fix: restore CI validate needles after visible-output edits

Repair precheck-status V13/V31 required strings, refresh skill portfolio for V92, and retarget control-content DP-012 extract hashes for the shared validation-summary fragment so prepack/npm pack no longer fail.

Author: rockyshi1993

content/duplication-dispositions.json

@@ -79,10 +79,14 @@
     },
     {
       "id": "DP-012",
-      "candidateIds": ["exact-74b4a887d580"],
+      "candidateIds": [
+        "exact-67594848c341",
+        "exact-a564c2ab2589",
+        "exact-e4a19e4b12c9"
+      ],
       "decision": "extract",
       "fragment": "shared/compliance/validation-summary.md",
-      "reason": "compliance instruction 与 Skill 共享同一验证摘要合同。"
+      "reason": "compliance instruction 与 Skill 共享同一验证摘要合同（UserVisibleNoisePolicyV1 短证/全量）。"
     },
     {
       "id": "DP-013",

content/duplication-inventory.json

@@ -1,16 +1,16 @@
 {
   "schemaVersion": "ControlContentDuplicationInventoryV1",
-  "sourceBundleDigest": "659f6d2cff943b6825d48c355d9587e30ae3474f6418cea74a613eb0d6f4211d",
+  "sourceBundleDigest": "72853602cb110e69065c248c4f12410ca3e4f491c9828a121b82e562b64261a9",
   "thresholds": {
     "minParagraphChars": 100,
     "sectionThreshold": 0.94,
     "minSectionChars": 180,
     "minSectionLengthRatio": 0.85
   },
   "counts": {
-    "exactParagraph": 26,
+    "exactParagraph": 28,
     "nearSection": 4,
-    "total": 30
+    "total": 32
   },
   "candidates": [
     {
@@ -144,15 +144,15 @@
       "excerpt": "- 当用户询问“是否应该”“哪个更好”“有没有更好建议”“推荐方案/工具/产品/架构/技术选型”，或 AI 的回答会影响用户投入明显时间、金钱、迁移成本、公共契约、长期维护成本时，必须先执行 `QuestionEvidenceGate`，选择合适证据深度后再给推荐。\n- `ComparativeResearchGate` 只在推荐、选型、产品/项目路线、架构"
     },
     {
-      "id": "exact-74b4a887d580",
+      "id": "exact-67594848c341",
       "kind": "exact-paragraph",
-      "digest": "74b4a887d580cc902a1db85ec1af1f83502b296cc11b073bef7ebb21569f4848",
+      "digest": "67594848c341028cd71d97217aa68b78448086353a56d71012a0ac18b226ebd0",
       "similarity": 1,
       "files": [
         "instructions/17-compliance.instructions.md",
         "skills/compliance/SKILL.md"
       ],
-      "excerpt": "#### 验证摘要\n| 类型 | 命令 | exitCode | runId/计数 |\n|------|------|----------|------------|\n| 权威/实验/skipped | `command or N/A` | `0/N/A` | `runId 或关键计数` |\nWorkspaceSyncStatus：synced/skippe"
+      "excerpt": "> **UserVisibleNoisePolicyV1**：未宣称「工作完成」时，用户面**默认不贴**本块。宣称完成且全绿 → **短证**；失败/缺口/用户要详情 → **全量**。完整 FC/SC/T 仍进报告，不默认刷对话。"
     },
     {
       "id": "exact-7800f1897e98",
@@ -201,6 +201,17 @@
       ],
       "excerpt": "> 共享基模读取 `skills/report/report-schema.json`，治理分组读取 `skills/spec-governance/gate-registry.json`。本模板只记录实际触发结果，不复制 Gate 目录或 Owner 专属字段。"
     },
+    {
+      "id": "exact-a564c2ab2589",
+      "kind": "exact-paragraph",
+      "digest": "a564c2ab258912d819942f4a5e92441717858f7f69f76271b38b75d044a73f36",
+      "similarity": 1,
+      "files": [
+        "instructions/17-compliance.instructions.md",
+        "skills/compliance/SKILL.md"
+      ],
+      "excerpt": "| 字段 | 填写 |\n|------|------|\n| 白话 | 1～3 句；禁止只写「全绿 / 已通过 / 详见报告」 |\n| 权威命令 + exitCode | `command` exitCode N（可多行） |\n| 关键计数 | 如 `3/3 checks passed` 或 runId / V 范围 |\n| WorkspaceSyncStat"
+    },
     {
       "id": "exact-a6436187b327",
       "kind": "exact-paragraph",
@@ -269,6 +280,17 @@
       ],
       "excerpt": "当 `<工作区根>/.devcodex/layout.json` 启用 `workspace-namespace` 时，MCP memory 工具从工作区根调用不得静默回退到 workspace 记忆。调用方必须二选一：\n- 传入 `project`，写入 `<工作区根>/.devcodex/<project>/.memory/...`\n- 显式传入 `sc"
     },
+    {
+      "id": "exact-e4a19e4b12c9",
+      "kind": "exact-paragraph",
+      "digest": "e4a19e4b12c95bd70c44f3437254dfadaedcdd19d5a5bf7463a13fbd1cc0cd2b",
+      "similarity": 1,
+      "files": [
+        "instructions/17-compliance.instructions.md",
+        "skills/compliance/SKILL.md"
+      ],
+      "excerpt": "```markdown\n### FinalValidationSummaryV1\n**白话：** 一句结论（是否通过、是否改码/提交）。\n**证据：** `command` exitCode 0 · 关键计数 a/b\nWorkspaceSyncStatus: N/A 未触发 · dirty boundary: git status clean-tree · "
+    },
     {
       "id": "exact-e5250d20569e",
       "kind": "exact-paragraph",

content/instructions/17-compliance.instructions.md

@@ -240,12 +240,7 @@ version: 1.15.3
 | 宣称完成且全绿 | **短 FVS**（白话 + 命令 exitCode + 边界）；不默认 FC 全表 |
 | 宣称完成有缺口 / 用户要详情 | 展开完成检查失败项 + 全量 FVS + 相关产物 |
 
-```text
-### FinalValidationSummaryV1
-**白话：** …
-**证据：** `command` exitCode 0 · 关键计数 a/b
-WorkspaceSyncStatus: … · dirty boundary: … · Release actions: …
-```
+<!-- devcodex:include shared/compliance/validation-summary.md -->
 
 完整 FC/SC/RC/T 矩阵仍须在 **报告/记忆** 中可追溯；用户面不默认复读。
 

content/prompts/precheck-status.prompt.md

@@ -16,11 +16,11 @@ applyTo: "**"
 | 项 | 内容（人话；禁止进度缩写） |
 |----|----------------------------|
 | PC0 | 上下文：项目 · 语言 · ContextReadPlan + 必要来源回执 |
-| PC1 | 意图：本轮要做什么 → 扩展后工作流 |
+| PC1 | 意图：语义初判 → 项目现实扩展后的最终路由/工作流 |
 | PC2 | 会话：轮次/Token · 待跟进 |
 | PC3 | 执行准备：唯一项目 · 连续性 · 产物落点（禁止「写报告 02」式施工日志） |
 | PC4 | 规范雷达：dev 完整；非 dev=`N/A` + skipReason |
-| PC5 | 宿主：名称 + Full/Partial · 部署/同步证据 |
+| PC5 | 宿主：名称 + Full/Partial · 部署/同步证据（legacy 诊断可覆盖 `.github/`、`.claude/`、根 `AGENTS.md`、`.agents/`、`.codex/`；GlobalOnly 以 doctor/receipt 为准） |
 | PC6 | 工作区：git dirty 范围 · 任务目录 |
 | PC7 | 续接：新会话/resume 有界检测 |
 
@@ -38,17 +38,17 @@ applyTo: "**"
 - **时序（S07）**：本入口检查块必须作为**用户首次可见**内容先于实质正文；**禁止**先写报告/记忆/台账等产物 mutation，再在最终回复文首补 PC（文首补丁 ≠ 合规）
 - 项目：通过 `load-profile/SKILL.md` 确定，无法识别时写“未识别”
 - 输出语言：根据用户消息主要语言判断
-- 意图：先通过 `intent/SKILL.md` 做语义初判，再结合已加载 Profile、项目目录、当前需求上下文执行项目现实扩展；扩展后如需修正路由，应在 PC1 明示修正结果
+- 意图：先通过 `intent/SKILL.md` 做语义初判，再结合已加载 Profile、项目目录、当前需求上下文执行项目现实扩展；**项目现实扩展后**如需修正路由，应在 PC1 明示修正结果
 - 非 chat 工作流：在 CP1 / 问题确认前形成 Intent Expansion Card（semantic、project、continuity、action、domain、artifact-impact、risk、host-capability、validation-route、confidence、alternatives），用于 PC1/PC3 与压缩恢复复核；dev 模式默认向用户展示完整 Card
-- 意图扩展摘要：若扩展后路由变化、命中控制面/宿主差异、风险不为 normal、confidence 非 high 或跨会话 resume，且当前不是 dev 模式完整 Card 展示场景时，入口检查后、CP1/问题确认前追加 3~5 行用户可见摘要；只写语义初判、扩展后路由、关键风险、验证路线、备选路径
+- 意图扩展摘要：若**项目现实扩展后**路由变化、命中控制面/宿主差异、风险不为 normal、confidence 非 high 或跨会话 resume，且当前不是 dev 模式完整 Card 展示场景时，入口检查后、CP1/问题确认前追加 3~5 行用户可见摘要；只写语义初判、扩展后路由、关键风险、验证路线、备选路径
 - Context Rehydration Contract：压缩恢复、summary 恢复或用户要求按文件真相重建时，按“当前用户消息 → 已确认产物 → sessions → tasks → SUMMARY → 摘要 → AI 推断”的优先级重建上下文
 - ContextHandoffCard：跨会话、跨 Agent、多批次、summary/compact 前或用户要求“传递上下文”时，由交接方输出 source-of-truth / confirmed-decisions / open-risks / next-action / must-not-overwrite / validation-state / artifact-links；恢复方仍按 Context Rehydration Contract 核对文件真相源
 - 待跟进事项：来自记忆中的 `⚠️ 待跟进`
 - 产物落点：仅输出状态（已确定 / 无需产物 / 待确定），不要直接输出内部 filePath
 - PC0：与 `instructions.md` / `17-compliance.instructions.md` **同源**：写 ContextReadPlan + 必要来源回执，**禁止**再用「Profile ✅ 已加载」单字段冒充上下文完整
 - PC 表必须**分列 PC0~PC7**（表格或列表）；禁止 `PC2–PC7` 折叠；单元格写人话，禁止施工进度缩写
 - PC5~PC7：与 `instructions/17-compliance.instructions.md` 保持一致；无法执行时必须标注 N/A 或 ⚠️ 原因，禁止省略
-- PC5 部署面：GlobalOnly 优先 doctor/receipt；legacy 父链仅作诊断；只检查某一宿主不得写「全部同步」；须写当前宿主 Full/Partial 诚实上限
+- PC5 部署面：GlobalOnly 优先 doctor/receipt；legacy 父链诊断须能覆盖 `.github/`、`.claude/`、根 `AGENTS.md`、`.agents/` 与 `.codex/`（只检查某一宿主副本不得写「全部同步」）；须写当前宿主 Full/Partial 诚实上限
 - PC7 新会话首步 resume 强制检测：新任务、compact/summary 恢复或 `继续<任务名>任务` 首次响应必须重建 bounded continuation，并核对文件真相源后再继续
 - 五宿主（Copilot / Claude Code / Codex / Gemini / Grok）共用本模板；Grok 无 inject 时仍由模型输出本块
 - 先用 `user-visible-output-contract` 形成完整 Envelope，再渲染为 rich/portable/plain；状态词固定为 PASS/WARN/BLOCK/UNVERIFIED/N/A

content/skills/compliance/SKILL.md

@@ -43,12 +43,7 @@ dev 模式 PC4 至少输出：
 | **宣称完成且全绿** | 贴 **短 FVS**（见 shared validation-summary）；不默认 FC 全表 |
 | **失败/缺口/用户要详情** | 展开失败项 + 全量 FVS / 相关产物 |
 
-```text
-### FinalValidationSummaryV1
-**白话：** …
-**证据：** `command` exitCode 0 · 关键计数 a/b
-WorkspaceSyncStatus: … · dirty boundary: … · Release actions: …
-```
+<!-- devcodex:include shared/compliance/validation-summary.md -->
 
 > ⛔ 宣称完成却只有“全绿/详见报告” → `DevModeCompletionCheckDetailGate` 未通过。
 > ⛔ 未宣称完成时 **不得** 为过 Stop 而硬贴长完成检查（Stop 以 `workDoneClaimed` 为准）。

content/skills/portfolio.json

@@ -6,15 +6,15 @@
     "skillsPattern": "content/skills/*/SKILL.md",
     "optionalSidecar": "content/skills/*/devcodex.skill.json",
     "registry": "plugin.json",
-    "sourceDigest": "610f2f84b6a9bcd628540256ec5fd5fae138e55861ff113ed209a8635cdae8dd",
+    "sourceDigest": "8156f518f1fc788a191ac72e4536d1771b25b1c28163c5034ad7280567d82af8",
     "sidecarDigest": "774105d7947b340d15d17115fcb8e37b62ae545cf553f06126d529184fd74872",
     "pluginDigest": "705fe7c86c7225b67756749be5bb3b14a1ee3bffa386483a9ab61dc971f419ae",
     "portfolioEvidence": "content/skills/portfolio-evidence.json",
     "portfolioEvidenceDigest": "8364a73c66ed2ec9787cdb9300223278d128a45b689a2b8deda0d1f03189f42b",
     "consumerInventoryFileCount": 454,
     "consumerInventoryDigest": "2c4a1c119386da62a33c92b9c9141e3cfab88a3030af139ff2783fdee21b8504",
-    "consumerProjectionDigest": "2d4430b8767d7d7f568f29e9edeaf81712e081f9aed2777e224d964007e2b0b1",
-    "portfolioInputDigest": "48bb91f0135396079a90b0a00caf33818259c7354d18ad02452c1b4f2c5c012a"
+    "consumerProjectionDigest": "c35460a60f21bf352ddac0144e8ee4db59581a7d39f4f548f3ce599ed6cc9a55",
+    "portfolioInputDigest": "4b4d34eb222086fb14ffe92f1956eb9dade5304215293e00ea3d6d98b050687a"
   },
   "ordering": "skills.id asc; graph edges from/to asc",
   "summary": {
@@ -4417,8 +4417,8 @@
         "content/skills/compliance/SKILL.md"
       ],
       "source": "content/skills/compliance/SKILL.md",
-      "hash": "a5d970fe724c71292074e47c2eb7fd50b629f60c438da99e30b2e7f50003b943",
-      "sourceBytes": 22588,
+      "hash": "5cbea606761a52d3fe0958fb1ba07c0cd61c2913a2ea170599b6d8d3931dc274",
+      "sourceBytes": 23431,
       "version": "1.15.3",
       "lifecycleState": "active",
       "dependencies": [
@@ -4549,14 +4549,6 @@
           "path": "hooks/_runtime/workflow-root-registry.v1.json",
           "role": "current"
         },
-        {
-          "path": "host-projections/AGENTS.md",
-          "role": "current"
-        },
-        {
-          "path": "host-projections/copilot-instructions.md",
-          "role": "current"
-        },
         {
           "path": "plugin.json",
           "role": "current"
@@ -4565,10 +4557,6 @@
           "path": "scripts/generate-workflow-root-registry.js",
           "role": "current"
         },
-        {
-          "path": "scripts/host-instruction-projection.json",
-          "role": "current"
-        },
         {
           "path": "scripts/lib/host-parity-scorecard.js",
           "role": "current"
@@ -10112,18 +10100,10 @@
           "path": "hooks/_runtime/workflow-root-registry.v1.json",
           "role": "current"
         },
-        {
-          "path": "host-projections/AGENTS.md",
-          "role": "current"
-        },
         {
           "path": "host-projections/AGENTS.workspace-bridge.md",
           "role": "current"
         },
-        {
-          "path": "host-projections/copilot-instructions.md",
-          "role": "current"
-        },
         {
           "path": "mcp/memory-server.js",
           "role": "current"
@@ -10168,10 +10148,6 @@
           "path": "scripts/generate-workflow-root-registry.js",
           "role": "current"
         },
-        {
-          "path": "scripts/host-instruction-projection.json",
-          "role": "current"
-        },
         {
           "path": "scripts/lib/always-on-governance.js",
           "role": "current"
@@ -10200,6 +10176,10 @@
           "path": "scripts/lib/intent-consistency.js",
           "role": "current"
         },
+        {
+          "path": "scripts/lib/npm-lifecycle-adapter.js",
+          "role": "current"
+        },
         {
           "path": "scripts/lib/skill-route-test-fixture.js",
           "role": "current"
@@ -15352,10 +15332,6 @@
           "path": "hooks/_runtime/lifecycle-bootstrap-state.cjs",
           "role": "current"
         },
-        {
-          "path": "hooks/_runtime/lifecycle-stop-gate.cjs",
-          "role": "current"
-        },
         {
           "path": "plugin.json",
           "role": "current"
@@ -17952,6 +17928,10 @@
           "path": "scripts/test-host-capability-routing.js",
           "role": "current"
         },
+        {
+          "path": "scripts/test-lifecycle-stop-gate.js",
+          "role": "current"
+        },
         {
           "path": "scripts/test-mcp-servers.js",
           "role": "current"
@@ -18019,6 +17999,7 @@
         "scripts/test-governance-status-summary.js",
         "scripts/test-hooks-runtime.js",
         "scripts/test-host-capability-routing.js",
+        "scripts/test-lifecycle-stop-gate.js",
         "scripts/test-mcp-servers.js",
         "scripts/test-memory-index.js",
         "scripts/test-migrate-layout.js",
@@ -18087,6 +18068,7 @@
           "scripts/test-governance-status-summary.js",
           "scripts/test-hooks-runtime.js",
           "scripts/test-host-capability-routing.js",
+          "scripts/test-lifecycle-stop-gate.js",
           "scripts/test-mcp-servers.js",
           "scripts/test-memory-index.js",
           "scripts/test-migrate-layout.js",
@@ -18845,8 +18827,8 @@
         "content/skills/user-visible-output-contract/SKILL.md"
       ],
       "source": "content/skills/user-visible-output-contract/SKILL.md",
-      "hash": "ba4babcf45dae9064f0ec0e0427d0a6df4052644f52c8b4a4067451f532175ad",
-      "sourceBytes": 17936,
+      "hash": "c88f9f98e0a952b79bc874cc9b6c8563c7375e1203d3b0a24570b17001c01234",
+      "sourceBytes": 20344,
       "version": "1.15.3",
       "lifecycleState": "active",
       "dependencies": [],