cc-devflow 4.5.14 → 4.5.16

package/docs/examples/local-handoff/ROADMAP.md

@@ -3,7 +3,7 @@
 ## Roadmap Meta
 
 - Roadmap version: `roadmap.v3`
-- Skill version: `5.3.0`
+- Skill version: `5.3.1`
 - Status: `active`
 - Last updated: `2026-04-16`
 - Owner / decider: `product-owner`
@@ -52,6 +52,7 @@
 
 ## Implementation Tracking
 - Roadmap state source: `roadmap.json`
+- Output language: en
 
 <!-- roadmap-tracking:start -->
 | RM-ID | Item | Stage | Priority | Primary Capability | Secondary Capabilities | Expected Spec Delta | Depends On | Status | REQ | Progress |
@@ -59,7 +60,6 @@
 | RM-020 | Add an audit-log export summary download | Stage 2 | P1 | cap-audit-log-export | - | codify local export summary truth | - | Local handoff | REQ-003 | 100% |
 <!-- roadmap-tracking:end -->
 
-
 ## Technical Architecture
 
 ```mermaid

package/docs/examples/local-handoff/changes/REQ-003-audit-log-export/task.md

@@ -4,8 +4,9 @@
 
 - Requirement version: `REQ-003.v1`
 - Design version: `design.v1`
-- CC-Plan skill version: `3.10.2`
+- CC-Plan skill version: `3.10.7`
 - Work branch: `REQ/003-audit-log-export`
+- Output language: en
 - Source roadmap item: `RM-020`
 - Source roadmap version: `roadmap.v3`
 
@@ -24,6 +25,39 @@
   - Testing decisions: test through admin panel action and visible row data
   - Out of scope: JSON export, scheduled reporting, shared reporting backend
 - AI Leverage Decision Lens: boil-lake; complete visible-row CSV export while keeping reporting-platform work outside the lake
+- Product / Creative Discovery:
+  - Worth doing: admins already copy audit summaries into weekly reports by hand.
+  - Desired product shape: download the same visible rows they are reviewing, without a reporting setup flow.
+  - Narrowest wedge: CSV export for currently visible summary rows.
+  - 10x / better version: scheduled reporting and shared report history, deferred outside this requirement.
+  - Do-nothing consequence: weekly reporting stays manual and easy to miscopy.
+- Second-Move Review:
+  - First good move: add a download action to the audit summary panel.
+  - Simpler move: export only visible rows in CSV and avoid backend report storage.
+  - Better architecture: later introduce a reporting backend for scheduled exports.
+  - Selected move: visible-row CSV download from the existing panel.
+  - Rejected tradeoff: no JSON export, scheduling, or shared reporting backend here.
+- ASCII Branch Chain Analysis:
+  Language rule: connector tokens stay ASCII; node text follows `Output language: en`.
+
+```text
+Requirement Impact Chain
+REQ: CSV export for visible audit summary rows
+|-- Upstream source: RM-020 + weekly admin reporting workflow
+|-- Current code path: src/admin/AuditSummaryPanel.tsx
+|   |-- caller: admin audit summary panel action
+|   |-- data or state: visible summary rows
+|   `-- deepest affected layer: browser download/blob boundary for CSV output
+|-- Required change: export visible rows without adding reporting storage
+`-- Verification seam: src/admin/AuditSummaryPanel.test.tsx
+
+Business Impact Chain
+OUTCOME: admins stop manually copying weekly audit summaries
+|-- Direct behavior impact: visible rows become a downloadable CSV
+|-- Downstream impact: local handoff documents proof for CSV-only export
+|-- Risk branch: exported rows diverge from visible panel state
+`-- Non-goal branch: JSON export, scheduled reporting, and backend history stay out
+```
 - Read first:
   - `design.md`
   - `src/admin/AuditSummaryPanel.tsx`
@@ -42,8 +76,8 @@
 ClaudeCode / Codex 执行本计划时，必须把本文件当成任务模板合同，而不是普通 TODO 列表。
 
 - Template source: `assets/TASKS_TEMPLATE.md`
-- CLI resolver: all workflow commands must run through `.claude/skills/cc-dev/scripts/resolve-cc-devflow.sh` or `.codex/skills/cc-dev/scripts/resolve-cc-devflow.sh`; if it cannot prove `query workflow-context` and `next-change-key`, stop blocked.
-- Context first: run resolved CLI `query workflow-context --change <changeId> --change-key <changeKey> --cwd <repo-root> --data-only --no-trace --compact` before opening deep sections.
+- CLI resolver: all workflow commands must run through `.claude/skills/cc-dev/scripts/resolve-cc-devflow.sh` or `.codex/skills/cc-dev/scripts/resolve-cc-devflow.sh`; if it cannot prove `next-change-key`, stop blocked.
+- Context first: read this `task.md`, current Git status/history, and PR or handoff truth when present before opening deep sections.
 - Task selection: use `scripts/select-ready-tasks.sh --tasks devflow/changes/<change-key>/task.md`.
 - Task block rule: read the full task block before coding; title-only execution is invalid.
 - Completion rule: after verification and review gates pass, run `scripts/mark-task-complete.sh --tasks devflow/changes/<change-key>/task.md --task <task-id>`.
@@ -54,8 +88,8 @@ DEVFLOW=".claude/skills/cc-dev/scripts/resolve-cc-devflow.sh"
 if [[ ! -f "$DEVFLOW" && -f ".codex/skills/cc-dev/scripts/resolve-cc-devflow.sh" ]]; then
   DEVFLOW=".codex/skills/cc-dev/scripts/resolve-cc-devflow.sh"
 fi
-bash "$DEVFLOW" require query workflow-context next-change-key
-bash "$DEVFLOW" query workflow-context --change <changeId> --change-key <changeKey> --cwd <repo-root> --data-only --no-trace --compactSCRIPT_ROOT=".claude/skills/cc-do/scripts"
+bash "$DEVFLOW" require next-change-key
+SCRIPT_ROOT=".claude/skills/cc-do/scripts"
 if [[ ! -d "$SCRIPT_ROOT" && -d ".codex/skills/cc-do/scripts" ]]; then
   SCRIPT_ROOT=".codex/skills/cc-do/scripts"
 fi

package/docs/examples/local-handoff/roadmap.json

@@ -5,12 +5,17 @@
   },
   "meta": {
     "roadmapVersion": "roadmap.v2",
-    "skillVersion": "5.3.0",
+    "skillVersion": "5.3.1",
     "status": "active",
     "lastUpdated": "2026-04-19",
     "currentFocusStage": "Stage 2"
   },
   "context": {
+    "projectDirectionMode": "",
+    "projectDirectionRationale": "",
+    "directionQuestionsSelected": [],
+    "directionQuestionsSkipped": [],
+    "directionGuardrailsApplied": [],
     "planningPosture": "",
     "evidenceMaturity": "",
     "canonicalTerms": [],
@@ -123,7 +128,7 @@
   },
   "backlogMeta": {
     "roadmapVersion": "roadmap.v2",
-    "skillVersion": "4.4.1",
+    "skillVersion": "5.3.1",
     "currentFocusStage": "Stage 2"
   },
   "dependencyHandoff": {

package/docs/examples/pdca-loop/BACKLOG.md

@@ -5,10 +5,11 @@
 ## Backlog Meta
 
 - Roadmap version: `roadmap.v1`
-- Skill version: `5.3.0`
+- Skill version: `5.3.1`
 - Last synced: `2026-04-19`
 - Current focus stage: `Stage 1`
 - Roadmap state source: `roadmap.json`
+- Output language: en
 
 ## Queue
 
@@ -23,6 +24,16 @@
 - Parallel-ready next wave: none until RM-001 proves the copy action shape
 - Notes on blockers: RM-002 should reuse the final interaction contract from RM-001, not race it
 
+## Project Direction Handoff
+
+- Project direction mode: -
+- Direction mode rationale: -
+- Direction-specific questions selected: -
+- Direction-specific questions skipped: -
+- Direction guardrails applied: -
+- Planning posture: -
+- Evidence maturity: -
+
 ## Ready For Req-Plan
 
 - RM-001:

package/docs/examples/pdca-loop/README.md

@@ -4,7 +4,7 @@
 
 - Example version: `1.0.0`
 - Last reviewed: `2026-04-17`
-- Bound skills: `cc-roadmap@5.3.0`, `cc-plan@3.10.2`, `cc-do@1.7.2`, `cc-check@1.12.2`, `cc-act@1.9.1`
+- Bound skills: `cc-roadmap@5.3.1`, `cc-plan@3.10.7`, `cc-do@1.7.2`, `cc-check@1.12.2`, `cc-act@1.9.2`
 
 This folder shows one minimal but complete `cc-roadmap -> cc-plan -> cc-do -> cc-check -> cc-act` loop.
 

package/docs/examples/pdca-loop/ROADMAP.md

@@ -3,7 +3,7 @@
 ## Roadmap Meta
 
 - Roadmap version: `roadmap.v1`
-- Skill version: `5.3.0`
+- Skill version: `5.3.1`
 - Status: `active`
 - Last updated: `2026-04-15`
 - Owner / decider: `product-owner`
@@ -138,6 +138,7 @@
 
 ## Implementation Tracking
 - Roadmap state source: `roadmap.json`
+- Output language: en
 
 <!-- roadmap-tracking:start -->
 | RM-ID | Item | Stage | Priority | Primary Capability | Secondary Capabilities | Expected Spec Delta | Depends On | Status | REQ | Progress |
@@ -147,7 +148,6 @@
 | RM-004 | Collaboration analytics follow-up | - | - | - | - | - | - | - | - | - |
 <!-- roadmap-tracking:end -->
 
-
 ## Technical Architecture
 
 ```mermaid

package/docs/examples/pdca-loop/changes/REQ-001-copy-invite-link/task.md

@@ -4,8 +4,9 @@
 
 - Requirement version: `REQ-001.v1`
 - Design version: `design.v1`
-- CC-Plan skill version: `3.10.2`
+- CC-Plan skill version: `3.10.7`
 - Work branch: `REQ/001-copy-invite-link`
+- Output language: en
 - Source roadmap item: `RM-001`
 - Source roadmap version: `roadmap.v1`
 
@@ -25,6 +26,39 @@
   - Testing decisions: test through share dialog behavior
   - Out of scope: invite generation, role controls, analytics, clipboard fallback redesign
 - AI Leverage Decision Lens: boil-lake; complete the bounded same-dialog copy lake instead of stopping at a happy-path button
+- Product / Creative Discovery:
+  - Worth doing: users already have an invite URL but must manually select and copy it.
+  - Desired product shape: one click in the existing dialog, then visible copied confirmation.
+  - Narrowest wedge: copy the currently visible invite URL only.
+  - 10x / better version: richer share workflow with fallback feedback and analytics, deferred outside this requirement.
+  - Do-nothing consequence: users keep making copy mistakes in the most common sharing path.
+- Second-Move Review:
+  - First good move: add a copy button beside the invite URL.
+  - Simpler move: expose only a browser copy action and keep all share-service contracts unchanged.
+  - Better architecture: centralize all share actions behind a richer share command model later.
+  - Selected move: same-dialog copy action with minimal copied-state confirmation.
+  - Rejected tradeoff: no broader share-service redesign in this bounded requirement.
+- ASCII Branch Chain Analysis:
+  Language rule: connector tokens stay ASCII; node text follows `Output language: en`.
+
+```text
+Requirement Impact Chain
+REQ: one-click copy for the visible invite URL
+|-- Upstream source: RM-001 + share dialog user story
+|-- Current code path: src/features/share/ShareDialog.tsx
+|   |-- caller: share dialog render flow
+|   |-- data or state: existing invite URL prop + copied confirmation state
+|   `-- deepest affected layer: clipboard boundary in the dialog behavior
+|-- Required change: add copy action without changing share-service contracts
+`-- Verification seam: src/features/share/ShareDialog.test.tsx
+
+Business Impact Chain
+OUTCOME: users can share invites without manual text selection
+|-- Direct behavior impact: visible invite URL copies from the dialog
+|-- Downstream impact: tests and PR brief document the bounded copy behavior
+|-- Risk branch: clipboard boundary failure or stale copied-state feedback
+`-- Non-goal branch: invite generation, analytics, and fallback redesign stay out
+```
 - Read first:
   - `design.md`
   - `src/features/share/ShareDialog.tsx`
@@ -45,8 +79,8 @@
 ClaudeCode / Codex 执行本计划时，必须把本文件当成任务模板合同，而不是普通 TODO 列表。
 
 - Template source: `assets/TASKS_TEMPLATE.md`
-- CLI resolver: all workflow commands must run through `.claude/skills/cc-dev/scripts/resolve-cc-devflow.sh` or `.codex/skills/cc-dev/scripts/resolve-cc-devflow.sh`; if it cannot prove `query workflow-context` and `next-change-key`, stop blocked.
-- Context first: run resolved CLI `query workflow-context --change <changeId> --change-key <changeKey> --cwd <repo-root> --data-only --no-trace --compact` before opening deep sections.
+- CLI resolver: all workflow commands must run through `.claude/skills/cc-dev/scripts/resolve-cc-devflow.sh` or `.codex/skills/cc-dev/scripts/resolve-cc-devflow.sh`; if it cannot prove `next-change-key`, stop blocked.
+- Context first: read this `task.md`, current Git status/history, and PR or handoff truth when present before opening deep sections.
 - Task selection: use `scripts/select-ready-tasks.sh --tasks devflow/changes/<change-key>/task.md`.
 - Task block rule: read the full task block before coding; title-only execution is invalid.
 - Completion rule: after verification and review gates pass, run `scripts/mark-task-complete.sh --tasks devflow/changes/<change-key>/task.md --task <task-id>`.
@@ -57,8 +91,8 @@ DEVFLOW=".claude/skills/cc-dev/scripts/resolve-cc-devflow.sh"
 if [[ ! -f "$DEVFLOW" && -f ".codex/skills/cc-dev/scripts/resolve-cc-devflow.sh" ]]; then
   DEVFLOW=".codex/skills/cc-dev/scripts/resolve-cc-devflow.sh"
 fi
-bash "$DEVFLOW" require query workflow-context next-change-key
-bash "$DEVFLOW" query workflow-context --change <changeId> --change-key <changeKey> --cwd <repo-root> --data-only --no-trace --compactSCRIPT_ROOT=".claude/skills/cc-do/scripts"
+bash "$DEVFLOW" require next-change-key
+SCRIPT_ROOT=".claude/skills/cc-do/scripts"
 if [[ ! -d "$SCRIPT_ROOT" && -d ".codex/skills/cc-do/scripts" ]]; then
   SCRIPT_ROOT=".codex/skills/cc-do/scripts"
 fi

package/docs/examples/pdca-loop/roadmap.json

@@ -5,12 +5,17 @@
   },
   "meta": {
     "roadmapVersion": "roadmap.v1",
-    "skillVersion": "5.3.0",
+    "skillVersion": "5.3.1",
     "status": "active",
     "lastUpdated": "2026-04-19",
     "currentFocusStage": "Stage 1"
   },
   "context": {
+    "projectDirectionMode": "",
+    "projectDirectionRationale": "",
+    "directionQuestionsSelected": [],
+    "directionQuestionsSkipped": [],
+    "directionGuardrailsApplied": [],
     "planningPosture": "",
     "evidenceMaturity": "",
     "canonicalTerms": [],
@@ -193,7 +198,7 @@
   },
   "backlogMeta": {
     "roadmapVersion": "roadmap.v1",
-    "skillVersion": "4.4.1",
+    "skillVersion": "5.3.1",
     "currentFocusStage": "Stage 1"
   },
   "dependencyHandoff": {

package/lib/skill-runtime/CLAUDE.md

@@ -4,7 +4,7 @@
 职责分组
 入口层: `cli.js` 负责命令分发，`index.js` 提供给测试和内部脚本的稳定聚合入口。
 基础层: `schemas.js`、`store.js`、`paths.js` 管住契约、持久化与路径规则，避免执行层重复造轮子。
-状态层: `artifacts.js`、`lifecycle.js`、`query.js`、`workflow-context.js`、`review.js`、`team-state.js` 维护运行时真相源与只读查询。
+状态层: `artifacts.js`、`lifecycle.js`、`review.js`、`team-state.js` 维护运行时真相源。
 规划与交接: `planner.js`、`intent.js`、`delegation.js` 把任务解析、handoff 生成和 team/workspace 委派收口成统一语义。
 阶段操作: `operations/` 是唯一 stage 入口目录；具体阶段边界见 `operations/CLAUDE.md`。
 测试布局: `__tests__/` 紧贴模块放置单元、回归与集成测试；顶层 `test/` 不再承载 `skill-runtime` 私有测试。

package/lib/skill-runtime/__tests__/cli-bootstrap.integration.test.js

@@ -133,8 +133,9 @@ describe('cc-devflow cli distribution bootstrap', () => {
     const doctorResult = runCli(['config', 'doctor', '--cwd', repoRoot], repoRoot);
 
     expect(resolveResult.status).toBe(0);
-    expect(resolveResult.stdout).toContain('CC-DevFlow Output Policy');
+    expect(resolveResult.stdout).toContain('CC-DevFlow 输出策略');
     expect(resolveResult.stdout).toContain('Output language: zh-CN');
+    expect(resolveResult.stdout).toContain('标题、正文、占位符、证据说明和 PR/body 草稿必须使用简体中文');
     expect(resolveResult.stdout).toContain('Trace:');
     expect(resolveResult.stdout).toContain('project');
     expect(getResult.status).toBe(0);

package/lib/skill-runtime/__tests__/config.test.js

@@ -100,8 +100,13 @@ describe('skill-runtime user config', () => {
       'cli:output.document_language:zh-CN'
     ]);
     expect(result.policy).toContain('Output language: zh-CN');
-    expect(result.policy).toContain('Machine-enforced fields: output.document_language');
-    expect(result.policy).toContain('Agent advisory preferences');
+    expect(result.policy).toContain('## CC-DevFlow 输出策略');
+    expect(result.policy).toContain('机器强制字段：output.document_language');
+    expect(result.policy).toContain('标题、正文、占位符、证据说明和 PR/body 草稿必须使用简体中文');
+    expect(result.policy).toContain('必须在元数据或首屏包含 `Output language: zh-CN`');
+    expect(result.policy).toContain('Agent 建议偏好：');
+    expect(result.policy).not.toContain('Machine-enforced fields');
+    expect(result.policy).not.toContain('Durable Markdown artifacts must include');
   });
 
   test('rejects unsupported language and unknown top-level fields', () => {

package/lib/skill-runtime/config.js

@@ -24,6 +24,35 @@ const DEFAULT_CONFIG = {
   agent_preferences: {}
 };
 
+const OUTPUT_POLICY_TEXT = {
+  en: {
+    title: '## CC-DevFlow Output Policy',
+    enforced: 'Machine-enforced fields: output.document_language',
+    language: (language) => `- Output language: ${language}`,
+    markdown: (language) =>
+      `- Durable planning, review, and handoff Markdown must write headings, prose, placeholders, evidence summaries, and PR/body drafts in ${language}.`,
+    literals:
+      '- Keep code, commands, paths, schema keys, API names, dependency names, and Conventional Commit type/scope literals unchanged.',
+    metadata: (language) =>
+      `- Durable Markdown artifacts must include \`Output language: ${language}\` in their metadata or first screen.`,
+    preferences: 'Agent advisory preferences:',
+    none: '- none'
+  },
+  'zh-CN': {
+    title: '## CC-DevFlow 输出策略',
+    enforced: '机器强制字段：output.document_language',
+    language: (language) => `- Output language: ${language}`,
+    markdown:
+      '- 所有 durable planning、review、handoff Markdown 的标题、正文、占位符、证据说明和 PR/body 草稿必须使用简体中文。',
+    literals:
+      '- 代码、命令、路径、schema key、API 名、依赖名、Conventional Commit 的 type/scope 等机器字面量保持原文。',
+    metadata: (language) =>
+      `- Durable Markdown artifact 必须在元数据或首屏包含 \`Output language: ${language}\`。`,
+    preferences: 'Agent 建议偏好：',
+    none: '- 无'
+  }
+};
+
 function isPlainObject(value) {
   return Boolean(value) && typeof value === 'object' && !Array.isArray(value);
 }
@@ -221,21 +250,24 @@ function formatPreferenceLines(value, prefix = '') {
 
 function renderOutputPolicy(config = DEFAULT_CONFIG) {
   const documentLanguage = config.output?.document_language || DEFAULT_CONFIG.output.document_language;
+  const text = OUTPUT_POLICY_TEXT[documentLanguage] || OUTPUT_POLICY_TEXT.en;
   const lines = [
-    '## CC-DevFlow Output Policy',
+    text.title,
     '',
-    'Machine-enforced fields: output.document_language',
-    `- Output language: ${documentLanguage}`,
-    `- Durable Markdown artifacts must include \`Output language: ${documentLanguage}\` in their metadata or first screen.`,
+    text.enforced,
+    text.language(documentLanguage),
+    typeof text.markdown === 'function' ? text.markdown(documentLanguage) : text.markdown,
+    text.literals,
+    text.metadata(documentLanguage),
     '',
-    'Agent advisory preferences:'
+    text.preferences
   ];
 
   const preferences = config.agent_preferences || {};
   const preferenceLines = formatPreferenceLines(preferences);
 
   if (preferenceLines.length === 0) {
-    lines.push('- none');
+    lines.push(text.none);
   } else {
     lines.push(...preferenceLines);
   }

package/lib/skill-runtime/index.js

@@ -1,25 +1,17 @@
 /**
  * [INPUT]: 依赖 skill runtime 基础模块。
- * [OUTPUT]: 统一导出 config/path/store/query helpers。
+ * [OUTPUT]: 统一导出 config/path/store helpers。
  * [POS]: skill runtime 模块聚合出口，只保留 CLI 仍需的轻量能力。
  * [PROTOCOL]: 变更时更新此头部，然后检查 CLAUDE.md
  */
 
 const store = require('./store');
-const query = require('./query');
-const queryRegistry = require('./query-registry');
-const errors = require('./errors');
-const trace = require('./trace');
 const paths = require('./paths');
 const config = require('./config');
 const archiveChange = require('./archive-change');
 
 module.exports = {
   ...store,
-  ...query,
-  ...queryRegistry,
-  ...errors,
-  ...trace,
   ...paths,
   ...config,
   ...archiveChange

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "cc-devflow",
-    "version": "4.5.14",
+    "version": "4.5.16",
     "description": "Multi-platform CLI and skill pack for agent coding",
     "main": "bin/cc-devflow.js",
     "bin": {

package/lib/skill-runtime/errors.js

@@ -1,39 +0,0 @@
-/**
- * [INPUT]: 接收 runtime/query/compiler 边界抛出的错误或失败字段。
- * [OUTPUT]: 生成可序列化 named error，保留 refs 与 rescue action。
- * [POS]: skill runtime 的失败语义层，避免用 null/false/string 表达可恢复失败。
- * [PROTOCOL]: 变更时更新此头部，然后检查 CLAUDE.md
- */
-
-class SkillRuntimeError extends Error {
-  constructor(name, message, options = {}) {
-    super(message);
-    this.name = name;
-    this.artifactRefs = options.artifactRefs || [];
-    this.rescueAction = options.rescueAction || 'inspect-task-md-and-git-history';
-    this.details = options.details || {};
-  }
-}
-
-function namedError(name, message, options = {}) {
-  return new SkillRuntimeError(name, message, options);
-}
-
-function serializeError(error, fallbackName = 'SkillRuntimeError') {
-  const name = error?.name || fallbackName;
-  const message = error?.message || String(error || 'Unknown runtime error');
-
-  return {
-    name,
-    message,
-    artifactRefs: error?.artifactRefs || [],
-    rescueAction: error?.rescueAction || 'inspect-task-md-and-git-history',
-    details: error?.details || {}
-  };
-}
-
-module.exports = {
-  SkillRuntimeError,
-  namedError,
-  serializeError
-};

package/lib/skill-runtime/query-registry.js

@@ -1,101 +0,0 @@
-/**
- * [INPUT]: 接收 query id、repoRoot/changeId 与只读 handler registry。
- * [OUTPUT]: 返回 typed query result：ok/data 或 named error，并附 trace。
- * [POS]: skill runtime 的查询分发表，只读派生 task.md 与 Git 语义。
- * [PROTOCOL]: 变更时更新此头部，然后检查 CLAUDE.md
- */
-
-const fs = require('fs');
-
-const { namedError, serializeError } = require('./errors');
-const { createTrace } = require('./trace');
-
-function resolveArtifactRefs(entry, context, key) {
-  const refs = entry[key];
-  return typeof refs === 'function' ? refs(context) : refs || [];
-}
-
-function createQueryRegistry(entries) {
-  const registry = new Map(entries.map((entry) => [entry.id, entry]));
-
-  function listQueryIds() {
-    return [...registry.keys()].sort();
-  }
-
-  async function runQuery(queryId, context = {}) {
-    const entry = registry.get(queryId);
-
-    if (!entry) {
-      const supported = listQueryIds();
-      const error = namedError(
-        'UnknownQueryError',
-        `Unknown query id: ${queryId}`,
-        {
-          rescueAction: `use one of: ${supported.join(', ')}`,
-          details: { supported }
-        }
-      );
-
-      return {
-        ok: false,
-        queryId,
-        error: serializeError(error),
-        trace: createTrace({
-          event: 'query.unknown',
-          changeId: context.changeId,
-          nextAction: 'choose-supported-query'
-        })
-      };
-    }
-
-    let artifactRefs = [];
-    try {
-      artifactRefs = resolveArtifactRefs(entry, context, 'artifactRefs');
-      const requiredArtifactRefs = resolveArtifactRefs(entry, context, 'requiredArtifactRefs');
-      const missingRefs = requiredArtifactRefs.filter((ref) => !fs.existsSync(ref));
-      if (missingRefs.length > 0) {
-        throw namedError(
-          'MissingQueryArtifactError',
-          `Missing required query artifact: ${missingRefs.join(', ')}`,
-          {
-            artifactRefs: missingRefs,
-            rescueAction: 'create task.md before running this query'
-          }
-        );
-      }
-
-      return {
-        ok: true,
-        queryId,
-        data: await entry.handler(context),
-        trace: createTrace({
-          event: `query.${queryId}`,
-          changeId: context.changeId,
-          artifactRefs,
-          nextAction: entry.nextAction || 'read-query-result'
-        })
-      };
-    } catch (error) {
-      return {
-        ok: false,
-        queryId,
-        error: serializeError(error),
-        trace: createTrace({
-          event: `query.${queryId}.failed`,
-          changeId: context.changeId,
-          artifactRefs,
-          nextAction: error.rescueAction || 'inspect-task-md-and-git-history'
-        })
-      };
-    }
-  }
-
-  return {
-    listQueryIds,
-    runQuery
-  };
-}
-
-module.exports = {
-  createQueryRegistry
-};