claude-mem-lite 2.18.0 → 2.19.0

package/.claude-plugin/marketplace.json

@@ -10,7 +10,7 @@
   "plugins": [
     {
       "name": "claude-mem-lite",
-      "version": "2.18.0",
+      "version": "2.19.0",
       "source": "./",
       "description": "Lightweight persistent memory system for Claude Code — FTS5 search, episode batching, error-triggered recall"
     }

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-mem-lite",
-  "version": "2.18.0",
+  "version": "2.19.0",
   "description": "Lightweight persistent memory system for Claude Code — FTS5 search, episode batching, error-triggered recall",
   "author": {
     "name": "sdsrss"

package/README.md

@@ -100,6 +100,12 @@ The original sends **everything to the LLM and hopes it filters well**. claude-m
 - **LLM concurrency control** -- File-based semaphore limits background workers to 2 concurrent LLM calls, preventing resource contention
 - **stdin overflow protection** -- Hook input truncated at 256KB with regex-based action salvage for oversized tool outputs
 - **Cross-session handoff** -- Captures session state (request, completed work, next steps, key files) on `/clear` or `/exit`, then injects context when the next session detects continuation intent via explicit keywords or FTS5 term overlap
+- **In-place observation updates** -- `mem_update` tool modifies existing observations atomically (field update + FTS text rebuild + vector re-computation in one transaction), preserving original IDs and references
+- **Bulk export** -- `mem_export` tool exports observations as JSON or JSONL, with project/type/date filtering and 1000-row pagination cap with batch guidance
+- **FTS integrity management** -- `mem_fts_check` tool verifies FTS5 index health or rebuilds indexes on demand, useful after database recovery or when search results seem wrong
+- **Atomic multi-table writes** -- `saveObservation` wraps observations + observation_files + observation_vectors INSERTs in a single `db.transaction()`, preventing orphaned rows on crash
+- **Modular NLP pipeline** -- Synonym maps, stop words, scoring constants, and query building extracted into focused modules (`synonyms.mjs`, `stop-words.mjs`, `scoring-sql.mjs`, `nlp.mjs`) for independent testing and maintenance
+- **Porter-aligned PRF** -- Pseudo-relevance feedback terms are now stemmed with the same Porter algorithm used by FTS5, ensuring PRF expansion terms match the search index
 
 ## Platform Support
 
@@ -148,7 +154,7 @@ Source files stay in the cloned repo. Update via `git pull && node install.mjs i
 ### What happens during installation
 
 1. **Install dependencies** -- `npm install --omit=dev` (compiles native `better-sqlite3`)
-2. **Register MCP server** -- `mem` server with 9 tools (search, timeline, get, save, stats, delete, compress, maintain, registry)
+2. **Register MCP server** -- `mem` server with 12 tools (search, timeline, get, save, update, stats, delete, compress, maintain, registry, export, fts_check)
 3. **Configure hooks** -- `PostToolUse`, `SessionStart`, `Stop`, `UserPromptSubmit` lifecycle hooks
 4. **Create data directory** -- `~/.claude-mem-lite/` (hidden) for database, runtime, and managed resource files
 5. **Auto-migrate** -- If `~/.claude-mem/` (original claude-mem) or `~/claude-mem-lite/` (pre-v0.5 unhidden) exists, migrates database and runtime files to `~/.claude-mem-lite/`, preserving the original untouched
@@ -204,10 +210,13 @@ rm -rf ~/claude-mem-lite/   # pre-v0.5 unhidden (if not auto-moved)
 | `mem_timeline` | Browse observations chronologically around an anchor point. |
 | `mem_get` | Retrieve full details for specific observation IDs (includes importance and related_ids). |
 | `mem_save` | Manually save a memory/observation. |
+| `mem_update` | Update an existing observation in-place. Preserves original ID and references. |
 | `mem_stats` | View statistics: counts, type distribution, top projects, daily activity. |
 | `mem_delete` | Delete observations by ID with preview/confirm workflow. FTS5 cleanup is automatic. |
 | `mem_compress` | Compress old low-value observations into weekly summaries to reduce noise. |
 | `mem_maintain` | Memory maintenance: scan for duplicates/stale/broken items, then execute cleanup/dedup/rebuild_vectors operations. |
+| `mem_export` | Export observations as JSON or JSONL for backup or migration. Filters by project, type, date range. |
+| `mem_fts_check` | Check FTS5 index integrity or rebuild indexes. Use when search results seem wrong or after DB recovery. |
 | `mem_registry` | Manage resource registry: search for skills/agents by need, list resources, view stats, import/remove tools, reindex. |
 
 ### Skill Commands (in Claude Code chat)
@@ -238,7 +247,8 @@ Five core tables with FTS5 virtual tables for search:
 id, memory_session_id, project, type, title, subtitle,
 text, narrative, concepts, facts, files_read, files_modified,
 importance, related_ids, created_at, created_at_epoch,
-lesson_learned, minhash_sig, access_count, compressed_into, search_aliases
+lesson_learned, minhash_sig, access_count, compressed_into, search_aliases,
+branch, superseded_at, superseded_by, last_accessed_at
 ```
 
 **session_summaries** -- LLM-generated session summaries
@@ -265,6 +275,11 @@ project, type, session_id, working_on, completed, unfinished,
 key_files, key_decisions, match_keywords, created_at_epoch
 ```
 
+**observation_files** -- Normalized file membership for efficient file-based recall
+```
+obs_id, filename
+```
+
 **observation_vectors** -- TF-IDF vector embeddings for hybrid search
 ```
 observation_id, vector (BLOB Float32Array), vocab_version, created_at_epoch
@@ -422,7 +437,16 @@ claude-mem-lite/
   tool-schemas.mjs     # Shared Zod schemas for MCP tool validation
   tfidf.mjs            # TF-IDF vector engine: tokenization, vocabulary building, vector computation, cosine similarity, RRF merge
   tier.mjs             # Temporal tier system: activity-based time window classification
-  utils.mjs            # Shared utilities: FTS5 query building, BM25 weight constants, MinHash dedup, secret scrubbing, CJK synonym extraction
+  utils.mjs            # Re-export hub: backward-compatible surface for all utility modules
+  nlp.mjs              # FTS5 query building: synonym expansion, CJK bigrams, sanitization
+  scoring-sql.mjs      # BM25 weight constants and type-differentiated decay half-lives
+  stop-words.mjs       # Shared base stop-word set for all NLP/search modules
+  synonyms.mjs         # Unified synonym source: SYNONYM_MAP (bidirectional) + DISPATCH_SYNONYMS
+  project-utils.mjs    # Shared project name resolution with in-process cache
+  secret-scrub.mjs     # API key, token, PEM, and credential pattern redaction
+  format-utils.mjs     # String formatting: truncate, typeIcon, date/time/week formatting
+  hash-utils.mjs       # MinHash signatures, Jaccard similarity for dedup
+  bash-utils.mjs       # Bash output significance detection: errors, tests, builds, deploys
   # Resource registry
   registry.mjs         # Resource registry DB: schema, CRUD, FTS5, invocation tracking
   registry-retriever.mjs # FTS5 retrieval with synonym expansion and composite scoring

package/README.zh-CN.md

@@ -144,7 +144,7 @@ node install.mjs install
 ### 安装过程
 
 1. **安装依赖** -- `npm install --omit=dev`（编译原生 `better-sqlite3`）
-2. **注册 MCP 服务器** -- `mem` 服务器，包含 7 个工具（search、timeline、get、save、stats、delete、compress）
+2. **注册 MCP 服务器** -- `mem` 服务器，包含 12 个工具（search、timeline、get、save、update、stats、delete、compress、maintain、registry、export、fts_check）
 3. **配置钩子** -- `PostToolUse`、`PreToolUse`、`SessionStart`、`Stop`、`UserPromptSubmit` 生命周期钩子
 4. **创建数据目录** -- `~/.claude-mem-lite/`（隐藏目录），存放数据库、运行时和托管资源文件
 5. **自动迁移** -- 自动检测 `~/.claude-mem/`（原版 claude-mem）或 `~/claude-mem-lite/`（v0.5 前的非隐藏目录），将数据库和运行时文件迁移到 `~/.claude-mem-lite/`，原目录保持不变
@@ -200,9 +200,14 @@ rm -rf ~/claude-mem-lite/   # v0.5 前的非隐藏目录（如未自动迁移）
 | `mem_timeline` | 围绕锚点按时间顺序浏览观察。 |
 | `mem_get` | 获取指定观察 ID 的完整详情（包含重要度和关联 ID）。 |
 | `mem_save` | 手动保存记忆/观察。 |
+| `mem_update` | 原地更新已有观察，保留原始 ID 和引用关系。 |
 | `mem_stats` | 查看统计：计数、类型分布、热门项目、每日活动。 |
 | `mem_delete` | 按 ID 删除观察，支持预览/确认工作流。FTS5 自动清理。 |
 | `mem_compress` | 压缩旧的低价值观察为每周摘要，减少噪声。 |
+| `mem_maintain` | 记忆维护：扫描重复/过期/损坏条目，执行清理/去重/向量重建操作。 |
+| `mem_export` | 导出观察为 JSON 或 JSONL 格式，支持按项目、类型、日期范围过滤。 |
+| `mem_fts_check` | 检查 FTS5 索引完整性或重建索引。搜索结果异常或数据库恢复后使用。 |
+| `mem_registry` | 管理资源注册表：按需搜索技能/代理、列表、统计、导入/移除、重索引。 |
 
 ### 技能命令（在 Claude Code 聊天中使用）
 
@@ -441,7 +446,16 @@ claude-mem-lite/
   hook-semaphore.mjs   # LLM 并发控制：基于文件的信号量
   schema.mjs           # 数据库 schema：表、迁移、FTS5 的单一事实来源
   tool-schemas.mjs     # 共享 Zod schema，用于 MCP 工具校验
-  utils.mjs            # 共享工具：FTS5 查询构建、MinHash 去重、秘密擦除
+  utils.mjs            # 重导出中心：所有工具模块的向后兼容入口
+  nlp.mjs              # FTS5 查询构建：同义词扩展、CJK 二元组、查询清洗
+  scoring-sql.mjs      # BM25 权重常量和类型差异化衰减半衰期
+  stop-words.mjs       # 共享基础停用词集
+  synonyms.mjs         # 统一同义词源：SYNONYM_MAP（双向）+ DISPATCH_SYNONYMS
+  project-utils.mjs    # 共享项目名解析（含进程内缓存）
+  secret-scrub.mjs     # API 密钥、令牌、PEM 证书等凭据模式擦除
+  format-utils.mjs     # 字符串格式化：截断、类型图标、日期/时间格式化
+  hash-utils.mjs       # MinHash 签名、Jaccard 相似度（去重用）
+  bash-utils.mjs       # Bash 输出显著性检测：错误、测试、构建、部署
   # 智能调度
   dispatch.mjs         # 三级调度编排：快速过滤、上下文信号、FTS5、Haiku
   dispatch-inject.mjs  # 注入模板渲染：skill/agent 推荐

package/bash-utils.mjs

@@ -0,0 +1,109 @@
+// claude-mem-lite: Bash command analysis and file path extraction
+// Extracted from utils.mjs for focused responsibility
+
+import { basename } from 'path';
+
+/**
+ * Detect significance signals in a Bash command and its response.
+ * Checks for errors, test runs, builds, git operations, and deployments.
+ * @param {object} input Tool input with command field
+ * @param {string} response Command output text
+ * @returns {{isError: boolean, isTest: boolean, isBuild: boolean, isGit: boolean, isDeploy: boolean, isSignificant: boolean}}
+ */
+export function detectBashSignificance(input, response) {
+  const cmd = (input.command || '').toLowerCase();
+  // Skip error keyword matching when the command is a read/search operation
+  // (grep output naturally contains matched keywords like "error")
+  const isSearchCmd = /\b(grep|rg|ag|ack|cat|head|tail|less|more|find|locate|wc|file|which|type)\b/i.test(cmd);
+  const isError = !isSearchCmd
+    && /\berror\b|\bERR!|fail(ed|ure)?|exception|panic|traceback|errno|enoent|command not found/i.test(response)
+    && response.length > 15;
+  // Match actual test runner invocations, not commands that merely reference "test" as a keyword
+  const isTest = /\b(npm\s+test|npm\s+run\s+test|yarn\s+test|pnpm\s+test|pnpm\s+run\s+test|bun\s+test|go\s+test|cargo\s+test)\b/i.test(cmd)
+    || /\b(jest|pytest|vitest|mocha|cypress|playwright)\b/i.test(cmd);
+  const isBuild = /\b(build|compile|tsc|webpack|vite|rollup|esbuild|make|cargo)\b/i.test(cmd);
+  const isGit = /\bgit\s+(commit|merge|rebase|cherry-pick|push)\b/i.test(cmd);
+  const isDeploy = /\b(deploy|docker|kubectl|terraform)\b/i.test(cmd);
+  return {
+    isError, isTest, isBuild, isGit, isDeploy,
+    isSignificant: isError || isTest || isBuild || isGit || isDeploy,
+  };
+}
+
+const ERROR_STOP_WORDS = new Set([
+  'error', 'failed', 'cannot', 'could', 'with', 'from', 'that', 'this',
+  'have', 'been', 'were', 'does', 'will', 'would', 'should', 'must',
+  'true', 'false', 'null', 'undefined', 'function', 'return', 'const',
+  'node', 'require', 'stack', 'trace',
+]);
+
+/**
+ * Extract discriminative keywords from a failed command and its error output.
+ * Filters out common stop words to produce useful FTS5 search terms.
+ * @param {string} cmd The command that was executed
+ * @param {string} response The error output text
+ * @returns {string[]|null} Array of 1-6 keywords or null if none found
+ */
+export function extractErrorKeywords(cmd, response) {
+  const words = new Set();
+  const cmdParts = cmd.split(/[\s/\\|&;]+/).filter(w => w.length > 2 && !/^-/.test(w));
+  for (const w of cmdParts.slice(0, 3)) {
+    const lw = w.toLowerCase();
+    if (!ERROR_STOP_WORDS.has(lw)) words.add(lw);
+  }
+  const errLines = response.split('\n').filter(l =>
+    /error|fail|exception|cannot|not found|undefined|null/i.test(l)
+  ).slice(0, 3);
+  for (const line of errLines) {
+    const tokens = line.replace(/[^a-zA-Z0-9_.-]/g, ' ').split(/\s+/)
+      .filter(w => w.length > 3 && !/^\d+$/.test(w));
+    for (const t of tokens.slice(0, 5)) {
+      const lt = t.toLowerCase();
+      if (!ERROR_STOP_WORDS.has(lt)) words.add(lt);
+    }
+  }
+  const result = [...words].slice(0, 6);
+  return result.length >= 1 ? result : null;
+}
+
+// ─── File Paths ──────────────────────────────────────────────────────────────
+
+/**
+ * Extract file paths from tool input (file_path, path, filePath, or command args).
+ * Deduplicates and excludes /dev/, /proc/, and /tmp/ paths.
+ * @param {object} input Tool input object
+ * @returns {string[]} Unique array of file paths
+ */
+export function extractFilePaths(input) {
+  const paths = [];
+  if (input.file_path) paths.push(input.file_path);
+  if (input.path) paths.push(input.path);
+  if (input.filePath) paths.push(input.filePath);
+  if (input.command) {
+    // Match absolute paths; extension optional to support Makefile, Dockerfile etc.
+    const match = input.command.match(/(?:^|\s)(\/[\w./-]+\w)/g);
+    if (match) {
+      for (const m of match) {
+        const p = m.trim();
+        if (!p.startsWith('/dev/') && !p.startsWith('/proc/') && !p.startsWith('/tmp/')
+          // Skip single-component paths like /exit, /clear — likely slash commands, not files
+          && (p.indexOf('/', 1) !== -1 || /\.\w+$/.test(p))) {
+          paths.push(p);
+        }
+      }
+    }
+  }
+  return [...new Set(paths)];
+}
+
+// ─── Episode Logic ───────────────────────────────────────────────────────────
+
+/**
+ * Strip test/spec/e2e suffixes from a filename for sibling matching.
+ * Example: auth.test.ts → auth.ts, auth.spec.js → auth.js
+ * @param {string} filePath File path to strip
+ * @returns {string} Basename with test suffix removed
+ */
+export function stripTestSuffix(filePath) {
+  return basename(filePath).replace(/\.(test|spec|e2e)\./i, '.');
+}

package/format-utils.mjs

@@ -0,0 +1,71 @@
+// claude-mem-lite: String formatting and display utilities
+// Extracted from utils.mjs for focused responsibility
+
+/**
+ * Truncate a string to a maximum length, replacing newlines with spaces.
+ * @param {string} str Input string
+ * @param {number} [max=80] Maximum character length
+ * @returns {string} Truncated string with ellipsis if needed
+ */
+export function truncate(str, max = 80) {
+  if (!str) return '';
+  str = str.replace(/\n/g, ' ').trim();
+  return str.length > max ? str.slice(0, max - 1) + '\u2026' : str;
+}
+
+/**
+ * Map observation type to its display emoji icon.
+ * @param {string} type Observation type (decision, bugfix, feature, etc.)
+ * @returns {string} Emoji icon for the type
+ */
+export function typeIcon(type) {
+  const icons = { decision: '\uD83D\uDFE1', bugfix: '\uD83D\uDD34', feature: '\uD83D\uDFE2', refactor: '\uD83D\uDD35', discovery: '\uD83D\uDD0D', change: '\uD83D\uDCDD' };
+  return icons[type] || '\u26AA';
+}
+
+// ─── Date Formatting ─────────────────────────────────────────────────────────
+
+const MONTHS = ['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec'];
+
+/**
+ * Format an ISO date string as "Mon DD HH:MM" for compact display.
+ * @param {string} iso ISO 8601 date string
+ * @returns {string} Formatted date or empty string
+ */
+export function fmtDate(iso) {
+  if (!iso) return '';
+  const d = new Date(iso);
+  const mon = MONTHS[d.getUTCMonth()];
+  const day = d.getUTCDate();
+  const h = String(d.getUTCHours()).padStart(2, '0');
+  const m = String(d.getUTCMinutes()).padStart(2, '0');
+  return `${mon} ${day} ${h}:${m}`;
+}
+
+/**
+ * Format an ISO date string as "HH:MM" for time-only display.
+ * @param {string} iso ISO 8601 date string
+ * @returns {string} Formatted time or empty string
+ */
+export function fmtTime(iso) {
+  if (!iso) return '';
+  const d = new Date(iso);
+  return `${String(d.getUTCHours()).padStart(2, '0')}:${String(d.getUTCMinutes()).padStart(2, '0')}`;
+}
+
+// ─── ISO Week ────────────────────────────────────────────────────────────────
+
+/**
+ * Convert an epoch timestamp to an ISO week key string (e.g. "2026-W06").
+ * @param {number} epochMs Epoch timestamp in milliseconds
+ * @returns {string} ISO week key in format "YYYY-Wnn"
+ */
+export function isoWeekKey(epochMs) {
+  const d = new Date(epochMs);
+  const tmp = new Date(Date.UTC(d.getUTCFullYear(), d.getUTCMonth(), d.getUTCDate()));
+  tmp.setUTCDate(tmp.getUTCDate() + 4 - (tmp.getUTCDay() || 7));
+  const yearStart = new Date(Date.UTC(tmp.getUTCFullYear(), 0, 1));
+  const weekNum = Math.ceil(((tmp - yearStart) / 86400000 + 1) / 7);
+  const isoYear = tmp.getUTCFullYear();
+  return `${isoYear}-W${String(weekNum).padStart(2, '0')}`;
+}

package/hash-utils.mjs

@@ -0,0 +1,77 @@
+// claude-mem-lite: Hashing and similarity utilities
+// Extracted from utils.mjs for focused responsibility
+
+/**
+ * Compute word-level Jaccard similarity between two strings.
+ * @param {string} a First string
+ * @param {string} b Second string
+ * @returns {number} Similarity score between 0 and 1
+ */
+export function jaccardSimilarity(a, b) {
+  if (!a || !b) return 0;
+  // Strip trailing punctuation from tokens to match MinHash normalization
+  // (prevents "server.rs," ≠ "server.rs" dedup failures)
+  const norm = s => s.toLowerCase().split(/\s+/).map(t => t.replace(/[,;:!?]+$/, ''));
+  const setA = new Set(norm(a));
+  const setB = new Set(norm(b));
+  let intersection = 0;
+  for (const w of setA) { if (setB.has(w)) intersection++; }
+  const union = setA.size + setB.size - intersection;
+  return union === 0 ? 0 : intersection / union;
+}
+
+// ─── MinHash Signatures ──────────────────────────────────────────────────
+
+// FNV-1a hash: fast, non-cryptographic, ~10x faster than SHA-256 for MinHash
+function fnv1a(str) {
+  let hash = 0x811c9dc5; // FNV offset basis (32-bit)
+  for (let i = 0; i < str.length; i++) {
+    hash ^= str.charCodeAt(i);
+    hash = Math.imul(hash, 0x01000193); // FNV prime
+    hash >>>= 0; // Keep as uint32
+  }
+  return hash;
+}
+
+/**
+ * Compute a MinHash signature for approximate set similarity.
+ * Returns null for texts with fewer than 3 tokens.
+ * @param {string} text Input text to hash
+ * @param {number} [numHashes=64] Number of hash functions
+ * @returns {string|null} Hex-encoded MinHash signature or null
+ */
+export function computeMinHash(text, numHashes = 64) {
+  if (!text || typeof text !== 'string') return null;
+  const tokens = text.toLowerCase().replace(/[^a-z0-9\s]/g, ' ').split(/\s+/)
+    .filter(t => t.length > 2);
+  // Require at least 3 tokens for meaningful signature (avoids high collision on short texts)
+  if (tokens.length < 3) return null;
+
+  const mins = new Array(numHashes).fill(0xFFFFFFFF);
+  for (const token of tokens) {
+    for (let i = 0; i < numHashes; i++) {
+      const val = fnv1a(`${i}-${token}`);
+      if (val < mins[i]) mins[i] = val;
+    }
+  }
+  return mins.map(v => v.toString(16).padStart(8, '0')).join('');
+}
+
+/**
+ * Estimate Jaccard similarity from two MinHash signatures.
+ * @param {string} sig1 First hex-encoded MinHash signature
+ * @param {string} sig2 Second hex-encoded MinHash signature
+ * @returns {number} Estimated Jaccard similarity between 0 and 1
+ */
+export function estimateJaccardFromMinHash(sig1, sig2) {
+  if (!sig1 || !sig2) return 0;
+  if (sig1.length !== sig2.length) return 0;
+  const numHashes = sig1.length / 8;
+  if (numHashes === 0) return 0;
+  let matches = 0;
+  for (let i = 0; i < numHashes; i++) {
+    const offset = i * 8;
+    if (sig1.slice(offset, offset + 8) === sig2.slice(offset, offset + 8)) matches++;
+  }
+  return matches / numHashes;
+}

package/hook-llm.mjs

@@ -102,48 +102,51 @@ export function saveObservation(obs, projectOverride, sessionIdOverride, externa
 
     const { conceptsText, factsText, textField } = buildFtsTextField(obs);
 
-    const result = db.prepare(`
-      INSERT INTO observations (memory_session_id, project, text, type, title, subtitle, narrative, concepts, facts, files_read, files_modified, importance, minhash_sig, lesson_learned, search_aliases, branch, created_at, created_at_epoch)
-      VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
-    `).run(
-      sessionId, project,
-      textField, obs.type, obs.title, obs.subtitle || '',
-      obs.narrative || '',
-      conceptsText,
-      factsText,
-      JSON.stringify(obs.filesRead || []),
-      JSON.stringify(obs.files || []),
-      obs.importance ?? 1,
-      minhashSig,
-      obs.lessonLearned || null,
-      obs.searchAliases || null,
-      getCurrentBranch(),
-      now.toISOString(), now.getTime()
-    );
-    const savedId = Number(result.lastInsertRowid);
-
-    // Populate observation_files junction table (non-critical)
-    if (savedId && obs.files && obs.files.length > 0) {
-      try {
+    // Atomic: observation INSERT + observation_files + vector in one transaction
+    const savedId = db.transaction(() => {
+      const result = db.prepare(`
+        INSERT INTO observations (memory_session_id, project, text, type, title, subtitle, narrative, concepts, facts, files_read, files_modified, importance, minhash_sig, lesson_learned, search_aliases, branch, created_at, created_at_epoch)
+        VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+      `).run(
+        sessionId, project,
+        textField, obs.type, obs.title, obs.subtitle || '',
+        obs.narrative || '',
+        conceptsText,
+        factsText,
+        JSON.stringify(obs.filesRead || []),
+        JSON.stringify(obs.files || []),
+        obs.importance ?? 1,
+        minhashSig,
+        obs.lessonLearned || null,
+        obs.searchAliases || null,
+        getCurrentBranch(),
+        now.toISOString(), now.getTime()
+      );
+      const id = Number(result.lastInsertRowid);
+
+      // Populate observation_files junction table
+      if (id && obs.files && obs.files.length > 0) {
         const insertFile = db.prepare('INSERT OR IGNORE INTO observation_files (obs_id, filename) VALUES (?, ?)');
         for (const f of obs.files) {
-          if (typeof f === 'string' && f.length > 0) insertFile.run(savedId, f);
+          if (typeof f === 'string' && f.length > 0) insertFile.run(id, f);
         }
-      } catch (e) { debugCatch(e, 'saveObservation-obsFiles'); }
-    }
+      }
 
-    // Write TF-IDF vector (non-critical)
-    try {
-      const vocab = getVocabulary(db);
-      if (vocab) {
-        const vecText = [obs.title || '', obs.narrative || '', (Array.isArray(obs.concepts) ? obs.concepts.join(' ') : '')].filter(Boolean).join(' ');
-        const vec = computeVector(vecText, vocab);
-        if (vec) {
-          db.prepare('INSERT OR REPLACE INTO observation_vectors (observation_id, vector, vocab_version, created_at_epoch) VALUES (?, ?, ?, ?)')
-            .run(savedId, Buffer.from(vec.buffer), vocab.version, Date.now());
+      // Write TF-IDF vector (non-critical — catch inside transaction to avoid rollback)
+      try {
+        const vocab = getVocabulary(db);
+        if (vocab) {
+          const vecText = [obs.title || '', obs.narrative || '', (Array.isArray(obs.concepts) ? obs.concepts.join(' ') : '')].filter(Boolean).join(' ');
+          const vec = computeVector(vecText, vocab);
+          if (vec) {
+            db.prepare('INSERT OR REPLACE INTO observation_vectors (observation_id, vector, vocab_version, created_at_epoch) VALUES (?, ?, ?, ?)')
+              .run(id, Buffer.from(vec.buffer), vocab.version, Date.now());
+          }
         }
-      }
-    } catch (e) { debugCatch(e, 'saveObservation-vector'); }
+      } catch (e) { debugCatch(e, 'saveObservation-vector'); }
+
+      return id;
+    })();
 
     return savedId;
   } finally {

package/hook-update.mjs

@@ -200,6 +200,8 @@ const SOURCE_FILES = [
   'registry.mjs', 'registry-scanner.mjs', 'registry-indexer.mjs',
   'registry-retriever.mjs', 'resource-discovery.mjs',
   'install.mjs', 'install-metadata.mjs', 'mem-cli.mjs', 'tier.mjs', 'tfidf.mjs',
+  'nlp.mjs', 'synonyms.mjs', 'scoring-sql.mjs', 'stop-words.mjs', 'project-utils.mjs',
+  'secret-scrub.mjs', 'format-utils.mjs', 'hash-utils.mjs', 'bash-utils.mjs',
 ];
 const SWITCHABLE_PATHS = [...SOURCE_FILES, 'scripts', 'registry', 'node_modules'];
 

package/install.mjs

@@ -206,7 +206,8 @@ async function install() {
     'registry.mjs', 'registry-scanner.mjs', 'registry-indexer.mjs',
     'registry-retriever.mjs', 'resource-discovery.mjs',
     'install-metadata.mjs', 'mem-cli.mjs', 'tier.mjs', 'tfidf.mjs',
-    'nlp.mjs', 'scoring-sql.mjs', 'stop-words.mjs',
+    'nlp.mjs', 'synonyms.mjs', 'scoring-sql.mjs', 'stop-words.mjs', 'project-utils.mjs',
+    'secret-scrub.mjs', 'format-utils.mjs', 'hash-utils.mjs', 'bash-utils.mjs',
   ];
 
   if (IS_DEV) {

package/mem-cli.mjs

@@ -4,6 +4,7 @@
 
 import { ensureDb, DB_PATH } from './schema.mjs';
 import { sanitizeFtsQuery, relaxFtsQueryToOr, truncate, typeIcon, inferProject, jaccardSimilarity, computeMinHash, scrubSecrets, cjkBigrams, OBS_BM25, TYPE_DECAY_CASE, getCurrentBranch } from './utils.mjs';
+import { resolveProject } from './project-utils.mjs';
 import { TIER_CASE_SQL, tierSqlParams } from './tier.mjs';
 import { getVocabulary, computeVector, vectorSearch, rrfMerge, VECTOR_SCAN_LIMIT } from './tfidf.mjs';
 import { basename, join } from 'path';
@@ -60,20 +61,6 @@ function fmtDateShort(iso) {
   return iso.slice(0, 10); // YYYY-MM-DD
 }
 
-// ─── Project Resolution ──────────────────────────────────────────────────────
-
-function resolveProject(db, name) {
-  if (!name) return name;
-  if (name.includes('--')) return name;
-  const suffixed = db.prepare(
-    'SELECT project FROM observations WHERE project LIKE ? GROUP BY project ORDER BY COUNT(*) DESC LIMIT 1'
-  ).get(`%--${name}`);
-  if (suffixed) return suffixed.project;
-  const inferred = inferProject();
-  if (inferred.endsWith(`--${name}`)) return inferred;
-  return name;
-}
-
 // ─── Commands ────────────────────────────────────────────────────────────────
 
 function cmdSearch(db, args) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-mem-lite",
-  "version": "2.18.0",
+  "version": "2.19.0",
   "description": "Lightweight persistent memory system for Claude Code",
   "type": "module",
   "engines": {
@@ -46,6 +46,12 @@
     "skip-tools.mjs",
     "tool-schemas.mjs",
     "utils.mjs",
+    "project-utils.mjs",
+    "secret-scrub.mjs",
+    "format-utils.mjs",
+    "hash-utils.mjs",
+    "bash-utils.mjs",
+    "synonyms.mjs",
     "schema.mjs",
     "skill.md",
     "commands/mem.md",

package/project-utils.mjs

@@ -0,0 +1,37 @@
+// claude-mem-lite shared project resolution
+// Extracted from server.mjs and mem-cli.mjs to eliminate duplication
+
+import { inferProject } from './utils.mjs';
+
+const _cache = new Map();
+
+/**
+ * Resolve short project name to canonical "parent--base" form.
+ * Uses DB suffix match with in-process cache.
+ * @param {import('better-sqlite3').Database} db Database instance
+ * @param {string|null|undefined} name Project name to resolve
+ * @returns {string|null|undefined} Canonical project name
+ */
+export function resolveProject(db, name) {
+  if (!name) return name;
+  if (_cache.has(name)) return _cache.get(name);
+  // Already a canonical name (contains "--")? Use as-is.
+  if (name.includes('--')) { _cache.set(name, name); return name; }
+
+  // Short name: prefer the canonical "parent--name" form (from inferProject())
+  // which typically has far more data than manually-saved short names.
+  const suffixed = db.prepare(
+    'SELECT project FROM observations WHERE project LIKE ? GROUP BY project ORDER BY COUNT(*) DESC LIMIT 1'
+  ).get(`%--${name}`);
+  if (suffixed) { _cache.set(name, suffixed.project); return suffixed.project; }
+
+  // Fallback: synthesize canonical form from current directory
+  const inferred = inferProject();
+  if (inferred.endsWith(`--${name}`)) { _cache.set(name, inferred); return inferred; }
+
+  _cache.set(name, name);
+  return name;
+}
+
+/** Reset cache (for tests). */
+export function _resetProjectCache() { _cache.clear(); }