoxb-real-token"}`,
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/public.json")
+	runGit(t, repo, "commit", "-m", "add json config")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File == "docs/public.json" && item.Rule == "public_content_generic_credential" {
+			count++
+			for _, forbidden := range []string{
+				"real-json-token",
+				"real secret value",
+				"real-tenant-camel-token",
+				"real-github-token",
+				"real-vendor-key",
+				"xoxb-real-token",
+			} {
+				if strings.Contains(item.Excerpt, forbidden) {
+					t.Fatalf("JSON credential finding leaked value %q in excerpt %q", forbidden, item.Excerpt)
+				}
+			}
+		}
+	}
+	if count != 6 {
+		t.Fatalf("JSON credential findings = %d, want 6: %#v", count, got)
+	}
+}
+
+func TestCollectAllowsBenignJSONTokenFields(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "public.json"), "{}\n")
+	runGit(t, repo, "add", "docs/public.json")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "public.json"), strings.Join([]string{
+		`{"tokenizer":"cl100k_base"}`,
+		`{"token_count": 42}`,
+		`{"page_token":"next"}`,
+		`{"next_page_token":"next"}`,
+		`{"file_token":"file-example"}`,
+		`{"doc_token":"doc-example"}`,
+		`{"node_token":"node-example"}`,
+		`{"wiki_token":"wikcn_public_doc_example"}`,
+		`{"folder_token":"folder-example"}`,
+		`{"obj_token":"obj-example"}`,
+		`{"spreadsheet_token":"sheet-example"}`,
+		`{"parent_node_token":"parent-example"}`,
+		`{"origin_node_token":"origin-example"}`,
+		`{"drive_route_token":"route-example"}`,
+		`{"token":"<wiki_token>"}`,
+		`{"token":"wiki_token"}`,
+		`{"token_url":"https://example.com/oauth/token"}`,
+		`{"token_endpoint":"https://example.com/oauth/token"}`,
+		`{"token_format":"Bearer"}`,
+		`{"secret_name":"public-example-secret"}`,
+		`{"base_token":"base-example"}`,
+		`{"app_token":"app-example"}`,
+		`{"sync_token":"sync-example"}`,
+		`{"parent_token":"parent-example"}`,
+		`{"target_token":"target-example"}`,
+		`{"parent_file_token":"parent-file-example"}`,
+		`{"refresh_token_expires_in": 7200}`,
+		`{"access_token_expires_in": 7200}`,
+		`{"token_expires_in": 7200}`,
+		`{"token_status":"active"}`,
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/public.json")
+	runGit(t, repo, "commit", "-m", "add benign json token fields")
+
+	got := collectFromPreviousCommit(t, repo)
+	for _, item := range got {
+		if item.File == "docs/public.json" && item.Rule == "public_content_generic_credential" {
+			t.Fatalf("benign JSON token field should not be credential finding: %#v", got)
+		}
+	}
+}
+
+func TestCollectDetectsAngleWrappedRealisticCredentialValues(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), "base: true\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "base")
+	stripeLike := "sk_" + "live_1234567890abcdef"
+	patLike := "gh" + "p_1234567890abcdef1234567890abcdef1234"
+
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), strings.Join([]string{
+		"API_KEY: <" + stripeLike + ">",
+		"SECRET_TOKEN: <" + patLike + ">",
+		"CLIENT_SECRET: <real-client-secret-value>",
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "add credential config")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File == "docs/config.yaml" && item.Rule == "public_content_generic_credential" {
+			count++
+		}
+	}
+	if count != 3 {
+		t.Fatalf("angle-wrapped realistic credential findings = %d, want 3: %#v", count, got)
+	}
+}
+
+func TestCollectDetectsCredentialShapedValuesUnderBenignKeys(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "public.json"), "{}\n")
+	runGit(t, repo, "add", "docs/public.json")
+	runGit(t, repo, "commit", "-m", "base")
+	stripeLike := "sk_" + "live_1234567890abcdef"
+	patLike := "gh" + "p_1234567890abcdef1234567890abcdef1234"
+
+	writeFile(t, filepath.Join(repo, "docs", "public.json"), strings.Join([]string{
+		`{"access_token_expires_in":"` + patLike + `"}`,
+		`{"refresh_token_expires_in":"` + stripeLike + `"}`,
+		`{"client_secret_status":"real-client-secret-value"}`,
+		`{"client_secret_name":"real-client-secret-value"}`,
+		`{"app_token":"` + patLike + `"}`,
+		`{"sync_token":"` + stripeLike + `"}`,
+		`{"target_token":"real-client-secret-value"}`,
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/public.json")
+	runGit(t, repo, "commit", "-m", "add credential-shaped benign fields")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File == "docs/public.json" && item.Rule == "public_content_generic_credential" {
+			count++
+		}
+	}
+	if count != 7 {
+		t.Fatalf("credential-shaped benign-key findings = %d, want 7: %#v", count, got)
+	}
+}
+
+func TestCollectDetectsBareIdentifierCredentialsWithMetadataSuffixes(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), "base: true\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), strings.Join([]string{
+		"API_KEY_NAME: prod_key",
+		"CLIENT_SECRET_NAME: prod_secret",
+		"SECRET_STATUS: prod_secret",
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "add credential config")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File == "docs/config.yaml" && item.Rule == "public_content_generic_credential" {
+			count++
+		}
+	}
+	if count != 3 {
+		t.Fatalf("metadata-suffixed bare credential findings = %d, want 3: %#v", count, got)
+	}
+}
+
+func TestCollectDetectsAccessKeyCredentials(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), "base: true\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "base")
+	accessKey := "AK" + "IAIOSFODNN7EXAMPX"
+
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), strings.Join([]string{
+		"AWS_ACCESS_KEY_ID: " + accessKey,
+		"ACCESS_KEY_ID: " + accessKey,
+		"ACCESS_KEY: " + accessKey,
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "add access key config")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File != "docs/config.yaml" || item.Rule != "public_content_generic_credential" {
+			continue
+		}
+		count++
+		if strings.Contains(item.Excerpt, "AKIAIOSFODNN7EXAMPX") {
+			t.Fatalf("access key finding leaked value in excerpt %q", item.Excerpt)
+		}
+	}
+	if count != 3 {
+		t.Fatalf("access key credential findings = %d, want 3: %#v", count, got)
+	}
+}
+
+func TestCollectDetectsPrivateKeyAssignments(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), "base: true\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "base")
+
+	privateKey := "LS0tLS1CRUdJTiBQUklWQVRFIEtFWS0tLS0t"
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), strings.Join([]string{
+		"PRIVATE_KEY: " + privateKey,
+		"SSH_PRIVATE_KEY: " + privateKey,
+		"JWT_PRIVATE_KEY: " + privateKey,
+		"SIGNING_PRIVATE_KEY: " + privateKey,
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "add private key config")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File != "docs/config.yaml" || item.Rule != "public_content_generic_credential" {
+			continue
+		}
+		count++
+		if strings.Contains(item.Excerpt, privateKey) {
+			t.Fatalf("private key finding leaked value in excerpt %q", item.Excerpt)
+		}
+	}
+	if count != 4 {
+		t.Fatalf("private key assignment findings = %d, want 4: %#v", count, got)
+	}
+}
+
+func TestCollectDetectsCredentialValuesThatLookLikeBareIdentifiers(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), "base: true\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), strings.Join([]string{
+		"API_KEY_OPENAI: prod_key",
+		"CLIENT_SECRET_GOOGLE: prod_secret",
+		"TOKEN_GITHUB: github_token",
+		"APP_PASSWORD_PROD: prod_password",
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "add credential config")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File == "docs/config.yaml" && item.Rule == "public_content_generic_credential" {
+			count++
+		}
+	}
+	if count != 4 {
+		t.Fatalf("bare identifier credential findings = %d, want 4: %#v", count, got)
+	}
+}
+
+func TestCollectAllowsBenignUnquotedTokenFields(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), "base: true\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), strings.Join([]string{
+		"tokens: 128",
+		"token_type: bearer",
+		"max_tokens: 2000",
+		"completion_tokens: 200",
+		"prompt_tokens: 100",
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "add benign token config")
+
+	got := collectFromPreviousCommit(t, repo)
+	for _, item := range got {
+		if item.File == "docs/config.yaml" && item.Rule == "public_content_generic_credential" {
+			t.Fatalf("benign unquoted token field should not be credential finding: %#v", got)
+		}
+	}
+}
+
+func TestCollectDetectsCredentialPhraseBeforeEnvironmentSuffix(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), "base: true\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "config.yaml"), strings.Join([]string{
+		"API_KEY_OPENAI: real-openai-key",
+		"TOKEN_GITHUB: real-github-token",
+		"CLIENT_SECRET_GOOGLE: real-google-secret",
+		"SECRET_KEY_BASE: real-secret-key-base",
+		"APP_PASSWORD_PROD: real-prod-password",
+	}, "\n")+"\n")
+	runGit(t, repo, "add", "docs/config.yaml")
+	runGit(t, repo, "commit", "-m", "add credential config")
+
+	got := collectFromPreviousCommit(t, repo)
+	var count int
+	for _, item := range got {
+		if item.File != "docs/config.yaml" || item.Rule != "public_content_generic_credential" {
+			continue
+		}
+		count++
+		for _, forbidden := range []string{
+			"real-openai-key",
+			"real-github-token",
+			"real-google-secret",
+			"real-secret-key-base",
+			"real-prod-password",
+		} {
+			if strings.Contains(item.Excerpt, forbidden) {
+				t.Fatalf("credential finding leaked value %q in excerpt %q", forbidden, item.Excerpt)
+			}
+		}
+	}
+	if count != 5 {
+		t.Fatalf("credential suffix variants findings = %d, want 5: %#v", count, got)
+	}
+}
+
+func TestCollectDetectsPrivateKeyWhenOnlyEndIsAdded(t *testing.T) {
+	repo := newGitRepo(t)
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+"legacy-body\n")
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+"legacy-body\nnew-body\n"+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "complete key")
+
+	got := collectFromPreviousCommit(t, repo)
+	requireFinding(t, got, "docs/key.pem", "public_content_private_key_block")
+}
+
+func TestCollectDetectsPrivateKeyWhenOnlyBeginIsAdded(t *testing.T) {
+	repo := newGitRepo(t)
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), "legacy-body\n"+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+"legacy-body\n"+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "complete key")
+
+	got := collectFromPreviousCommit(t, repo)
+	requireFinding(t, got, "docs/key.pem", "public_content_private_key_block")
+}
+
+func TestCollectDetectsPrivateKeyWhenOnlyBodyIsAdded(t *testing.T) {
+	repo := newGitRepo(t)
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+"new-body\n"+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "add body")
+
+	got := collectFromPreviousCommit(t, repo)
+	requireFinding(t, got, "docs/key.pem", "public_content_private_key_block")
+}
+
+func TestCollectIgnoresUntouchedHistoricalPrivateKey(t *testing.T) {
+	repo := newGitRepo(t)
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+"legacy-body\n"+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+"legacy-body\n"+privateKeyEnd())
+	writeFile(t, filepath.Join(repo, "docs", "public.md"), "public docs update\n")
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "docs update")
+
+	got := collectFromPreviousCommit(t, repo)
+	for _, item := range got {
+		if item.File == "docs/key.pem" && item.Rule == "public_content_private_key_block" {
+			t.Fatalf("collector reported untouched historical private key: %#v", got)
+		}
+	}
+}
+
+func TestCollectIgnoresDeletedPrivateKeyLine(t *testing.T) {
+	repo := newGitRepo(t)
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+"legacy-body\n"+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "key.pem"), privateKeyBegin()+privateKeyEnd())
+	runGit(t, repo, "add", "docs/key.pem")
+	runGit(t, repo, "commit", "-m", "remove body")
+
+	got := collectFromPreviousCommit(t, repo)
+	for _, item := range got {
+		if item.File == "docs/key.pem" && item.Rule == "public_content_private_key_block" {
+			t.Fatalf("collector reported delete-only private key cleanup: %#v", got)
+		}
+	}
+}
+
+func TestCollectSkipsOnlyKnownQualityGateFixtureFiles(t *testing.T) {
+	repo := t.TempDir()
+	runGit(t, repo, "init")
+	runGit(t, repo, "config", "user.email", "test@example.com")
+	runGit(t, repo, "config", "user.name", "Test User")
+
+	writeFile(t, filepath.Join(repo, "README.md"), "base\n")
+	runGit(t, repo, "add", "README.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "internal", "qualitygate", "publiccontent", "collect_test.go"), "SECRET_TOKEN=fixture\n")
+	writeFile(t, filepath.Join(repo, "internal", "qualitygate", "publiccontent", "scan_test.go"), "SECRET_TOKEN=fixture\n")
+	writeFile(t, filepath.Join(repo, "internal", "qualitygate", "publiccontent", "scan.go"), "const privateKeyFixture = \""+privateKeyBeginPrefix+privateKeyMarker+"\"\n")
+	writeFile(t, filepath.Join(repo, "internal", "qualitygate", "publiccontent", "rules.go"), "markers := []string{\"generated with automation\"}\n")
+	writeFile(t, filepath.Join(repo, "tests", "e2e", "new-public-workflow.test.sh"), "SECRET_TOKEN=real-leak\n")
+	runGit(t, repo, "add", ".")
+	runGit(t, repo, "commit", "-m", "add scanner fixtures")
+
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{}`)
+
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		ChangedFrom:  "HEAD~1",
+		MetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+	var foundOrdinaryTestLeak bool
+	for _, item := range got {
+		switch item.File {
+		case "internal/qualitygate/publiccontent/collect_test.go",
+			"internal/qualitygate/publiccontent/scan.go",
+			"internal/qualitygate/publiccontent/scan_test.go",
+			"internal/qualitygate/publiccontent/rules.go":
+			t.Fatalf("collector scanned known fixture or detector implementation file: %#v", got)
+		}
+		if item.File == "tests/e2e/new-public-workflow.test.sh" && item.Rule == "public_content_generic_credential" {
+			foundOrdinaryTestLeak = true
+		}
+	}
+	if !foundOrdinaryTestLeak {
+		t.Fatalf("collector should still scan ordinary test files for real leaks: %#v", got)
+	}
+}
+
+func TestScanChangedFileDocumentsFixtureExclusions(t *testing.T) {
+	excluded := []string{
+		"internal/qualitygate/publiccontent/collect_test.go",
+		"internal/qualitygate/publiccontent/rules.go",
+		"internal/qualitygate/publiccontent/scan.go",
+		"internal/qualitygate/publiccontent/scan_test.go",
+	}
+	for _, file := range excluded {
+		if scanChangedFile(file) {
+			t.Fatalf("scanChangedFile(%q) = true, want false for detector fixture/implementation path", file)
+		}
+	}
+
+	included := []string{
+		"internal/qualitygate/publiccontent/new_test.go",
+		"tests/e2e/new-public-workflow.test.sh",
+		"docs/public.md",
+	}
+	for _, file := range included {
+		if !scanChangedFile(file) {
+			t.Fatalf("scanChangedFile(%q) = false, want true", file)
+		}
+	}
+}
+
+func TestCollectScansAddedLinesInSpecialPathNames(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "old.md"), "base\n")
+	runGit(t, repo, "add", ".")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "has space.md"), "SECRET_TOKEN=space-value\n")
+	writeFile(t, filepath.Join(repo, `weird"quote.md`), "SECRET_TOKEN=quote-value\n")
+	runGit(t, repo, "mv", "docs/old.md", "docs/new name.md")
+	writeFile(t, filepath.Join(repo, "docs", "new name.md"), "base\nSECRET_TOKEN=rename-value\n")
+	runGit(t, repo, "add", ".")
+	runGit(t, repo, "commit", "-m", "add special paths")
+
+	got := collectFromPreviousCommit(t, repo)
+	requireFinding(t, got, "docs/has space.md", "public_content_generic_credential")
+	requireFinding(t, got, `weird"quote.md`, "public_content_generic_credential")
+	requireFinding(t, got, "docs/new name.md", "public_content_generic_credential")
+}
+
+func TestCollectScansBranchNameAsWarning(t *testing.T) {
+	repo := t.TempDir()
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{"branch":"bot/public-doc-update"}`)
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		MetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+	if len(got) != 1 || got[0].Rule != "public_content_automation_branch" {
+		t.Fatalf("branch findings = %#v", got)
+	}
+}
+
+func TestCollectUsesExplicitBranchNameWhenDetached(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "README.md"), "base\n")
+	runGit(t, repo, "add", "README.md")
+	runGit(t, repo, "commit", "-m", "base")
+	runGit(t, repo, "checkout", "-b", "bot/public-doc-update")
+	writeFile(t, filepath.Join(repo, "docs.md"), "safe docs\n")
+	runGit(t, repo, "add", "docs.md")
+	runGit(t, repo, "commit", "-m", "docs")
+	head := strings.TrimSpace(string(runGitOutput(t, repo, "rev-parse", "HEAD")))
+	runGit(t, repo, "checkout", "--detach", head)
+
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{}`)
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		MetadataPath: metadataPath,
+		BranchName:   "bot/public-doc-update",
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+	requireFinding(t, got, "branch", "public_content_automation_branch")
+}
+
+func TestCollectUsesBranchEnvironmentWhenDetached(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "README.md"), "base\n")
+	runGit(t, repo, "add", "README.md")
+	runGit(t, repo, "commit", "-m", "base")
+	runGit(t, repo, "checkout", "-b", "bot/public-env-update")
+	writeFile(t, filepath.Join(repo, "docs.md"), "safe docs\n")
+	runGit(t, repo, "add", "docs.md")
+	runGit(t, repo, "commit", "-m", "docs")
+	head := strings.TrimSpace(string(runGitOutput(t, repo, "rev-parse", "HEAD")))
+	runGit(t, repo, "checkout", "--detach", head)
+	t.Setenv("GITHUB_HEAD_REF", "bot/public-env-update")
+
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{}`)
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		MetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+	requireFinding(t, got, "branch", "public_content_automation_branch")
+}
+
+func TestCollectPreservesFindingAttributionForChangedLines(t *testing.T) {
+	repo := newGitRepo(t)
+	writeFile(t, filepath.Join(repo, "docs", "auth.md"), "intro\n")
+	runGit(t, repo, "add", "docs/auth.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	writeFile(t, filepath.Join(repo, "docs", "auth.md"), "intro\nAuthorization: Bearer abcdefghijklmnopqrstuvwxyz\n")
+	runGit(t, repo, "add", "docs/auth.md")
+	runGit(t, repo, "commit", "-m", "add auth docs")
+
+	got := collectFromPreviousCommit(t, repo)
+	for _, item := range got {
+		if item.Rule == "public_content_bearer_header" {
+			if item.File != "docs/auth.md" || item.Line != 2 || item.Source != "file" {
+				t.Fatalf("changed-line attribution = %#v", item)
+			}
+			return
+		}
+	}
+	t.Fatalf("missing bearer finding: %#v", got)
+}
+
+func TestAppendUniqueFindingsDeduplicatesByRuleFileLineAndSource(t *testing.T) {
+	base := []Finding{newFinding("public_content_private_key_block", "docs/key.pem", 1, "file", "private key block")}
+	got := appendUniqueFindings(base,
+		newFinding("public_content_private_key_block", "docs/key.pem", 1, "file", "private key block"),
+		newFinding("public_content_private_key_block", "docs/key.pem", 2, "file", "private key block"),
+	)
+	if len(got) != 2 {
+		t.Fatalf("appendUniqueFindings len = %d, want 2: %#v", len(got), got)
+	}
+}
+
+func newGitRepo(t *testing.T) string {
+	t.Helper()
+	repo := t.TempDir()
+	runGit(t, repo, "init")
+	runGit(t, repo, "config", "user.email", "test@example.com")
+	runGit(t, repo, "config", "user.name", "Test User")
+	return repo
+}
+
+func privateKeyBegin() string {
+	return privateKeyBeginPrefix + privateKeyMarker + "\n"
+}
+
+func privateKeyEnd() string {
+	return privateKeyEndPrefix + privateKeyMarker + "\n"
+}
+
+func collectFromPreviousCommit(t *testing.T, repo string) []Finding {
+	t.Helper()
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{}`)
+	got, err := Collect(context.Background(), Options{
+		Repo:         repo,
+		ChangedFrom:  "HEAD~1",
+		MetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Collect() error = %v", err)
+	}
+	return got
+}
+
+func requireFinding(t *testing.T, got []Finding, file, rule string) {
+	t.Helper()
+	for _, item := range got {
+		if item.File == file && item.Rule == rule {
+			return
+		}
+	}
+	t.Fatalf("missing %s in %s findings: %#v", rule, file, got)
+}
+
+func TestCollectRequiresValidMetadataJSON(t *testing.T) {
+	repo := t.TempDir()
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	writeFile(t, metadataPath, `{"title":`)
+
+	_, err := Collect(context.Background(), Options{Repo: repo, MetadataPath: metadataPath})
+	if err == nil || !strings.Contains(err.Error(), "public content metadata") {
+		t.Fatalf("Collect() error = %v, want metadata parse error", err)
+	}
+}
+
+func runGit(t *testing.T, repo string, args ...string) {
+	t.Helper()
+	if len(args) > 0 && args[0] == "commit" {
+		args = append([]string{"commit", "--no-verify"}, args[1:]...)
+	}
+	cmd := exec.Command("git", args...)
+	cmd.Dir = repo
+	out, err := cmd.CombinedOutput()
+	if err != nil {
+		t.Fatalf("git %v failed: %v\n%s", args, err, out)
+	}
+}
+
+func runGitOutput(t *testing.T, repo string, args ...string) []byte {
+	t.Helper()
+	cmd := exec.Command("git", args...)
+	cmd.Dir = repo
+	out, err := cmd.CombinedOutput()
+	if err != nil {
+		t.Fatalf("git %v failed: %v\n%s", args, err, out)
+	}
+	return out
+}
+
+func writeFile(t *testing.T, path, data string) {
+	t.Helper()
+	if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(path, []byte(data), 0o644); err != nil {
+		t.Fatal(err)
+	}
+}

internal/qualitygate/publiccontent/comment_audit.go

@@ -0,0 +1,11 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+func ScanComment(kind, body string) []Finding {
+	if kind == "" {
+		kind = "comment"
+	}
+	return scanText(kind, "comment", body, false)
+}

internal/qualitygate/publiccontent/comment_audit_test.go

@@ -0,0 +1,19 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import "testing"
+
+func TestScanCommentAuditsPublishedCommentBodies(t *testing.T) {
+	got := ScanComment("issue_comment", `The published comment included /tmp/harness`+`-agent/run and CCM`+`-Harness: stage-4`)
+	rules := findingRules(got)
+	if !rules["public_content_harness_metadata"] || !rules["public_content_ccm_harness_trailer"] {
+		t.Fatalf("comment audit findings = %#v", got)
+	}
+	for _, item := range got {
+		if item.File != "issue_comment" {
+			t.Fatalf("comment finding file = %q, want issue_comment", item.File)
+		}
+	}
+}

internal/qualitygate/publiccontent/metadata.go

@@ -0,0 +1,45 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import (
+	"encoding/json"
+	"fmt"
+
+	"github.com/larksuite/cli/internal/vfs"
+)
+
+func LoadMetadata(path string) (Metadata, error) {
+	if path == "" {
+		return Metadata{}, nil
+	}
+	data, err := vfs.ReadFile(path)
+	if err != nil {
+		return Metadata{}, fmt.Errorf("public content metadata: %w", err)
+	}
+	if len(data) == 0 {
+		return Metadata{}, nil
+	}
+	var out Metadata
+	if err := json.Unmarshal(data, &out); err != nil {
+		return Metadata{}, fmt.Errorf("public content metadata: %w", err)
+	}
+	return out, nil
+}
+
+func scanMetadata(m Metadata) []Finding {
+	text := ""
+	if m.Title != "" {
+		text += "title: " + m.Title + "\n"
+	}
+	if m.Body != "" {
+		text += "body:\n" + m.Body + "\n"
+	}
+	if text == "" {
+		return nil
+	}
+	out := scanText("pull_request_metadata", "metadata", text, false)
+	out = append(out, semanticCandidate("pull_request_metadata", "metadata", text, 1)...)
+	return out
+}

internal/qualitygate/publiccontent/metadata_test.go

@@ -0,0 +1,22 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import (
+	"path/filepath"
+	"testing"
+)
+
+func TestLoadMetadataReadsTitleAndBody(t *testing.T) {
+	path := filepath.Join(t.TempDir(), "metadata.json")
+	writeFile(t, path, `{"title":"public change","body":"pass`+`word = \"example-password\""}`)
+
+	got, err := LoadMetadata(path)
+	if err != nil {
+		t.Fatalf("LoadMetadata() error = %v", err)
+	}
+	if got.Title != "public change" || got.Body == "" {
+		t.Fatalf("metadata = %#v", got)
+	}
+}

internal/qualitygate/publiccontent/rules.go

@@ -0,0 +1,441 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import (
+	"net/url"
+	"path/filepath"
+	"regexp"
+	"strings"
+
+	"github.com/larksuite/cli/internal/qualitygate/report"
+)
+
+var (
+	credentialAssignmentRE = regexp.MustCompile(`(?i)["']?\b[A-Za-z0-9_-]*(?:api[_-]?key|access[_-]?key|private[_-]?key|secret|password|passwd|token|webhook|access[_-]?token|client[_-]?secret)[A-Za-z0-9_-]*\b["']?\s*[:=]\s*(?:"((?:\\.|[^"\\])*)"|'((?:\\.|[^'\\])*)'|(\$\([^)]*\))|(\$\{\{[^}]+\}\})|([^"'\s,}\]]+))`)
+	jwtLikeRE              = regexp.MustCompile(`\b[A-Za-z0-9_-]{10,}\.[A-Za-z0-9_-]{10,}\.[A-Za-z0-9_-]{10,}\b`)
+	credentialURLRE        = regexp.MustCompile(`(?i)\b[a-z][a-z0-9+.-]*://[^/\s:@]*:[^@\s/]+@[^)\s]+`)
+	bearerHeaderRE         = regexp.MustCompile(`(?i)(?:\bAuthorization\s*:\s*Bearer\s+|["']Authorization["']\s*:\s*["']Bearer\s+)[A-Za-z0-9._+/=-]{12,}`)
+	semanticBearerHeaderRE = regexp.MustCompile(`(?i)(?:\bAuthorization\s*:\s*Bearer\s+[^"'\s,}\]]+|["']Authorization["']\s*:\s*["']Bearer\s+[^"'\\\s,}\]]+)`)
+	changeIDTrailerRE      = regexp.MustCompile(`(?i)^\s*Change-Id:\s*\S+`)
+	reviewedOnTrailerRE    = regexp.MustCompile(`(?i)^\s*Reviewed-on:\s*\S+`)
+	ccmHarnessTrailerRE    = regexp.MustCompile(`(?i)\bCCM-Harness:\s*\S+`)
+	privateIPv4RE          = regexp.MustCompile(`\b(?:10\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}|192\.168\.[0-9]{1,3}\.[0-9]{1,3}|172\.(?:1[6-9]|2[0-9]|3[0-1])\.[0-9]{1,3}\.[0-9]{1,3})\b`)
+	automationBranchRE     = regexp.MustCompile(`(?i)(^|/)(bot|automation)[-/]`)
+)
+
+func actionForRule(rule string) report.Action {
+	switch rule {
+	case "public_content_generic_credential",
+		"public_content_private_key_block",
+		"public_content_jwt_like_token",
+		"public_content_bearer_header",
+		"public_content_credential_url",
+		"public_content_change_id_trailer",
+		"public_content_reviewed_on_trailer",
+		"public_content_provenance_marker",
+		"public_content_detector_fingerprint",
+		"public_content_harness_metadata",
+		"public_content_ccm_harness_trailer":
+		return report.ActionReject
+	case "public_content_private_ipv4",
+		"public_content_automation_branch":
+		return report.ActionWarning
+	default:
+		return report.ActionWarning
+	}
+}
+
+func isPlaceholderValue(value string) bool {
+	trimmed := strings.Trim(value, `"'`)
+	normalized := strings.ToLower(trimmed)
+	if normalized == "" ||
+		normalized == "=" ||
+		percentWrappedPlaceholder(normalized) ||
+		angleWrappedPlaceholder(normalized) ||
+		urlWithAnglePlaceholder(normalized) ||
+		isCredentialReferenceValue(trimmed) {
+		return true
+	}
+	return namedPlaceholderValue(normalized)
+}
+
+func namedPlaceholderValue(value string) bool {
+	switch value {
+	case "...", "placeholder", "redacted", "<redacted>", "xxxx", "test-secret":
+		return true
+	}
+	return strings.Contains(value, "cli_example") || allXPlaceholder(value)
+}
+
+func allXPlaceholder(value string) bool {
+	if len(value) < 4 {
+		return false
+	}
+	for _, r := range value {
+		if r != 'x' {
+			return false
+		}
+	}
+	return true
+}
+
+func urlWithAnglePlaceholder(value string) bool {
+	if !strings.Contains(value, "://") ||
+		!strings.Contains(value, "<") ||
+		!strings.Contains(value, ">") {
+		return false
+	}
+	return !urlRemainderLooksCredentialLike(removeAnglePlaceholders(value))
+}
+
+func removeAnglePlaceholders(value string) string {
+	var out strings.Builder
+	for len(value) > 0 {
+		start := strings.Index(value, "<")
+		if start < 0 {
+			out.WriteString(value)
+			break
+		}
+		out.WriteString(value[:start])
+		end := strings.Index(value[start+1:], ">")
+		if end < 0 {
+			out.WriteString(value[start:])
+			break
+		}
+		value = value[start+end+2:]
+	}
+	return out.String()
+}
+
+func urlRemainderLooksCredentialLike(value string) bool {
+	normalized := strings.ToLower(value)
+	for _, marker := range []string{
+		"secret",
+		"token",
+		"password",
+		"passwd",
+		"api_key",
+		"apikey",
+		"private_key",
+		"privatekey",
+		"client_secret",
+		"clientsecret",
+	} {
+		if strings.Contains(normalized, marker) {
+			return true
+		}
+	}
+	for _, part := range strings.FieldsFunc(normalized, func(r rune) bool {
+		return !((r >= 'a' && r <= 'z') || (r >= '0' && r <= '9') || r == '_' || r == '-')
+	}) {
+		if credentialShapedIdentifier(part) || longCredentialSegment(part) {
+			return true
+		}
+	}
+	return false
+}
+
+func longCredentialSegment(value string) bool {
+	if len(value) < 16 {
+		return false
+	}
+	var hasLetter, hasDigit bool
+	for _, r := range value {
+		switch {
+		case r >= 'a' && r <= 'z':
+			hasLetter = true
+		case r >= '0' && r <= '9':
+			hasDigit = true
+		case r == '_' || r == '-':
+		default:
+			return false
+		}
+	}
+	return hasLetter || hasDigit
+}
+
+func isCredentialReferenceValue(value string) bool {
+	normalized := strings.ToLower(value)
+	switch {
+	case strings.HasPrefix(normalized, "${{"):
+		return githubExpressionReference(normalized)
+	case strings.HasPrefix(normalized, "$("):
+		return !commandSubstitutionLooksCredentialLike(normalized)
+	case strings.HasPrefix(normalized, "process.env."):
+		return credentialReferenceIdentifier(strings.TrimPrefix(normalized, "process.env."))
+	case strings.HasPrefix(normalized, "${"):
+		return credentialReferenceIdentifier(strings.TrimSuffix(strings.TrimPrefix(normalized, "${"), "}"))
+	case strings.HasPrefix(value, "$"):
+		return credentialReferenceIdentifier(strings.TrimPrefix(normalized, "$"))
+	default:
+		return false
+	}
+}
+
+func commandSubstitutionLooksCredentialLike(value string) bool {
+	if !strings.HasPrefix(value, "$(") || !strings.HasSuffix(value, ")") {
+		return false
+	}
+	inner := strings.TrimSuffix(strings.TrimPrefix(value, "$("), ")")
+	for _, part := range strings.FieldsFunc(inner, func(r rune) bool {
+		return !((r >= 'a' && r <= 'z') || (r >= '0' && r <= '9') || r == '_' || r == '-')
+	}) {
+		if credentialShapedIdentifier(part) || longCredentialSegment(part) {
+			return true
+		}
+	}
+	return false
+}
+
+func githubExpressionReference(value string) bool {
+	if !strings.HasPrefix(value, "${{") || !strings.HasSuffix(value, "}}") {
+		return false
+	}
+	expr := strings.TrimSpace(strings.TrimSuffix(strings.TrimPrefix(value, "${{"), "}}"))
+	switch {
+	case strings.HasPrefix(expr, "secrets."):
+		return dottedReferenceIdentifier(strings.TrimPrefix(expr, "secrets."))
+	case strings.HasPrefix(expr, "env."):
+		return dottedReferenceIdentifier(strings.TrimPrefix(expr, "env."))
+	case strings.HasPrefix(expr, "vars."):
+		return dottedReferenceIdentifier(strings.TrimPrefix(expr, "vars."))
+	case expr == "github.token":
+		return true
+	default:
+		return false
+	}
+}
+
+func dottedReferenceIdentifier(value string) bool {
+	if value == "" {
+		return false
+	}
+	for _, part := range strings.Split(value, ".") {
+		if !referenceIdentifier(part) {
+			return false
+		}
+	}
+	return true
+}
+
+func credentialReferenceIdentifier(value string) bool {
+	return referenceIdentifier(value) && !credentialShapedIdentifier(value)
+}
+
+func referenceIdentifier(value string) bool {
+	if value == "" {
+		return false
+	}
+	for i, r := range value {
+		switch {
+		case r >= 'a' && r <= 'z':
+		case r >= '0' && r <= '9' && i > 0:
+		case r == '_' && i > 0:
+		default:
+			return false
+		}
+	}
+	return true
+}
+
+func angleWrappedPlaceholder(value string) bool {
+	if len(value) < 3 || !strings.HasPrefix(value, "<") || !strings.HasSuffix(value, ">") {
+		return false
+	}
+	return anglePlaceholderIdentifier(strings.Trim(value, "<>"))
+}
+
+func percentWrappedPlaceholder(value string) bool {
+	if len(value) < 3 || !strings.HasPrefix(value, "%") || !strings.HasSuffix(value, "%") {
+		return false
+	}
+	inner := strings.Trim(value, "%")
+	return delimitedPlaceholderIdentifier(inner) && !credentialShapedIdentifier(inner)
+}
+
+func delimitedPlaceholderIdentifier(value string) bool {
+	if value == "" {
+		return false
+	}
+	for _, r := range value {
+		if (r >= 'a' && r <= 'z') || (r >= '0' && r <= '9') || r == '_' || r == '-' {
+			continue
+		}
+		return false
+	}
+	return true
+}
+
+func anglePlaceholderIdentifier(value string) bool {
+	if value == "" {
+		return false
+	}
+	for _, r := range value {
+		if (r >= 'a' && r <= 'z') || (r >= '0' && r <= '9') || r == '_' || r == '-' {
+			continue
+		}
+		return false
+	}
+	if credentialShapedIdentifier(value) {
+		return false
+	}
+	switch value {
+	case "token",
+		"id",
+		"userid",
+		"openid",
+		"key",
+		"secret",
+		"password",
+		"api-key",
+		"user-id",
+		"open-id",
+		"client-secret",
+		"access-token",
+		"refresh-token",
+		"auth-token",
+		"bearer-token",
+		"session-token",
+		"service-token":
+		return true
+	}
+	for _, suffix := range []string{"_token", "_id", "_key", "_secret", "_password"} {
+		if strings.HasSuffix(value, suffix) {
+			return true
+		}
+	}
+	for _, suffix := range []string{"-token", "-id", "-key", "-secret", "-password"} {
+		if strings.HasSuffix(value, suffix) {
+			return true
+		}
+	}
+	return false
+}
+
+func credentialShapedValue(value string) bool {
+	normalized := strings.ToLower(strings.Trim(value, `"'<>`))
+	return credentialShapedIdentifier(normalized)
+}
+
+func credentialShapedIdentifier(value string) bool {
+	switch {
+	case strings.HasPrefix(value, "sk_live_"),
+		strings.HasPrefix(value, "sk_test_"),
+		strings.HasPrefix(value, "ghp_"),
+		strings.HasPrefix(value, "gho_"),
+		strings.HasPrefix(value, "ghu_"),
+		strings.HasPrefix(value, "github_pat_"),
+		strings.HasPrefix(value, "xoxb_"),
+		strings.HasPrefix(value, "xoxp_"),
+		strings.HasPrefix(value, "xoxa_"):
+		return true
+	case strings.HasPrefix(value, "real-") &&
+		(strings.Contains(value, "secret") ||
+			strings.Contains(value, "token") ||
+			strings.Contains(value, "key") ||
+			strings.Contains(value, "password")):
+		return true
+	default:
+		return false
+	}
+}
+
+func resourceTokenPlaceholderValue(value string) bool {
+	normalized := strings.ToLower(strings.Trim(value, `"'`))
+	switch normalized {
+	case "wiki_token",
+		"folder_token",
+		"obj_token",
+		"spreadsheet_token",
+		"file_token",
+		"doc_token",
+		"node_token",
+		"parent_node_token",
+		"origin_node_token",
+		"drive_route_token":
+		return true
+	default:
+		return minuteTokenFixturePlaceholder(normalized)
+	}
+}
+
+func minuteTokenFixturePlaceholder(value string) bool {
+	if value == "minute_no_meta" {
+		return true
+	}
+	suffix, ok := strings.CutPrefix(value, "minute_")
+	if !ok || suffix == "" {
+		return false
+	}
+	for _, r := range suffix {
+		if r < '0' || r > '9' {
+			return false
+		}
+	}
+	return true
+}
+
+func provenanceMarker(line string) bool {
+	normalized := strings.ToLower(line)
+	markers := []string{
+		"generat" + "ed by tool",
+		"creat" + "ed by tool",
+		"generat" + "ed by automation",
+		"creat" + "ed by automation",
+		"machine-" + "generated",
+		"generated with automated",
+		"generated with automation",
+		"🤖 generated",
+	}
+	for _, marker := range markers {
+		if strings.Contains(normalized, marker) {
+			return true
+		}
+	}
+	if strings.HasPrefix(normalized, "co-authored-by:") &&
+		(strings.Contains(normalized, "<bot@") ||
+			strings.Contains(normalized, " bot@") ||
+			strings.Contains(normalized, "[bot]") ||
+			strings.Contains(normalized, "automation") ||
+			strings.Contains(normalized, "automated-code-assistant")) {
+		return true
+	}
+	return false
+}
+
+// Detector fingerprint checks are intentionally scoped to public rule/config
+// files. They do not try to hide this package's implementation; they prevent
+// publishing reusable detector identifiers in external-facing rule bundles.
+func isDetectorRuleFile(path string) bool {
+	normalized := filepath.ToSlash(path)
+	base := filepath.Base(normalized)
+	return base == ".gitleaks.toml" ||
+		strings.Contains(normalized, "public-rules/") ||
+		strings.Contains(normalized, "public_rules/")
+}
+
+func detectorFingerprint(line string) bool {
+	normalized := strings.ToLower(line)
+	fingerprints := []string{
+		strings.Join([]string{"public", "content", "leakage"}, "-"),
+		strings.Join([]string{"public", "content", "detector"}, "-"),
+		"publiccontent",
+	}
+	for _, fingerprint := range fingerprints {
+		if strings.Contains(normalized, fingerprint) {
+			return true
+		}
+	}
+	return false
+}
+
+func redactCredentialURL(raw string) string {
+	u, err := url.Parse(raw)
+	if err != nil || u.User == nil {
+		return "<credential-url>"
+	}
+	u.User = url.UserPassword("<user>", "<redacted>")
+	return u.String()
+}

internal/qualitygate/publiccontent/scan.go

@@ -0,0 +1,797 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import (
+	"fmt"
+	"path/filepath"
+	"sort"
+	"strings"
+	"unicode"
+)
+
+const (
+	privateKeyBeginPrefix = "-----" + "BEGIN "
+	privateKeyEndPrefix   = "-----" + "END "
+	privateKeyMarker      = "PRIVATE " + "KEY-----"
+)
+
+func ScanFile(path string, data []byte) []Finding {
+	return scanText(filepath.ToSlash(path), "file", string(data), isDetectorRuleFile(path))
+}
+
+func semanticCandidate(file, source, text string, line int) []Finding {
+	excerpt := redactedSemanticExcerpt(text)
+	if excerpt == "" {
+		return nil
+	}
+	return []Finding{newFinding("public_content_semantic_candidate", file, line, source, excerpt)}
+}
+
+func scanText(file, source, text string, detectorFile bool) []Finding {
+	var out []Finding
+	lines := strings.Split(text, "\n")
+	inPrivateKey := false
+	privateKeyLine := 0
+	for i, line := range lines {
+		lineNo := i + 1
+		if strings.Contains(line, privateKeyBeginPrefix) && strings.Contains(line, privateKeyMarker) {
+			inPrivateKey = true
+			privateKeyLine = lineNo
+		}
+		if inPrivateKey && strings.Contains(line, privateKeyEndPrefix) && strings.Contains(line, privateKeyMarker) {
+			out = append(out, newFinding("public_content_private_key_block", file, privateKeyLine, source, "private key block"))
+			inPrivateKey = false
+		}
+		for _, match := range credentialAssignmentRE.FindAllStringSubmatch(line, -1) {
+			if !isCredentialAssignmentMatch(match[0]) {
+				continue
+			}
+			value := credentialAssignmentValue(match)
+			keyName, _ := normalizedCredentialAssignmentKey(match[0])
+			if value == "" ||
+				isNonSecretLiteralValue(value) ||
+				isBenignCodeCredentialExpression(file, value) ||
+				isPlaceholderValue(value) ||
+				isResourceTokenPlaceholderAssignment(keyName, value) {
+				continue
+			}
+			if looksLikeEqualityComparison(value) {
+				continue
+			}
+			out = append(out, newFinding("public_content_generic_credential", file, lineNo, source, redactAssignment(match[0])))
+		}
+		for _, match := range jwtLikeRE.FindAllString(line, -1) {
+			if isSchemaDottedIdentifier(line, match) {
+				continue
+			}
+			out = append(out, newFinding("public_content_jwt_like_token", file, lineNo, source, redactToken(match)))
+		}
+		for range bearerHeaderRE.FindAllString(line, -1) {
+			out = append(out, newFinding("public_content_bearer_header", file, lineNo, source, "Authorization: Bearer <redacted>"))
+		}
+		for _, match := range credentialURLRE.FindAllString(line, -1) {
+			if isPlaceholderCredentialURL(match) {
+				continue
+			}
+			out = append(out, newFinding("public_content_credential_url", file, lineNo, source, redactCredentialURL(match)))
+		}
+		for _, match := range privateIPv4RE.FindAllString(line, -1) {
+			out = append(out, newFinding("public_content_private_ipv4", file, lineNo, source, match))
+		}
+		if source == "branch" && automationBranchRE.MatchString(line) {
+			out = append(out, newFinding("public_content_automation_branch", file, lineNo, source, "automation branch marker"))
+		}
+		switch {
+		case changeIDTrailerRE.MatchString(line):
+			out = append(out, newFinding("public_content_change_id_trailer", file, lineNo, source, "Change-Id: <redacted>"))
+		case reviewedOnTrailerRE.MatchString(line):
+			out = append(out, newFinding("public_content_reviewed_on_trailer", file, lineNo, source, "Reviewed-on: <redacted>"))
+		case ccmHarnessTrailerRE.MatchString(line):
+			out = append(out, newFinding("public_content_ccm_harness_trailer", file, lineNo, source, "CCM-Harness: <redacted>"))
+		}
+		if provenanceMarker(line) {
+			out = append(out, newFinding("public_content_provenance_marker", file, lineNo, source, "provenance marker"))
+		}
+		if strings.Contains(line, "/tmp/harness-agent") {
+			out = append(out, newFinding("public_content_harness_metadata", file, lineNo, source, "/tmp/harness-agent"))
+		}
+		if detectorFile && detectorFingerprint(line) {
+			out = append(out, newFinding("public_content_detector_fingerprint", file, lineNo, source, "public detector fingerprint"))
+		}
+	}
+	sort.SliceStable(out, func(i, j int) bool {
+		if out[i].File != out[j].File {
+			return out[i].File < out[j].File
+		}
+		if out[i].Line != out[j].Line {
+			return out[i].Line < out[j].Line
+		}
+		return out[i].Rule < out[j].Rule
+	})
+	return out
+}
+
+func isCredentialAssignmentMatch(match string) bool {
+	name, value, ok := normalizedCredentialAssignment(match)
+	if !ok {
+		return false
+	}
+	if isWebhookCredentialKey(name) && webhookAssignmentValueLooksCredentialLike(value) {
+		return true
+	}
+	if isBenignTokenField(name) && !credentialShapedValue(value) {
+		return false
+	}
+	return isExplicitCredentialKey(name)
+}
+
+func normalizedCredentialAssignmentKey(match string) (string, bool) {
+	key, _, ok := normalizedCredentialAssignment(match)
+	return key, ok
+}
+
+func normalizedCredentialAssignment(match string) (string, string, bool) {
+	key, ok := credentialAssignmentKey(match)
+	if !ok {
+		return "", "", false
+	}
+	key = strings.TrimSpace(key)
+	if key == "" {
+		return "", "", false
+	}
+	submatches := credentialAssignmentRE.FindStringSubmatch(match)
+	return normalizedCredentialKey(strings.Trim(key, `"'`)), credentialAssignmentValue(submatches), true
+}
+
+func normalizedCredentialKey(key string) string {
+	key = strings.TrimSpace(key)
+	var out []rune
+	var prev rune
+	for i, r := range key {
+		if r == '-' {
+			r = '_'
+		}
+		if i > 0 && isCredentialKeyBoundary(prev, r) {
+			out = append(out, '_')
+		}
+		out = append(out, unicode.ToLower(r))
+		prev = r
+	}
+	key = string(out)
+	key = strings.ReplaceAll(key, "-", "_")
+	return key
+}
+
+func isCredentialKeyBoundary(prev, current rune) bool {
+	if prev == '_' || current == '_' {
+		return false
+	}
+	return (unicode.IsLower(prev) || unicode.IsDigit(prev)) && unicode.IsUpper(current)
+}
+
+func isBenignTokenField(key string) bool {
+	if isTokenMetricField(key) ||
+		isTokenMetadataField(key) ||
+		isResourceTokenField(key) ||
+		isPaginationOrSyncTokenField(key) {
+		return true
+	}
+	return false
+}
+
+func isTokenMetricField(key string) bool {
+	switch key {
+	case "tokenizer",
+		"token_count",
+		"tokens",
+		"max_tokens",
+		"completion_tokens",
+		"prompt_tokens":
+		return true
+	default:
+		return false
+	}
+}
+
+func isTokenMetadataField(key string) bool {
+	switch key {
+	case "access_token_expires_in",
+		"refresh_token_expires_in",
+		"token_expires_in",
+		"token_status",
+		"token_type",
+		"token_url",
+		"token_endpoint",
+		"token_format",
+		"secret_name":
+		return true
+	default:
+		return false
+	}
+}
+
+func isPaginationOrSyncTokenField(key string) bool {
+	switch key {
+	case "page_token",
+		"next_page_token",
+		"sync_token":
+		return true
+	default:
+		return false
+	}
+}
+
+func isResourceTokenField(key string) bool {
+	if !strings.HasSuffix(key, "_token") {
+		return false
+	}
+	prefix := strings.TrimSuffix(key, "_token")
+	switch prefix {
+	case "app",
+		"base",
+		"board",
+		"doc",
+		"drive_route",
+		"file",
+		"folder",
+		"host_node",
+		"minute",
+		"node",
+		"obj",
+		"origin_node",
+		"parent",
+		"parent_file",
+		"parent_node",
+		"share",
+		"spreadsheet",
+		"target",
+		"wiki":
+		return true
+	default:
+		return false
+	}
+}
+
+func isResourceTokenPlaceholderAssignment(key, value string) bool {
+	switch {
+	case key == "client_token" && idempotencyTokenPlaceholderValue(value):
+		return true
+	case key == "retry_without_token" && numericStringPlaceholderValue(value):
+		return true
+	case tokenLikePlaceholderKey(key):
+		return tokenLikePlaceholderValue(value)
+	default:
+		return false
+	}
+}
+
+func tokenLikePlaceholderKey(key string) bool {
+	return key == "token" ||
+		strings.HasSuffix(key, "_token") ||
+		strings.HasSuffix(key, "-token")
+}
+
+func tokenLikePlaceholderValue(value string) bool {
+	normalized := strings.ToLower(strings.Trim(value, `"'`))
+	if normalized == "" || credentialShapedIdentifier(normalized) {
+		return false
+	}
+	return resourceTokenPlaceholderValue(value) ||
+		isPlaceholderValue(value) ||
+		normalized == "token" ||
+		strings.Contains(normalized, "...") ||
+		strings.Contains(normalized, "xxx") ||
+		strings.Contains(normalized, "_or_") ||
+		strings.HasSuffix(normalized, "_token") ||
+		strings.HasPrefix(normalized, ".")
+}
+
+func idempotencyTokenPlaceholderValue(value string) bool {
+	return numericStringPlaceholderValue(value) || uuidStringPlaceholderValue(value)
+}
+
+func uuidStringPlaceholderValue(value string) bool {
+	normalized := strings.Trim(value, `"'`)
+	parts := strings.Split(normalized, "-")
+	if len(parts) != 5 {
+		return false
+	}
+	for i, part := range parts {
+		want := []int{8, 4, 4, 4, 12}[i]
+		if len(part) != want {
+			return false
+		}
+		for _, r := range part {
+			if (r >= '0' && r <= '9') ||
+				(r >= 'a' && r <= 'f') ||
+				(r >= 'A' && r <= 'F') {
+				continue
+			}
+			return false
+		}
+	}
+	return true
+}
+
+func numericStringPlaceholderValue(value string) bool {
+	normalized := strings.Trim(value, `"'`)
+	if normalized == "" {
+		return false
+	}
+	for _, r := range normalized {
+		if r < '0' || r > '9' {
+			return false
+		}
+	}
+	return true
+}
+
+func isBenignCodeCredentialExpression(file, value string) bool {
+	normalized := strings.TrimSpace(value)
+	if strings.HasPrefix(normalized, "regexp.MustCompile(") {
+		return true
+	}
+	if !sourceCodeFile(file) || quotedLiteral(value) || credentialShapedValue(value) {
+		return false
+	}
+	return codeReferenceExpression(normalized)
+}
+
+func sourceCodeFile(file string) bool {
+	switch filepath.Ext(file) {
+	case ".go", ".py":
+		return true
+	default:
+		return false
+	}
+}
+
+func quotedLiteral(value string) bool {
+	normalized := strings.TrimSpace(value)
+	return len(normalized) >= 2 &&
+		((strings.HasPrefix(normalized, `"`) && strings.HasSuffix(normalized, `"`)) ||
+			(strings.HasPrefix(normalized, `'`) && strings.HasSuffix(normalized, `'`)))
+}
+
+func codeReferenceExpression(value string) bool {
+	if value == "" {
+		return false
+	}
+	for _, marker := range []string{".", "(", ")", "[", "]", "{"} {
+		if strings.Contains(value, marker) {
+			return true
+		}
+	}
+	return codeIdentifier(value) && !credentialNameFragment(value)
+}
+
+func codeIdentifier(value string) bool {
+	for i, r := range value {
+		switch {
+		case r >= 'a' && r <= 'z':
+		case r >= 'A' && r <= 'Z':
+		case r == '_' && i > 0:
+		case r >= '0' && r <= '9' && i > 0:
+		default:
+			return false
+		}
+	}
+	return true
+}
+
+func credentialNameFragment(value string) bool {
+	normalized := strings.ToLower(value)
+	for _, marker := range []string{"secret", "token", "password", "passwd", "key"} {
+		if strings.Contains(normalized, marker) {
+			return true
+		}
+	}
+	return false
+}
+
+func isSchemaDottedIdentifier(line, match string) bool {
+	return strings.Contains(line, "schema ") && strings.Contains(match, "_")
+}
+
+func isNonSecretLiteralValue(value string) bool {
+	switch strings.ToLower(strings.TrimSpace(strings.Trim(value, `"'`))) {
+	case "true", "false", "null", "nil", "{", "[":
+		return true
+	default:
+		return false
+	}
+}
+
+func isWebhookCredentialKey(key string) bool {
+	return strings.Contains(strings.ReplaceAll(key, "_", ""), "webhook")
+}
+
+func webhookAssignmentValueLooksCredentialLike(value string) bool {
+	normalized := strings.ToLower(strings.Trim(value, `"'`))
+	if normalized == "" || isPlaceholderValue(normalized) || isNonSecretLiteralValue(normalized) {
+		return false
+	}
+	return urlRemainderLooksCredentialLike(removeAnglePlaceholders(normalized)) ||
+		credentialShapedIdentifier(strings.Trim(normalized, "$"))
+}
+
+func isExplicitCredentialKey(key string) bool {
+	compact := strings.ReplaceAll(key, "_", "")
+	switch compact {
+	case "token",
+		"accesstoken",
+		"refreshtoken",
+		"authtoken",
+		"bearertoken",
+		"sessiontoken",
+		"servicetoken",
+		"apikey",
+		"accesskey",
+		"privatekey",
+		"apisecret",
+		"secret",
+		"secretkey",
+		"clientsecret",
+		"password",
+		"passwd":
+		return true
+	}
+	for _, phrase := range []string{
+		"accesstoken",
+		"refreshtoken",
+		"authtoken",
+		"bearertoken",
+		"sessiontoken",
+		"servicetoken",
+		"bottoken",
+		"apikey",
+		"accesskey",
+		"privatekey",
+		"apisecret",
+		"clientsecret",
+		"secretkey",
+	} {
+		if strings.Contains(compact, phrase) {
+			return true
+		}
+	}
+	parts := credentialKeyParts(key)
+	for _, phrase := range [][2]string{
+		{"access", "token"},
+		{"refresh", "token"},
+		{"auth", "token"},
+		{"bearer", "token"},
+		{"session", "token"},
+		{"service", "token"},
+		{"bot", "token"},
+		{"api", "key"},
+		{"access", "key"},
+		{"private", "key"},
+		{"api", "secret"},
+		{"client", "secret"},
+		{"secret", "key"},
+	} {
+		if hasAdjacentCredentialParts(parts, phrase[0], phrase[1]) {
+			return true
+		}
+	}
+	for _, part := range parts {
+		switch part {
+		case "token", "secret", "password", "passwd":
+			return true
+		}
+	}
+	for _, suffix := range []string{
+		"token",
+		"accesstoken",
+		"refreshtoken",
+		"authtoken",
+		"bearertoken",
+		"sessiontoken",
+		"servicetoken",
+		"bottoken",
+		"apikey",
+		"accesskey",
+		"privatekey",
+		"apisecret",
+		"clientsecret",
+		"secret",
+		"secretkey",
+		"password",
+		"passwd",
+	} {
+		if strings.HasSuffix(compact, suffix) {
+			return true
+		}
+	}
+	for _, suffix := range []string{
+		"_access_token",
+		"_refresh_token",
+		"_auth_token",
+		"_bearer_token",
+		"_session_token",
+		"_service_token",
+		"_api_key",
+		"_access_key",
+		"_private_key",
+		"_api_secret",
+		"_client_secret",
+		"_secret",
+		"_secret_key",
+		"_password",
+		"_passwd",
+	} {
+		if strings.HasSuffix(key, suffix) {
+			return true
+		}
+	}
+	return false
+}
+
+func credentialKeyParts(key string) []string {
+	var parts []string
+	for _, part := range strings.Split(key, "_") {
+		if part != "" {
+			parts = append(parts, part)
+		}
+	}
+	return parts
+}
+
+func hasAdjacentCredentialParts(parts []string, first, second string) bool {
+	for i := 0; i+1 < len(parts); i++ {
+		if parts[i] == first && parts[i+1] == second {
+			return true
+		}
+	}
+	return false
+}
+
+func credentialAssignmentValue(match []string) string {
+	for _, value := range match[1:] {
+		if value != "" {
+			return value
+		}
+	}
+	return ""
+}
+
+func looksLikeEqualityComparison(value string) bool {
+	return strings.HasPrefix(strings.TrimSpace(value), "=")
+}
+
+func isPlaceholderCredentialURL(raw string) bool {
+	userInfo, ok := credentialURLUserInfo(raw)
+	if !ok {
+		return false
+	}
+	_, password, ok := strings.Cut(userInfo, ":")
+	if !ok {
+		return false
+	}
+	return credentialURLPasswordPlaceholder(password)
+}
+
+func credentialURLPasswordPlaceholder(password string) bool {
+	normalized := strings.ToLower(password)
+	decoded := strings.ReplaceAll(normalized, "%3c", "<")
+	decoded = strings.ReplaceAll(decoded, "%3e", ">")
+	switch decoded {
+	case "placeholder", "redacted", "<redacted>", "xxxx":
+		return true
+	}
+	return angleWrappedPlaceholder(decoded) || percentWrappedPlaceholder(decoded)
+}
+
+func credentialURLUserInfo(raw string) (string, bool) {
+	schemeIdx := strings.Index(raw, "://")
+	if schemeIdx < 0 {
+		return "", false
+	}
+	rest := raw[schemeIdx+len("://"):]
+	atIdx := strings.Index(rest, "@")
+	if atIdx < 0 {
+		return "", false
+	}
+	return rest[:atIdx], true
+}
+
+func newFinding(rule, file string, line int, source, excerpt string) Finding {
+	return Finding{
+		Rule:       rule,
+		Action:     actionForRule(rule),
+		File:       file,
+		Line:       line,
+		Source:     source,
+		Excerpt:    excerpt,
+		Message:    messageForRule(rule),
+		Suggestion: suggestionForRule(rule),
+	}
+}
+
+func messageForRule(rule string) string {
+	switch rule {
+	case "public_content_generic_credential":
+		return "public contribution contains a generic credential assignment"
+	case "public_content_private_key_block":
+		return "public contribution contains a private key block"
+	case "public_content_jwt_like_token":
+		return "public contribution contains a JWT-like token"
+	case "public_content_bearer_header":
+		return "public contribution contains an Authorization bearer token"
+	case "public_content_credential_url":
+		return "public contribution contains credentials embedded in a URL"
+	case "public_content_private_ipv4":
+		return "public contribution contains a private-network IP address"
+	case "public_content_automation_branch":
+		return "public contribution uses an automation-shaped branch name"
+	case "public_content_change_id_trailer":
+		return "public contribution contains a Change-Id trailer"
+	case "public_content_reviewed_on_trailer":
+		return "public contribution contains a Reviewed-on trailer"
+	case "public_content_provenance_marker":
+		return "public contribution contains a prohibited provenance marker"
+	case "public_content_detector_fingerprint":
+		return "public rule/config content exposes public detector fingerprints"
+	case "public_content_harness_metadata":
+		return "public contribution contains visible harness pipeline metadata"
+	case "public_content_ccm_harness_trailer":
+		return "public contribution contains a CCM-Harness trailer"
+	case "public_content_semantic_candidate":
+		return "public contribution contains text for semantic public content review"
+	default:
+		return "public contribution contains content that should not be published"
+	}
+}
+
+func suggestionForRule(rule string) string {
+	switch actionForRule(rule) {
+	case "REJECT":
+		return "remove the value from the public contribution and replace it with a non-sensitive placeholder"
+	default:
+		return "remove private workflow metadata before publishing the public contribution"
+	}
+}
+
+func redactAssignment(match string) string {
+	key, ok := credentialAssignmentKey(match)
+	if !ok {
+		return "<credential-assignment>"
+	}
+	return fmt.Sprintf("%s= <redacted>", strings.TrimSpace(key))
+}
+
+func credentialAssignmentKey(match string) (string, bool) {
+	idx := -1
+	for _, sep := range []string{":", "="} {
+		if candidate := strings.Index(match, sep); candidate >= 0 && (idx < 0 || candidate < idx) {
+			idx = candidate
+		}
+	}
+	if idx < 0 {
+		return "", false
+	}
+	return match[:idx], true
+}
+
+func redactToken(_ string) string {
+	return "<jwt-like-token>"
+}
+
+func redactedSemanticExcerpt(text string) string {
+	normalized := strings.Join(strings.Fields(text), " ")
+	if normalized == "" {
+		return ""
+	}
+	signals := semanticSignals(normalized)
+	if len(signals) == 0 {
+		return ""
+	}
+	sanitized := truncateRunes(sanitizeSemanticExcerpt(text), 600)
+	return fmt.Sprintf("semantic signals: %s; excerpt: %q", strings.Join(signals, ","), sanitized)
+}
+
+func semanticSignals(normalized string) []string {
+	lower := strings.ToLower(normalized)
+	var signals []string
+	add := func(signal string) {
+		for _, existing := range signals {
+			if existing == signal {
+				return
+			}
+		}
+		signals = append(signals, signal)
+	}
+
+	hasPrivateScope := strings.Contains(lower, "private") || strings.Contains(lower, "internal-only")
+	hasRequestMetadata := strings.Contains(lower, "request header") || strings.Contains(lower, "request headers") || strings.Contains(lower, "authorization header") || strings.Contains(lower, "metadata header")
+	hasTrustBoundary := strings.Contains(lower, "spoof") || strings.Contains(lower, "trust") || strings.Contains(lower, "risk scoring") || strings.Contains(lower, "classification")
+	hasRoadmap := strings.Contains(lower, "roadmap") || strings.Contains(lower, "migration") || strings.Contains(lower, "rollout") || strings.Contains(lower, "cutover") || strings.Contains(lower, "unpublished")
+	hasTiming := strings.Contains(lower, "target date") || strings.Contains(lower, "friday") || strings.Contains(lower, "monday") || strings.Contains(lower, "tuesday") || strings.Contains(lower, "wednesday") || strings.Contains(lower, "thursday") || strings.Contains(lower, "customer-visible")
+	hasImplementation := strings.Contains(lower, "server-side") || strings.Contains(lower, "implementation")
+
+	if hasPrivateScope && hasRequestMetadata && hasTrustBoundary {
+		add("private_scope")
+		add("request_metadata")
+		add("trust_boundary_detail")
+	}
+	if hasRoadmap && (hasPrivateScope || hasTiming) {
+		add("roadmap_detail")
+		if hasPrivateScope {
+			add("private_scope")
+		}
+		if hasTiming {
+			add("roadmap_timing")
+		}
+	}
+	if hasPrivateScope && hasImplementation && hasTrustBoundary {
+		add("private_scope")
+		add("implementation_detail")
+		add("trust_boundary_detail")
+	}
+
+	return signals
+}
+
+func sanitizeSemanticExcerpt(text string) string {
+	text = redactPrivateKeyBlocks(text)
+	text = credentialAssignmentRE.ReplaceAllStringFunc(text, sanitizeCredentialAssignment)
+	text = strings.ReplaceAll(text, `<redacted>"`, `<redacted>`)
+	text = strings.ReplaceAll(text, `<redacted>'`, `<redacted>`)
+	text = semanticBearerHeaderRE.ReplaceAllString(text, "Authorization: Bearer <redacted>")
+	text = jwtLikeRE.ReplaceAllString(text, "<jwt-like-token>")
+	text = credentialURLRE.ReplaceAllStringFunc(text, sanitizeCredentialURL)
+	return strings.Join(strings.Fields(text), " ")
+}
+
+func redactPrivateKeyBlocks(text string) string {
+	lines := strings.Split(text, "\n")
+	var out []string
+	inPrivateKey := false
+	for _, line := range lines {
+		if strings.Contains(line, privateKeyBeginPrefix) && strings.Contains(line, privateKeyMarker) {
+			out = append(out, "<private-key-block>")
+			inPrivateKey = true
+			if strings.Contains(line, privateKeyEndPrefix) && strings.Contains(line, privateKeyMarker) {
+				inPrivateKey = false
+			}
+			continue
+		}
+		if inPrivateKey {
+			if strings.Contains(line, privateKeyEndPrefix) && strings.Contains(line, privateKeyMarker) {
+				inPrivateKey = false
+			}
+			continue
+		}
+		out = append(out, line)
+	}
+	return strings.Join(out, "\n")
+}
+
+func sanitizeCredentialAssignment(match string) string {
+	key, ok := credentialAssignmentKey(match)
+	if !ok {
+		return "<credential-assignment>"
+	}
+	return strings.TrimSpace(key) + "=<redacted>"
+}
+
+func sanitizeCredentialURL(raw string) string {
+	redacted := redactCredentialURL(raw)
+	redacted = strings.ReplaceAll(redacted, "%3Cuser%3E", "<user>")
+	redacted = strings.ReplaceAll(redacted, "%3Credacted%3E", "<redacted>")
+	return redacted
+}
+
+func truncateRunes(text string, limit int) string {
+	if limit <= 0 {
+		return ""
+	}
+	runes := []rune(text)
+	if len(runes) <= limit {
+		return text
+	}
+	return string(runes[:limit]) + "..."
+}

internal/qualitygate/publiccontent/types.go

@@ -0,0 +1,30 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package publiccontent
+
+import "github.com/larksuite/cli/internal/qualitygate/report"
+
+type Options struct {
+	Repo         string
+	ChangedFrom  string
+	MetadataPath string
+	BranchName   string
+}
+
+type Metadata struct {
+	Title  string `json:"title"`
+	Body   string `json:"body"`
+	Branch string `json:"branch"`
+}
+
+type Finding struct {
+	Rule       string
+	Action     report.Action
+	File       string
+	Line       int
+	Source     string
+	Excerpt    string
+	Message    string
+	Suggestion string
+}

internal/qualitygate/rules/dryrun.go

@@ -174,8 +174,9 @@ type materializedExample struct {
 }
 
 type placeholderContext struct {
-	FlagName  string
-	FlagUsage string
+	FlagName    string
+	FlagUsage   string
+	FlagDefault string
 }
 
 func materializePlaceholderExample(raw string, cmd manifest.Command) (materializedExample, bool) {
@@ -247,6 +248,7 @@ func placeholderContextForFlag(name string, flag *manifest.Flag) placeholderCont
 	ctx := placeholderContext{FlagName: name}
 	if flag != nil {
 		ctx.FlagUsage = flag.Usage
+		ctx.FlagDefault = flag.DefValue
 	}
 	return ctx
 }
@@ -309,11 +311,17 @@ func fakeValueForPlaceholder(raw string, ctx placeholderContext) (string, bool)
 	if name == "" {
 		return "", false
 	}
+	if value, ok := fakeNumericValueForPlaceholder(name, ctx); ok {
+		return value, true
+	}
+	if value, ok := fakeContextualURLValueForPlaceholder(name, ctx); ok {
+		return value, true
+	}
 	if value, ok := fakeValueFromPlaceholderName(name); ok {
 		return value, true
 	}
 	if isGenericPlaceholderName(name) {
-		return fakeValueFromUsageHint(ctx.FlagUsage)
+		return fakeValueFromContextHint(ctx)
 	}
 	return "", false
 }
@@ -336,16 +344,26 @@ func fakeValueFromPlaceholderName(name string) (string, bool) {
 		return "file_test123", true
 	case hasPlaceholderToken(tokens, "file") && hasPlaceholderToken(tokens, "token"):
 		return "file_test123", true
+	case hasPlaceholderToken(tokens, "folder") && hasPlaceholderToken(tokens, "token"):
+		return "fld_test123", true
 	case hasPlaceholderToken(tokens, "image", "img"):
 		return "img_test123", true
 	case hasPlaceholderToken(tokens, "app"):
 		return "app_test123", true
+	case hasPlaceholderToken(tokens, "draft"):
+		return "draft_test123", true
+	case hasPlaceholderToken(tokens, "label"):
+		return "label_test123", true
+	case hasPlaceholderToken(tokens, "share"):
+		return "share_test123", true
 	case hasPlaceholderToken(tokens, "doc", "document"):
 		return "doc_test123", true
 	case hasPlaceholderToken(tokens, "sheet", "spreadsheet"):
 		return "shtcn_test123", true
 	case hasPlaceholderToken(tokens, "base"):
 		return "base_test123", true
+	case hasPlaceholderToken(tokens, "space"):
+		return "space_test123", true
 	case hasPlaceholderToken(tokens, "table"):
 		return "tbl_test123", true
 	case hasPlaceholderToken(tokens, "view"):
@@ -377,17 +395,98 @@ func fakeValueFromPlaceholderName(name string) (string, bool) {
 	}
 }
 
-func fakeValueFromUsageHint(usage string) (string, bool) {
-	match := placeholderValuePattern.FindStringSubmatch(strings.ToLower(usage))
+func fakeValueFromContextHint(ctx placeholderContext) (string, bool) {
+	if value, ok := fakeNumericValueForPlaceholder("", ctx); ok {
+		return value, true
+	}
+	if value, ok := fakeContextualURLValueForPlaceholder("", ctx); ok {
+		return value, true
+	}
+	match := placeholderValuePattern.FindStringSubmatch(strings.ToLower(ctx.FlagUsage))
 	if len(match) != 2 || !knownTokenPrefix(match[1]) {
 		return "", false
 	}
 	return match[1] + "_test123", true
 }
 
+func fakeContextualURLValueForPlaceholder(name string, ctx placeholderContext) (string, bool) {
+	nameTokens := placeholderTokenSet(name)
+	flagName := strings.ReplaceAll(strings.ToLower(ctx.FlagName), "-", "_")
+	flagTokens := placeholderTokenSet(flagName)
+	if !hasPlaceholderToken(nameTokens, "url", "link") && !hasPlaceholderToken(flagTokens, "url", "link") {
+		return "", false
+	}
+	usage := strings.ToLower(ctx.FlagUsage)
+	if strings.Contains(usage, "lark") || strings.Contains(usage, "feishu") || strings.Contains(usage, "document url") {
+		return "https://example.feishu.cn/docx/doc_test123", true
+	}
+	return "", false
+}
+
+func fakeNumericValueForPlaceholder(name string, ctx placeholderContext) (string, bool) {
+	nameTokens := placeholderTokenSet(name)
+	flagName := strings.ReplaceAll(strings.ToLower(ctx.FlagName), "-", "_")
+	flagTokens := placeholderTokenSet(flagName)
+	usage := strings.ToLower(ctx.FlagUsage)
+
+	switch {
+	case placeholderTokenPair(nameTokens, "meeting", "id") || placeholderTokenPair(flagTokens, "meeting", "id"):
+		return "400000000001", true
+	case placeholderTokenPair(nameTokens, "meeting", "ids") || placeholderTokenPair(flagTokens, "meeting", "ids"):
+		return "400000000001", true
+	case placeholderTokenPair(nameTokens, "meeting", "no") || placeholderTokenPair(flagTokens, "meeting", "no"):
+		return "123456789", true
+	case placeholderTokenPair(nameTokens, "meeting", "number") || placeholderTokenPair(flagTokens, "meeting", "number"):
+		return "123456789", true
+	case hasPlaceholderToken(nameTokens, "timestamp") || hasPlaceholderToken(flagTokens, "timestamp") || strings.Contains(usage, "unix timestamp"):
+		return defaultPositiveInteger(ctx.FlagDefault, "1893456000"), true
+	case placeholderTokenPair(nameTokens, "page", "size") || placeholderTokenPair(flagTokens, "page", "size"):
+		return defaultPositiveInteger(ctx.FlagDefault, "20"), true
+	case placeholderTokenPair(nameTokens, "page", "limit") || placeholderTokenPair(flagTokens, "page", "limit"):
+		return defaultPositiveInteger(ctx.FlagDefault, "10"), true
+	case numericPlaceholderName(nameTokens) || numericPlaceholderName(flagTokens) || numericUsageHint(usage):
+		return defaultPositiveInteger(ctx.FlagDefault, "20"), true
+	default:
+		return "", false
+	}
+}
+
+func numericPlaceholderName(tokens map[string]bool) bool {
+	if len(tokens) == 0 || hasPlaceholderToken(tokens, "token", "format", "type", "status", "mode") {
+		return false
+	}
+	return hasPlaceholderToken(tokens,
+		"amount", "count", "depth", "height", "index", "length", "limit", "max",
+		"number", "revision", "size", "width",
+	)
+}
+
+func numericUsageHint(usage string) bool {
+	if usage == "" {
+		return false
+	}
+	return strings.Contains(usage, "positive integer") ||
+		strings.Contains(usage, "decimal integer") ||
+		strings.Contains(usage, "number of ") ||
+		strings.Contains(usage, "(number)")
+}
+
+func defaultPositiveInteger(raw, fallback string) string {
+	raw = strings.TrimSpace(raw)
+	if raw == "" || strings.HasPrefix(raw, "-") || raw == "0" {
+		return fallback
+	}
+	for _, r := range raw {
+		if r < '0' || r > '9' {
+			return fallback
+		}
+	}
+	return raw
+}
+
 func knownTokenPrefix(prefix string) bool {
 	switch prefix {
-	case "app", "base", "doc", "file", "fld", "img", "item", "meeting", "obcn", "oc", "od", "om", "ou", "page", "rec", "shtcn", "task", "tbl", "token", "viw", "wiki":
+	case "app", "base", "doc", "draft", "file", "fld", "img", "item", "label", "meeting", "obcn", "oc", "od", "om", "ou", "page", "rec", "share", "shtcn", "space", "task", "tbl", "token", "viw", "wiki":
 		return true
 	default:
 		return false
@@ -431,6 +530,10 @@ func hasPlaceholderToken(tokens map[string]bool, wants ...string) bool {
 	return false
 }
 
+func placeholderTokenPair(tokens map[string]bool, first, second string) bool {
+	return tokens[first] && tokens[second]
+}
+
 func hasUnresolvedDryRunPlaceholder(value string) bool {
 	if skillscan.HasPlaceholder(value) {
 		return true
@@ -623,6 +726,7 @@ func appendDryRunArg(raw string) ([]string, error) {
 		return nil, fmt.Errorf("not a lark-cli command")
 	}
 	argv = truncateShellTail(argv)
+	argv = forceDryRunJSONFormat(argv)
 	hasDryRunArg := false
 	dryRunEnabled := false
 	for _, arg := range argv[1:] {
@@ -642,6 +746,23 @@ func appendDryRunArg(raw string) ([]string, error) {
 	return append(argv[1:], "--dry-run"), nil
 }
 
+func forceDryRunJSONFormat(argv []string) []string {
+	for i := 1; i < len(argv); i++ {
+		arg := argv[i]
+		if arg == "--format" {
+			if i+1 < len(argv) && argv[i+1] == "pretty" {
+				argv[i+1] = "json"
+			}
+			return argv
+		}
+		if arg == "--format=pretty" {
+			argv[i] = "--format=json"
+			return argv
+		}
+	}
+	return argv
+}
+
 func truncateShellTail(argv []string) []string {
 	for i, arg := range argv {
 		if i == 0 {

internal/qualitygate/rules/dryrun_test.go

@@ -305,6 +305,161 @@ func TestRunDryRunsMaterializesInlinePlaceholderFlagValues(t *testing.T) {
 	}
 }
 
+func TestRunDryRunsMaterializesNumericPlaceholderFlagValues(t *testing.T) {
+	cliBin, argsPath := fakeDryRunCLI(t, `{"api":[{"method":"GET","url":"/open-apis/vc/v1/bots/events","params":{"meeting_id":"400000000001","page_size":50}}]}`)
+	m := manifest.Manifest{Commands: []manifest.Command{{
+		Path:     "vc +meeting-events",
+		Runnable: true,
+		Flags: []manifest.Flag{
+			{Name: "meeting-id", TakesValue: true, Usage: "meeting ID to query; must be a long positive integer, not a 9-digit meeting number"},
+			{Name: "page-size", TakesValue: true, Usage: "page size, 20-100 (default 50)", DefValue: "50"},
+			{Name: "dry-run"},
+		},
+	}}}
+	ex := skillscan.Example{
+		Raw:            "lark-cli vc +meeting-events --meeting-id <meeting_id> --page-size <page_size>",
+		SourceFile:     "skills/lark-vc-agent/SKILL.md",
+		Line:           120,
+		HasPlaceholder: true,
+	}
+
+	diags, facts := RunDryRuns(context.Background(), cliBin, m, []skillscan.Example{ex})
+	if len(diags) != 0 {
+		t.Fatalf("RunDryRuns() diagnostics = %#v", diags)
+	}
+	if len(facts) != 1 || !facts[0].Executable || facts[0].SkipReason != "" {
+		t.Fatalf("numeric placeholder example should be executable after materialization: %#v", facts)
+	}
+	wantArgs := []string{"vc", "+meeting-events", "--meeting-id", "400000000001", "--page-size", "50", "--dry-run"}
+	if gotArgs := readArgs(t, argsPath); !reflect.DeepEqual(gotArgs, wantArgs) {
+		t.Fatalf("fake CLI args = %#v, want %#v", gotArgs, wantArgs)
+	}
+}
+
+func TestRunDryRunsMaterializesNumericPlaceholdersInsideJSONFlags(t *testing.T) {
+	cliBin, argsPath := fakeDryRunCLI(t, `{"api":[{"method":"GET","url":"/open-apis/test","params":{"timestamp":"1893456000","count":"20"}}]}`)
+	m := manifest.Manifest{Commands: []manifest.Command{{
+		Path:     "api GET",
+		Runnable: true,
+		Flags: []manifest.Flag{
+			{Name: "params", TakesValue: true},
+			{Name: "dry-run"},
+		},
+	}}}
+	ex := skillscan.Example{
+		Raw:            `lark-cli api GET /open-apis/test --params '{"timestamp":"<timestamp>","count":"<count>"}'`,
+		SourceFile:     "skills/lark-demo/SKILL.md",
+		Line:           20,
+		HasPlaceholder: true,
+	}
+
+	diags, facts := RunDryRuns(context.Background(), cliBin, m, []skillscan.Example{ex})
+	if len(diags) != 0 {
+		t.Fatalf("RunDryRuns() diagnostics = %#v", diags)
+	}
+	if len(facts) != 1 || !facts[0].Executable || facts[0].SkipReason != "" {
+		t.Fatalf("JSON numeric placeholder example should be executable after materialization: %#v", facts)
+	}
+	wantArgs := []string{"api", "GET", "/open-apis/test", "--params", `{"timestamp":"1893456000","count":"20"}`, "--dry-run"}
+	if gotArgs := readArgs(t, argsPath); !reflect.DeepEqual(gotArgs, wantArgs) {
+		t.Fatalf("fake CLI args = %#v, want %#v", gotArgs, wantArgs)
+	}
+}
+
+func TestRunDryRunsMaterializesLarkDocumentURLPlaceholders(t *testing.T) {
+	cliBin, argsPath := fakeDryRunCLI(t, `{"api":[{"method":"GET","url":"/open-apis/drive/v1/metas/batch_query"}]}`)
+	m := manifest.Manifest{Commands: []manifest.Command{{
+		Path:     "drive +inspect",
+		Runnable: true,
+		Flags: []manifest.Flag{
+			{Name: "url", TakesValue: true, Usage: "Lark/Feishu document URL (docx, doc, sheet, bitable, wiki, file, folder, mindnote, slides)"},
+			{Name: "format", TakesValue: true},
+			{Name: "dry-run"},
+		},
+	}}}
+	ex := skillscan.Example{
+		Raw:            "lark-cli drive +inspect --url '<url>' --format json",
+		SourceFile:     "skills/lark-drive/references/lark-drive-workflow-permission-governance-commands.md",
+		Line:           15,
+		HasPlaceholder: true,
+	}
+
+	diags, facts := RunDryRuns(context.Background(), cliBin, m, []skillscan.Example{ex})
+	if len(diags) != 0 {
+		t.Fatalf("RunDryRuns() diagnostics = %#v", diags)
+	}
+	if len(facts) != 1 || !facts[0].Executable || facts[0].SkipReason != "" {
+		t.Fatalf("Lark URL placeholder example should be executable after materialization: %#v", facts)
+	}
+	wantArgs := []string{"drive", "+inspect", "--url", "https://example.feishu.cn/docx/doc_test123", "--format", "json", "--dry-run"}
+	if gotArgs := readArgs(t, argsPath); !reflect.DeepEqual(gotArgs, wantArgs) {
+		t.Fatalf("fake CLI args = %#v, want %#v", gotArgs, wantArgs)
+	}
+}
+
+func TestRunDryRunsMaterializesResourceIDPlaceholderFlagValues(t *testing.T) {
+	cliBin, argsPath := fakeDryRunCLI(t, `{"api":[{"method":"GET","url":"/open-apis/wiki/v2/spaces/space_test123/nodes"}]}`)
+	m := manifest.Manifest{Commands: []manifest.Command{{
+		Path:     "wiki +node-list",
+		Runnable: true,
+		Flags: []manifest.Flag{
+			{Name: "space-id", TakesValue: true, Usage: "wiki space ID"},
+			{Name: "page-token", TakesValue: true, Usage: "page token"},
+			{Name: "format", TakesValue: true},
+			{Name: "dry-run"},
+		},
+	}}}
+	ex := skillscan.Example{
+		Raw:            "lark-cli wiki +node-list --space-id <space_id> --page-token <PAGE_TOKEN> --format json",
+		SourceFile:     "skills/lark-wiki/references/lark-wiki-node-list.md",
+		Line:           24,
+		HasPlaceholder: true,
+	}
+
+	diags, facts := RunDryRuns(context.Background(), cliBin, m, []skillscan.Example{ex})
+	if len(diags) != 0 {
+		t.Fatalf("RunDryRuns() diagnostics = %#v", diags)
+	}
+	if len(facts) != 1 || !facts[0].Executable || facts[0].SkipReason != "" {
+		t.Fatalf("resource ID placeholder example should be executable after materialization: %#v", facts)
+	}
+	wantArgs := []string{"wiki", "+node-list", "--space-id", "space_test123", "--page-token", "page_test123", "--format", "json", "--dry-run"}
+	if gotArgs := readArgs(t, argsPath); !reflect.DeepEqual(gotArgs, wantArgs) {
+		t.Fatalf("fake CLI args = %#v, want %#v", gotArgs, wantArgs)
+	}
+}
+
+func TestRunDryRunsMaterializesResourcePlaceholdersInsideJSONFlags(t *testing.T) {
+	cliBin, argsPath := fakeDryRunCLI(t, `{"api":[{"method":"POST","url":"/open-apis/mail/v1/user_mailboxes/me/drafts/draft_test123/send"}]}`)
+	m := manifest.Manifest{Commands: []manifest.Command{{
+		Path:     "mail user_mailbox.drafts send",
+		Runnable: true,
+		Flags: []manifest.Flag{
+			{Name: "params", TakesValue: true},
+			{Name: "data", TakesValue: true},
+			{Name: "dry-run"},
+		},
+	}}}
+	ex := skillscan.Example{
+		Raw:            `lark-cli mail user_mailbox.drafts send --params '{"user_mailbox_id":"me","draft_id":"<draft_id>"}' --data '{"send_time":"<unix_timestamp>"}'`,
+		SourceFile:     "skills/lark-mail/references/lark-mail-send.md",
+		Line:           172,
+		HasPlaceholder: true,
+	}
+
+	diags, facts := RunDryRuns(context.Background(), cliBin, m, []skillscan.Example{ex})
+	if len(diags) != 0 {
+		t.Fatalf("RunDryRuns() diagnostics = %#v", diags)
+	}
+	if len(facts) != 1 || !facts[0].Executable || facts[0].SkipReason != "" {
+		t.Fatalf("JSON resource placeholder example should be executable after materialization: %#v", facts)
+	}
+	wantArgs := []string{"mail", "user_mailbox.drafts", "send", "--params", `{"user_mailbox_id":"me","draft_id":"draft_test123"}`, "--data", `{"send_time":"1893456000"}`, "--dry-run"}
+	if gotArgs := readArgs(t, argsPath); !reflect.DeepEqual(gotArgs, wantArgs) {
+		t.Fatalf("fake CLI args = %#v, want %#v", gotArgs, wantArgs)
+	}
+}
+
 func TestRunDryRunsSkipsUnknownFlagsBeforeDryRun(t *testing.T) {
 	m := manifest.Manifest{Commands: []manifest.Command{{
 		Path:     "im +chat-messages-list",
@@ -600,6 +755,51 @@ func TestAppendDryRunArgDoesNotDuplicate(t *testing.T) {
 	}
 }
 
+func TestAppendDryRunArgForcesJSONFormat(t *testing.T) {
+	got, err := appendDryRunArg("lark-cli vc +meeting-events --meeting-id 400000000001 --format pretty")
+	if err != nil {
+		t.Fatalf("appendDryRunArg() error = %v", err)
+	}
+	want := []string{"vc", "+meeting-events", "--meeting-id", "400000000001", "--format", "json", "--dry-run"}
+	if !reflect.DeepEqual(got, want) {
+		t.Fatalf("appendDryRunArg() = %#v, want %#v", got, want)
+	}
+}
+
+func TestAppendDryRunArgForcesInlineJSONFormat(t *testing.T) {
+	got, err := appendDryRunArg("lark-cli vc +meeting-events --meeting-id 400000000001 --format=pretty --dry-run")
+	if err != nil {
+		t.Fatalf("appendDryRunArg() error = %v", err)
+	}
+	want := []string{"vc", "+meeting-events", "--meeting-id", "400000000001", "--format=json", "--dry-run"}
+	if !reflect.DeepEqual(got, want) {
+		t.Fatalf("appendDryRunArg() = %#v, want %#v", got, want)
+	}
+}
+
+func TestAppendDryRunArgPreservesNonPrettyFormat(t *testing.T) {
+	for _, raw := range []string{
+		"lark-cli mail +watch --format data --dry-run",
+		"lark-cli export +events --format=ndjson --dry-run",
+		"lark-cli docs +fetch --format table",
+	} {
+		got, err := appendDryRunArg(raw)
+		if err != nil {
+			t.Fatalf("appendDryRunArg(%q) error = %v", raw, err)
+		}
+		for _, arg := range got {
+			if arg == "--format=json" {
+				t.Fatalf("appendDryRunArg(%q) unexpectedly rewrote inline format: %#v", raw, got)
+			}
+		}
+		for i, arg := range got {
+			if arg == "--format" && i+1 < len(got) && got[i+1] == "json" {
+				t.Fatalf("appendDryRunArg(%q) unexpectedly rewrote split format: %#v", raw, got)
+			}
+		}
+	}
+}
+
 func TestAppendDryRunArgForcesDryRunWhenExplicitlyDisabled(t *testing.T) {
 	got, err := appendDryRunArg("lark-cli docs +fetch --dry-run=false --doc abc")
 	if err != nil {

internal/qualitygate/rules/run.go

@@ -15,18 +15,20 @@ import (
 	manifestexamples "github.com/larksuite/cli/internal/qualitygate/examples"
 	"github.com/larksuite/cli/internal/qualitygate/facts"
 	"github.com/larksuite/cli/internal/qualitygate/manifest"
+	"github.com/larksuite/cli/internal/qualitygate/publiccontent"
 	"github.com/larksuite/cli/internal/qualitygate/report"
 	"github.com/larksuite/cli/internal/qualitygate/skillscan"
 	"github.com/larksuite/cli/internal/vfs"
 )
 
 type Options struct {
-	Repo             string
-	CLIBin           string
-	ChangedFrom      string
-	FactsOut         string
-	ManifestPath     string
-	CommandIndexPath string
+	Repo                      string
+	CLIBin                    string
+	ChangedFrom               string
+	FactsOut                  string
+	ManifestPath              string
+	CommandIndexPath          string
+	PublicContentMetadataPath string
 }
 
 func Run(ctx context.Context, opts Options) ([]report.Diagnostic, facts.Facts, error) {
@@ -98,9 +100,60 @@ func Run(ctx context.Context, opts Options) ([]report.Diagnostic, facts.Facts, e
 	if opts.ChangedFrom != "" {
 		diags = append(diags, errorDiags...)
 	}
+	publicContent, err := publiccontent.Collect(ctx, publiccontent.Options{
+		Repo:         opts.Repo,
+		ChangedFrom:  opts.ChangedFrom,
+		MetadataPath: opts.PublicContentMetadataPath,
+	})
+	if err != nil {
+		return nil, facts.Facts{}, err
+	}
+	diags = append(diags, publicContentDiagnostics(publicContent)...)
 	diags = filterPRDiagnostics(opts.Repo, opts.ChangedFrom, scope, m, diags)
 
-	return diags, facts.BuildWithCommandLookup(m, commandIndex, skillFacts, skillQualityFacts, errorFacts, exampleFacts, outputFacts, diags, scope.Files), nil
+	builtFacts := facts.BuildWithCommandLookup(m, commandIndex, skillFacts, skillQualityFacts, errorFacts, exampleFacts, outputFacts, diags, scope.Files)
+	return diags, facts.WithPublicContent(builtFacts, publicContentFacts(publicContent)), nil
+}
+
+func publicContentDiagnostics(items []publiccontent.Finding) []report.Diagnostic {
+	if len(items) == 0 {
+		return nil
+	}
+	out := make([]report.Diagnostic, 0, len(items))
+	for _, item := range items {
+		if item.Rule == "public_content_semantic_candidate" {
+			continue
+		}
+		out = append(out, report.Diagnostic{
+			Rule:       item.Rule,
+			Action:     item.Action,
+			File:       item.File,
+			Line:       item.Line,
+			Message:    item.Message,
+			Suggestion: item.Suggestion,
+		})
+	}
+	return out
+}
+
+func publicContentFacts(items []publiccontent.Finding) []facts.PublicContentFact {
+	if len(items) == 0 {
+		return nil
+	}
+	out := make([]facts.PublicContentFact, 0, len(items))
+	for _, item := range items {
+		out = append(out, facts.PublicContentFact{
+			Rule:       item.Rule,
+			Action:     item.Action,
+			File:       item.File,
+			Line:       item.Line,
+			Source:     item.Source,
+			Excerpt:    item.Excerpt,
+			Message:    item.Message,
+			Suggestion: item.Suggestion,
+		})
+	}
+	return out
 }
 
 func readManifestInput(path, kind, flag string) (manifest.Manifest, error) {
@@ -167,6 +220,9 @@ func filterPRDiagnostics(repo, changedFrom string, scope qdiff.Scope, m manifest
 }
 
 func prDiagnosticRelevant(repo string, changedFiles map[string]bool, commandScope diagnosticCommandScope, m manifest.Manifest, diag report.Diagnostic) bool {
+	if strings.HasPrefix(diag.Rule, "public_content_") {
+		return true
+	}
 	file := normalizeDiagnosticFile(repo, diag.File)
 	if file != "" && changedFiles[file] {
 		return true

internal/qualitygate/rules/run_test.go

@@ -189,6 +189,99 @@ description: Manage Drive comments with service command references.
 	}
 }
 
+func TestRunCollectsPublicContentFindingsIntoDiagnosticsAndFacts(t *testing.T) {
+	repo := t.TempDir()
+	runGit(t, repo, "init")
+	runGit(t, repo, "config", "user.email", "test@example.com")
+	runGit(t, repo, "config", "user.name", "Test User")
+	if err := vfs.WriteFile(filepath.Join(repo, "README.md"), []byte("# test\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	runGit(t, repo, "add", "README.md")
+	runGit(t, repo, "commit", "-m", "base")
+
+	if err := vfs.MkdirAll(filepath.Join(repo, "docs"), 0o755); err != nil {
+		t.Fatal(err)
+	}
+	publicDoc := "api_" + "key = \"example-public-key\"\n" +
+		"Public docs describe a pri" + "vate request header and trust classification detail.\n"
+	if err := vfs.WriteFile(filepath.Join(repo, "docs", "public.md"), []byte(publicDoc), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	runGit(t, repo, "add", "docs/public.md")
+	runGit(t, repo, "commit", "-m", "add public doc")
+
+	metadataPath := filepath.Join(repo, "pr-metadata.json")
+	if err := vfs.WriteFile(metadataPath, []byte(`{"title":"public docs","body":"Change`+`-Id: I0123456789abcdef0123456789abcdef01234567"}`), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	manifestPath := filepath.Join(repo, "command-manifest.json")
+	indexPath := filepath.Join(repo, "command-index.json")
+	m := manifest.Manifest{SchemaVersion: 1, Commands: []manifest.Command{{
+		Path:          "docs +fetch",
+		CanonicalPath: "docs +fetch",
+		Domain:        "docs",
+		Source:        manifest.SourceShortcut,
+	}}}
+	if err := manifest.WriteFile(manifestPath, manifest.KindCommandManifest, m); err != nil {
+		t.Fatal(err)
+	}
+	idx := manifest.Manifest{SchemaVersion: 1, Commands: append([]manifest.Command{}, m.Commands...)}
+	idx.Commands = append(idx.Commands, manifest.Command{
+		Path:          "drive files get",
+		CanonicalPath: "drive files get",
+		Domain:        "drive",
+		Source:        manifest.SourceService,
+		Generated:     true,
+		Runnable:      true,
+	})
+	if err := manifest.WriteFile(indexPath, manifest.KindCommandIndex, idx); err != nil {
+		t.Fatal(err)
+	}
+
+	diags, gotFacts, err := Run(context.Background(), Options{
+		Repo:                      repo,
+		CLIBin:                    "./lark-cli",
+		ChangedFrom:               "HEAD~1",
+		ManifestPath:              manifestPath,
+		CommandIndexPath:          indexPath,
+		PublicContentMetadataPath: metadataPath,
+	})
+	if err != nil {
+		t.Fatalf("Run() error = %v", err)
+	}
+	actions := map[string]report.Action{}
+	for _, diag := range diags {
+		actions[diag.Rule] = diag.Action
+	}
+	if actions["public_content_generic_credential"] != report.ActionReject {
+		t.Fatalf("generic credential diagnostic action = %q, diagnostics=%#v", actions["public_content_generic_credential"], diags)
+	}
+	if actions["public_content_change_id_trailer"] != report.ActionReject {
+		t.Fatalf("change-id diagnostic action = %q, diagnostics=%#v", actions["public_content_change_id_trailer"], diags)
+	}
+	if actions["public_content_semantic_candidate"] != "" {
+		t.Fatalf("semantic candidates should not become deterministic diagnostics: %#v", diags)
+	}
+	factRules := map[string]bool{}
+	for _, item := range gotFacts.PublicContent {
+		factRules[item.Rule] = true
+	}
+	for _, want := range []string{
+		"public_content_generic_credential",
+		"public_content_change_id_trailer",
+		"public_content_semantic_candidate",
+	} {
+		if !factRules[want] {
+			t.Fatalf("missing public content fact %s: %#v", want, gotFacts.PublicContent)
+		}
+	}
+	if len(gotFacts.PublicContent) < 3 {
+		t.Fatalf("public content facts = %#v", gotFacts.PublicContent)
+	}
+}
+
 func TestLoadBaseReferenceManifestReadsCommandGolden(t *testing.T) {
 	repo := t.TempDir()
 	runGit(t, repo, "init")
@@ -506,7 +599,7 @@ func TestNormalizeDiagnosticFileHandlesAbsoluteRepo(t *testing.T) {
 
 func runGit(t *testing.T, repo string, args ...string) {
 	t.Helper()
-	cmd := exec.Command("git", append([]string{"-C", repo}, args...)...)
+	cmd := exec.Command("git", append([]string{"-c", "core.hooksPath=/dev/null", "-C", repo}, args...)...)
 	cmd.Env = append(os.Environ(), "GIT_AUTHOR_DATE=2026-06-17T00:00:00Z", "GIT_COMMITTER_DATE=2026-06-17T00:00:00Z")
 	out, err := cmd.CombinedOutput()
 	if err != nil {