claude-memory-layer 1.0.11 → 1.0.13

package/AGENTS.md

@@ -0,0 +1,60 @@
+# Codex / AI Agent Instructions
+
+이 저장소는 **Claude Code용 메모리 플러그인 + CLI + 로컬 대시보드 서버**입니다.
+Codex(또는 다른 코딩 에이전트)가 작업할 때 아래 규칙을 우선으로 따르세요.
+
+## Quick Commands
+
+```bash
+npm install
+npm run build
+
+# 웹 대시보드(로컬)
+node dist/server/index.js
+# 또는
+npx claude-memory-layer dashboard
+```
+
+## Key Entry Points
+
+- `scripts/build.ts`: esbuild 기반 빌드. `dist/` 생성 + `.claude-plugin/` 및 `src/ui` 복사
+- `src/cli/index.ts`: CLI 엔트리포인트 (`bin`: `claude-memory-layer`)
+- `src/server/index.ts`: 대시보드/REST API 서버 (기본 포트 `37777`)
+- `src/server/api/chat.ts`: **외부 `claude` CLI**를 `spawn('claude', ...)`로 호출 (환경에 따라 미설치일 수 있음)
+
+## Local Side-Effects (주의)
+
+- `claude-memory-layer install` / `uninstall` 은 **사용자 머신의** `~/.claude/settings.json` 을 수정합니다.
+  - 자동 실행/테스트 목적으로 임의 실행하지 말고, 사용자가 명시적으로 요청할 때만 실행하세요.
+- 훅/CLI/서버는 기본적으로 `~/.claude-code/memory` 아래에 데이터를 읽고/씁니다.
+  - 메모리 스토리지를 삭제/초기화하는 동작은 요청 없이는 하지 마세요.
+
+## Common Pitfall: Self-dependency (반복 발생)
+
+`package.json`의 `dependencies`에 **자기 자신**을 추가하면 설치가 꼬입니다.
+
+- 금지 예: `"claude-memory-layer": "^1.0.x"`
+- 진단: `npm list claude-memory-layer` 결과가 `(empty)` 여야 정상
+- 복구:
+
+```bash
+rm package-lock.json && rm -rf node_modules
+npm install
+npm list claude-memory-layer
+```
+
+## Release / Publish Workflow
+
+1. Self-dependency 확인: `npm list claude-memory-layer` → `(empty)`
+2. 버전 업데이트: `npm version patch` 또는 `package.json` 수정
+3. 빌드: `npm run build`
+4. 배포: `npm publish --otp=<code>`
+5. 확인: `npm view claude-memory-layer versions`
+
+## Dev Notes
+
+- Node.js: `>=18` (ESM 프로젝트)
+- 스토리지: **SQLite(WAL) 기반**으로 훅(쓰기) + 서버(읽기) 동시 동작을 전제로 함
+- 테스트/정적분석 스크립트는 존재하지만, 현재 시점에서 전부 “항상 green”을 보장하지 않을 수 있으니
+  변경 전후로 실제 실행 결과를 확인하고 진행하세요.
+

package/README.md

@@ -10,6 +10,70 @@ Claude Memory Layer는 Claude Code에서 사용자와 AI 간의 모든 대화를
 - **프로젝트 맥락 이해**: 프로젝트별로 축적된 지식 활용
 - **개인화된 응답**: 사용자의 선호도와 패턴 학습
 
+## 빠른 시작 (신규 프로젝트 기준)
+
+아래 순서대로 하면 됩니다.
+
+### 0) 최초 1회(머신 전체) 설치
+
+```bash
+cd ~/workspace/claude-memory-layer
+npm install
+npm run build
+npx claude-memory-layer install
+```
+
+- `install`은 **한 번만** 하면 됩니다(Claude Code hooks 등록).
+- 이후 프로젝트별로 메모리 저장소가 자동 분리됩니다.
+
+### 1) 새 프로젝트에서 초기 메모리 생성
+
+```bash
+cd /path/to/your-project
+npx claude-memory-layer import
+```
+
+- 현재 프로젝트의 기존 Claude 세션(`~/.claude/projects/...`)을 읽어와 메모리로 적재합니다.
+- 벡터 임베딩까지 한 번에 처리됩니다.
+
+### 2) 사용 중 확인
+
+```bash
+# 프로젝트 메모리 검색
+npx claude-memory-layer search "인증 구조"
+
+# 통계 확인
+npx claude-memory-layer stats
+
+# 대시보드 실행
+npx claude-memory-layer dashboard
+```
+
+### 3) 다른 프로젝트에도 동일하게 적용?
+
+네. **완전히 동일한 흐름**을 각 프로젝트에서 반복하면 됩니다.
+
+```bash
+cd /path/to/another-project
+npx claude-memory-layer import
+npx claude-memory-layer search "배포 이슈"
+```
+
+프로젝트마다 내부적으로 별도 저장소(`~/.claude-code/memory/projects/<hash>`)를 사용하므로,
+기억이 자동으로 분리됩니다.
+
+### 4) 운영 팁
+
+- 특정 프로젝트를 명시하고 싶으면 대부분 명령에 `--project <path>` 사용 가능
+- 대규모 리임포트가 필요하면 `import --force` 사용
+- 백그라운드 worker가 못 처리한 임베딩은 `process`로 수동 처리
+- 상태 점검:
+  - `GET /health` (서버 헬스)
+  - `GET /api/health` (outbox pending/failed 포함 상세 헬스)
+  - `GET /api/stats/retrieval-traces` (검색→컨텍스트 채택 추적)
+
+---
+
 ## Features
 
 ### Core Features
@@ -31,6 +95,9 @@ Claude Memory Layer는 Claude Code에서 사용자와 AI 간의 모든 대화를
 - **PostToolUse Hook**: 도구 실행 결과 (Read, Write, Bash 등) 캡처 및 저장
 - **Private Tags**: `<private>` 태그로 민감 정보를 명시적으로 제외
 - **Progressive Disclosure**: 3-Layer 검색 (인덱스 → 타임라인 → 상세)으로 토큰 효율화
+- **memU-inspired Retrieval**: fast/deep 전략 + 스코프 필터(session prefix, canonical key prefix, metadata path)
+- **Append-only Markdown Mirror**: 저장 이벤트를 `memory/<namespace>/<category...>/YYYY-MM-DD.md`에도 동기 append (기본값: `namespace=default`, `category=uncategorized`, 경로 세그먼트 sanitize)
+  - `memory/_index.md` 인덱스를 자동 갱신
 - **Task Entity System**: Task를 Entity로 승격하여 세션 간 상태 추적
 - **Vector Outbox V2**: Transactional Outbox 패턴으로 DuckDB-LanceDB 정합성 보장
 - **Web Viewer UI**: localhost:37777 대시보드로 실시간 메모리 모니터링
@@ -100,6 +167,42 @@ Claude Code 내에서 사용할 수 있는 명령어:
 /memory-forget --session <id> --confirm
 ```
 
+### 기존 메모리 정리 Import (구조화)
+
+레거시 markdown 메모리를 읽어서 구조화 경로로 재저장(import)할 수 있습니다.
+
+```bash
+# 미리보기(실제 저장 없음)
+claude-memory-layer organize-import /path/to/legacy-memory --dry-run
+
+# 실제 import
+claude-memory-layer organize-import /path/to/legacy-memory --project /path/to/project
+
+# 일부만 import
+claude-memory-layer organize-import /path/to/legacy-memory --limit 100
+
+# source에 markdown이 없으면 자동 bootstrap(코드+git 분석)
+claude-memory-layer organize-import /path/to/empty-dir --project /path/to/project
+
+# bootstrap 강제 실행
+claude-memory-layer organize-import /path/to/memory --force-bootstrap --repo /path/to/project
+
+# 자동 bootstrap 비활성화
+claude-memory-layer organize-import /path/to/empty-dir --no-bootstrap-if-empty
+
+# markdown이 없는 초기 상태면 bootstrap 생성 + import
+claude-memory-layer organize-import /path/to/empty-dir --bootstrap --repo /path/to/codebase
+
+# bootstrap 강제 재생성 (기존 markdown 있어도)
+claude-memory-layer organize-import /path/to/legacy-memory --force-bootstrap --repo /path/to/codebase --out /path/to/legacy-memory/bootstrap-kb
+
+# 증분 bootstrap (기본값): 이전 manifest를 기준으로 변경분 중심 업데이트
+claude-memory-layer organize-import /path/to/legacy-memory --bootstrap --repo /path/to/codebase --incremental
+
+# 전체 재생성 bootstrap
+claude-memory-layer organize-import /path/to/legacy-memory --bootstrap --repo /path/to/codebase --no-incremental
+```
+
 ### CLI 명령어
 
 터미널에서 직접 사용:
@@ -127,8 +230,65 @@ npx claude-memory-layer list --project /path/to/project
 
 # 임베딩 수동 처리
 npx claude-memory-layer process
+
+# MongoDB 동기화 (옵션, 멀티 서버 협업)
+# - 여러 서버에서 같은 프로젝트를 개발할 때, 각 서버의 로컬 SQLite(events.sqlite) 이벤트를
+#   하나의 MongoDB로 모아 push/pull 동기화할 수 있습니다.
+# - 동일 프로젝트는 반드시 같은 project key로 실행해야 합니다.
+export CLAUDE_MEMORY_MONGO_URI="mongodb://USER:PASSWORD@HOST:PORT/"
+export CLAUDE_MEMORY_MONGO_DB="claude_memory_layer"
+export CLAUDE_MEMORY_MONGO_PROJECT="my-project"
+
+# 1회 동기화 (push+pull)
+npx claude-memory-layer mongo-sync
+
+# 지속 동기화 (주기적으로 push+pull)
+npx claude-memory-layer mongo-sync --watch --interval 30000
+```
+
+### memU-inspired Retrieval 사용 예시
+
+아래 예시는 SDK/서비스 레벨에서 `retrieveMemories()` 호출 시 적용되는 옵션입니다.
+
+```ts
+import { getMemoryServiceForProject } from './src/services/memory-service.js';
+
+const memory = getMemoryServiceForProject('/path/to/project');
+
+// 1) Fast: 키워드 기반 빠른 검색
+const fast = await memory.retrieveMemories('브리핑 포맷', {
+  strategy: 'fast',
+  topK: 5,
+  minScore: 0.6
+});
+
+// 2) Deep: 벡터 검색 + 키워드 오버랩 재정렬
+const deep = await memory.retrieveMemories('브리핑 포맷', {
+  strategy: 'deep',
+  topK: 10,
+  rerankWithKeyword: true
+});
+
+// 3) Scoped filter: 세션/타입/계층형 메타데이터로 범위 제한
+const scoped = await memory.retrieveMemories('아침 브리핑', {
+  strategy: 'deep',
+  scope: {
+    sessionIdPrefix: 'agent:main:',
+    eventTypes: ['user_prompt', 'agent_response'],
+    canonicalKeyPrefix: 'pref/briefing',
+    contentIncludes: ['아침'],
+    metadata: {
+      'scope.project.id': 'alpha'
+    }
+  }
+});
 ```
 
+팁:
+- `strategy: 'auto'`는 기본적으로 fallback 체인을 사용해 결과를 찾습니다.
+- 저지연 응답이 중요하면 `fast`, 정확도 우선이면 `deep`를 권장합니다.
+- 프로젝트 서비스(`getMemoryServiceForProject`) 기준 검색 스코프는 기본적으로 project-aware(엄격 모드)로 동작합니다.
+
 ## Privacy 기능
 
 ### Private Tags
@@ -250,6 +410,9 @@ claude-memory-layer dashboard
 - 벡터 검색 인터페이스
 - 저장소 통계 대시보드
 - Outbox 상태 모니터링
+- Retrieval Trace (검색 질의 → 후보/채택 수 → 최종 context 채택 IDs) 1:1 확인
+- 채택된 event ID 클릭으로 상세 모달 확인 + score breakdown(semantic/lexical/recency) 확인
+- 후보(candidate) event ID 스냅샷도 함께 노출되어 검색→채택 전 과정을 추적 가능
 
 ## 기존 대화 기록 임포트
 
@@ -429,8 +592,9 @@ Layer 1: Search Index (~50-100 tokens/result)
 
 ```
 ~/.claude-code/memory/
-├── events.duckdb     # 이벤트 저장소
-└── vectors/          # 벡터 임베딩
+├── projects/<hash>/events.sqlite  # 프로젝트별 이벤트 저장소 (primary)
+├── projects/<hash>/vectors/        # 프로젝트별 벡터 임베딩
+└── shared/                         # (옵션) 공유 지식 저장소
 ```
 
 Claude Code 세션 기록 위치:

package/bootstrap-kb/decisions/decisions.md

@@ -0,0 +1,244 @@
+# Decisions (extracted)
+
+- confidence: high
+
+## 2026-01-31 | feat: implement complete code-memory plugin
+- status: active (inferred)
+- source: commit:111f400e73f0568b192a5b1b37a88622fd44d648
+- author: Claude
+- changedFiles: 31
+- confidence: mid
+
+## 2026-01-31 | feat: add session history import functionality
+- status: active (inferred)
+- source: commit:22b1c43d9b9f095cd2a210f529ca8bee5a1100c5
+- author: Claude
+- changedFiles: 5
+- confidence: mid
+
+## 2026-02-01 | docs: Update README with comprehensive feature documentation (#3)
+- status: active (inferred)
+- source: commit:3662dfae8381fdf35923b115f58c362e091c474a
+- author: justinbuzzni
+- changedFiles: 1
+- confidence: mid
+
+## 2026-02-01 | feat: Add cross-project shared troubleshooting store
+- status: active (inferred)
+- source: commit:d04e562b5b011e0f07cdea0664db823f0caca5d0
+- author: justinbuzzni
+- changedFiles: 9
+- confidence: mid
+
+## 2026-02-01 | refactor: Extract DuckDB wrapper for consistent async API
+- status: active (inferred)
+- source: commit:65b0f64da3fd2c4b0b758b89f9d0f22301df23a7
+- author: justinbuzzni
+- changedFiles: 19
+- confidence: mid
+
+## 2026-02-01 | feat: Add web dashboard for memory visualization
+- status: active (inferred)
+- source: commit:665cf11bb9a01f1bcba51ec49f53382e95cd7e32
+- author: justinbuzzni
+- changedFiles: 5
+- confidence: mid
+
+## 2026-02-01 | refactor: Rename project from code-memory to claude-memory-layer
+- status: active (inferred)
+- source: commit:f35da2b92704d7d1b5655edac3516c64a829eeef
+- author: justinbuzzni
+- changedFiles: 7
+- confidence: mid
+
+## 2026-02-01 | feat: Migrate primary store from DuckDB to SQLite (WAL mode)
+- status: active (inferred)
+- source: commit:be36be70f69d42d2373ac26e6c0b8e8b53ca0340
+- author: justinbuzzni
+- changedFiles: 8
+- confidence: mid
+
+## 2026-02-01 | feat: Enhance dashboard, CLI, and memory graduation system
+- status: active (inferred)
+- source: commit:babfd15284754bc3dbb698fcc4c81efcf53fcd85
+- author: justinbuzzni
+- changedFiles: 13
+- confidence: mid
+
+## 2026-02-07 | feat: Add memory helpfulness tracking and dashboard sort controls
+- status: active (inferred)
+- source: commit:794640023652fc38b010e1ab46ca6a9a504a09e8
+- author: justinbuzzni
+- changedFiles: 9
+- confidence: mid
+
+## 2026-02-08 | feat: Add interactive dashboard with modals, navigation, and detail views
+- status: active (inferred)
+- source: commit:20c37b2cdffd271eaf04df9d2e87e5f23a4f853e
+- author: justinbuzzni
+- changedFiles: 5
+- confidence: mid
+
+## 2026-02-08 | fix: Align hooks with actual Claude Code input format and use lightweight service
+- status: active (inferred)
+- source: commit:a1d93af414821eb92d74cd02abb748849366ebfb
+- author: justinbuzzni
+- changedFiles: 4
+- confidence: mid
+
+## 2026-02-08 | feat: Add real-time progress display for import command
+- status: active (inferred)
+- source: commit:fd8f38af7cecf7d31a32a5aba9b1902178f986b1
+- author: justinbuzzni
+- changedFiles: 2
+- confidence: mid
+
+## 2026-02-08 | feat: Add turn-based event grouping, force reimport, and session registry support
+- status: active (inferred)
+- source: commit:938ec34bdf5fe1e4c43274424497ca081d62023a
+- author: justinbuzzni
+- changedFiles: 9
+- confidence: mid
+
+## 2026-02-08 | feat: Add per-project dashboard with project selector and turns API
+- status: active (inferred)
+- source: commit:af18cee5277324c78a9e1bca5873bc86ee5b2719
+- author: justinbuzzni
+- changedFiles: 12
+- confidence: mid
+
+## 2026-02-08 | fix: Handle vector store table race condition and auto-create SQLite directories
+- status: active (inferred)
+- source: commit:5b7401917611b50ee49e38bdf080983d1c763d5e
+- author: justinbuzzni
+- changedFiles: 2
+- confidence: mid
+
+## 2026-02-08 | feat: Add Ask Memory chat service to dashboard
+- status: active (inferred)
+- source: commit:a7b260f0ceeff6d768fefb675e6425f5b4717d7e
+- author: justinbuzzni
+- changedFiles: 5
+- confidence: mid
+
+## 2026-02-08 | feat: Widen chat panel and add conversation history with localStorage
+- status: active (inferred)
+- source: commit:15547d41e952350d095dfe3c01b2afa749638532
+- author: justinbuzzni
+- changedFiles: 3
+- confidence: mid
+
+## 2026-02-08 | fix: Knowledge Graph shows data even when no memories have been accessed
+- status: active (inferred)
+- source: commit:78b0956610eb7d22c28a821c67e10e6e9996c5be
+- author: justinbuzzni
+- changedFiles: 2
+- confidence: mid
+
+## 2026-02-20 | feat(ops): add heartbeat orchestrator and sync-gap auto-heal scripts
+- status: active (inferred)
+- source: commit:d56d22ba3271c1d51ea8bbaddb21f5f3d0a79d49
+- author: namsangboy
+- changedFiles: 7
+- confidence: mid
+
+## 2026-02-09 | feat: optional MongoDB sync for project events
+- status: active (inferred)
+- source: commit:11d3092e372836f784a760fc0b4269fd44e9159f
+- author: Justin
+- changedFiles: 9
+- confidence: mid
+
+## 2026-02-22 | feat: add Codex session history importer, MongoDB sync config, and AGENTS.md
+- status: active (inferred)
+- source: commit:4855172b8f92fff0c148c106de4e10c494d2bb09
+- author: Justin
+- changedFiles: 4
+- confidence: mid
+
+## 2026-02-24 | feat(memory): adopt memU-style scoped retrieval and staged ingest hooks
+- status: active (inferred)
+- source: commit:d278200d0f0666616a8556f2f25f9758b3ab9afd
+- author: justinbuzzni
+- changedFiles: 6
+- confidence: mid
+
+## 2026-02-24 | feat(memory): add append-only markdown mirror for categorized event archives
+- status: active (inferred)
+- source: commit:917ad7fc97a1c2de68637dfbe4b4cdd51776ec30
+- author: justinbuzzni
+- changedFiles: 4
+- confidence: mid
+
+## 2026-02-24 | feat: append-only markdown mirror in sqlite ingest flow
+- status: active (inferred)
+- source: commit:72250046152e851ed70476a8704ca38ca633ffab
+- author: justinbuzzni
+- changedFiles: 3
+- confidence: mid
+
+## 2026-02-24 | feat(memory): auto-refresh markdown mirror index
+- status: active (inferred)
+- source: commit:3a34d45a594affb034d6e416e3a29b6d2dcc20a5
+- author: justinbuzzni
+- changedFiles: 3
+- confidence: mid
+
+## 2026-02-24 | feat(cli): add organize-import command for structured markdown memory import
+- status: active (inferred)
+- source: commit:fe9ef67e5acabb81885a3c752a5b8525f80856a5
+- author: justinbuzzni
+- changedFiles: 3
+- confidence: mid
+
+## 2026-02-24 | feat(retrieval): add auto fallback chain fast→deep→scope-expand→summary
+- status: active (inferred)
+- source: commit:0dceb2d914e67e10d140848340f30e6fd07c4f67
+- author: justinbuzzni
+- changedFiles: 2
+- confidence: mid
+
+## 2026-02-24 | feat(ranking): add helpfulness-driven adaptive rerank weights v1
+- status: active (inferred)
+- source: commit:7de02967a1d59aea3e04ef1239585b93828c503a
+- author: justinbuzzni
+- changedFiles: 3
+- confidence: mid
+
+## 2026-02-24 | feat(retrieval): add ttl/decay ranking policy v1
+- status: active (inferred)
+- source: commit:46f2cfa2d99f0c8fb3a48709a97f62800f01d4fe
+- author: justinbuzzni
+- changedFiles: 2
+- confidence: mid
+
+## 2026-02-24 | feat(retrieval): add intent-rewrite deep retrieval merge flow
+- status: active (inferred)
+- source: commit:5fd950beee4d02f00b0c93be668a2e114ce9b257
+- author: justinbuzzni
+- changedFiles: 3
+- confidence: mid
+
+## 2026-02-24 | feat(retrieval): add graph-hop expansion with hop penalty
+- status: active (inferred)
+- source: commit:d1dfe03f881d1a76ecd9a916cbf04c3ff2a4bf7e
+- author: justinbuzzni
+- changedFiles: 2
+- confidence: mid
+
+## 2026-02-24 | feat(consolidation): add hierarchical summary→rule automation with report
+- status: active (inferred)
+- source: commit:7c84b5eacff733d8cb692eb9d47b7297a0872e43
+- author: justinbuzzni
+- changedFiles: 6
+- confidence: mid
+
+## 2026-02-25 | Feat/organize import bootstrap kb (#10)
+- status: active (inferred)
+- source: commit:db38c7e606a59478c03bac7e5bb5951954ca0813
+- author: justinbuzzni
+- changedFiles: 4
+- confidence: mid
+
+## Sources
+- source: git-log:since=180 days ago;max=1000

package/bootstrap-kb/glossary/glossary.md

@@ -0,0 +1,46 @@
+# Glossary (auto-extracted)
+
+- confidence: mid
+
+- context
+- spec
+- plan
+- memory
+- store
+- sync
+- worker
+- retriever
+- event
+- mirror
+- session
+- shared
+- vector
+- gap
+- history
+- matcher
+- sqlite
+- task
+- aligner
+- auto
+- bootstrap
+- canonical
+- config
+- consolidation
+- evidence
+- graduation
+- importer
+- ingest
+- interceptor
+- key
+- markdown
+- mongo
+- organizer
+- package
+- repo
+- resolver
+- search
+- stats
+- wrapper
+
+## Sources
+- source: repo-scan:/home/walter/workspace/opensource/claude-memory-layer

package/bootstrap-kb/modules/.claude-plugin.md

@@ -0,0 +1,22 @@
+# Module: .claude-plugin
+
+- responsibility: inferred from top-level path `.claude-plugin/`
+- files: 8
+- languages: Markdown, JSON
+- confidence: high
+
+## Entry Candidates
+- none detected
+
+## Related Commits (recent sample)
+- 2026-01-31 111f400e feat: implement complete code-memory plugin
+- 2026-01-31 22b1c43d feat: add session history import functionality
+- 2026-02-01 504ec3dd docs: Add citations and endless mode specification documents (#2)
+- 2026-02-01 f35da2b9 refactor: Rename project from code-memory to claude-memory-layer
+
+## Sources
+- source: repo-path:.claude-plugin/**
+- source: commit:111f400e73f0568b192a5b1b37a88622fd44d648
+- source: commit:22b1c43d9b9f095cd2a210f529ca8bee5a1100c5
+- source: commit:504ec3ddd7f451fabf2cf01e073fc96e65ae124e
+- source: commit:f35da2b92704d7d1b5655edac3516c64a829eeef

package/bootstrap-kb/modules/agents.md.md

@@ -0,0 +1,15 @@
+# Module: AGENTS.md
+
+- responsibility: inferred from top-level path `AGENTS.md/`
+- files: 1
+- languages: Markdown
+- confidence: low
+
+## Entry Candidates
+- none detected
+
+## Related Commits (recent sample)
+- none in selected range
+
+## Sources
+- source: repo-path:AGENTS.md/**

package/bootstrap-kb/modules/claude.md.md

@@ -0,0 +1,15 @@
+# Module: CLAUDE.md
+
+- responsibility: inferred from top-level path `CLAUDE.md/`
+- files: 1
+- languages: Markdown
+- confidence: low
+
+## Entry Candidates
+- none detected
+
+## Related Commits (recent sample)
+- none in selected range
+
+## Sources
+- source: repo-path:CLAUDE.md/**

package/bootstrap-kb/modules/context.md.md

@@ -0,0 +1,15 @@
+# Module: context.md
+
+- responsibility: inferred from top-level path `context.md/`
+- files: 1
+- languages: Markdown
+- confidence: low
+
+## Entry Candidates
+- none detected
+
+## Related Commits (recent sample)
+- none in selected range
+
+## Sources
+- source: repo-path:context.md/**

package/bootstrap-kb/modules/docs.md

@@ -0,0 +1,18 @@
+# Module: docs
+
+- responsibility: inferred from top-level path `docs/`
+- files: 2
+- languages: Markdown
+- confidence: mid
+
+## Entry Candidates
+- none detected
+
+## Related Commits (recent sample)
+- 2026-02-20 d56d22ba feat(ops): add heartbeat orchestrator and sync-gap auto-heal scripts
+- 2026-02-24 4447e64f docs(memory): document memU-inspired retrieval strategies and scoped filters
+
+## Sources
+- source: repo-path:docs/**
+- source: commit:d56d22ba3271c1d51ea8bbaddb21f5f3d0a79d49
+- source: commit:4447e64fa00065bfb364d38c597ad87484770891

package/bootstrap-kb/modules/handoff.md.md

@@ -0,0 +1,15 @@
+# Module: HANDOFF.md
+
+- responsibility: inferred from top-level path `HANDOFF.md/`
+- files: 1
+- languages: Markdown
+- confidence: low
+
+## Entry Candidates
+- none detected
+
+## Related Commits (recent sample)
+- none in selected range
+
+## Sources
+- source: repo-path:HANDOFF.md/**

package/bootstrap-kb/modules/package-lock.json.md

@@ -0,0 +1,15 @@
+# Module: package-lock.json
+
+- responsibility: inferred from top-level path `package-lock.json/`
+- files: 1
+- languages: JSON
+- confidence: low
+
+## Entry Candidates
+- none detected
+
+## Related Commits (recent sample)
+- none in selected range
+
+## Sources
+- source: repo-path:package-lock.json/**