Skip to content
Back to search

aivectormemory

aivectormemory 是一款基于 Model Context Protocol (MCP) 开发的OpenClaw、OpenCode、ClaudeCodeAI记忆管理工具。它专门为 Claude、OpenCode、Cursor 和 主流IDE 编程工具设计,通过向量数据库技术解决 AI 在不同对话会话中「健忘」的问题。aivectormemory: A lightweight MCP Server enabling persistent, cross-session memory for AI-powered IDEs via vector search.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/edlineas--aivectormemory/badge.svg)](/listings/edlineas--aivectormemory)

HTML

<a href="/listings/edlineas--aivectormemory"><img src="/api/listings/edlineas--aivectormemory/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Option 1: pip install (Recommended)

# Install
pip install aivectormemory

# Upgrade to latest version
pip install --upgrade aivectormemory

# Navigate to your project directory, one-click IDE setup
cd /path/to/your/project
run install

`run install` interactively guides you to select your IDE, auto-generating MCP config, Steering rules, and Hooks — no manual setup needed.

> **macOS users note**: > - If you get `externally-managed-environment` error, add `--break-system-packages` > - If you get `enable_load_extension` error, your Python doesn't support SQLite extension loading (macOS built-in Python and python.org installers don't support it). Use Homebrew Python instead: > ```bash > brew install python > /opt/homebrew/bin/python3 -m pip install aivectormemory > ```

Option 2: uvx (zero install)

No `pip install` needed, run directly:

cd /path/to/your/project
uvx aivectormemory install

> Requires [uv](https://docs.astral.sh/uv/getting-started/installation/) to be installed. `uvx` auto-downloads and runs the package — no manual installation needed.

Option 3: Manual configuration

{
  "mcpServers": {
    "aivectormemory": {
      "command": "run",
      "args": ["--project-dir", "/path/to/your/project"]
    }
  }
}

<details> <summary>📍 IDE Configuration File Locations</summary>

| IDE | Config Path | |-----|------------| | Kiro | `.kiro/settings/mcp.json` | | Cursor | `.cursor/mcp.json` | | Claude Code | `.mcp.json` | | Windsurf | `.windsurf/mcp.json` | | VSCode | `.vscode/mcp.json` | | Trae | `.trae/mcp.json` | | OpenCode | `opencode.json` | | Codex | `.codex/config.toml` |

</details>

For Codex, use project-scoped TOML instead of JSON:

[mcp_servers.aivectormemory]
command = "run"
args = ["--project-dir", "/path/to/your/project"]

> Codex only loads project-scoped `.codex/config.toml` after the repository is marked as a trusted project.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 632 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. 🌐 [简体中文](docs/README.zh-CN.md) | [繁體中文](docs/README.zh-TW.md) | English | [Español](docs/README.es.md) | [Deutsch](docs/README.de.md) | [Français](docs/README.fr.md) | [日本語](docs/README.ja.md)
  2. <p align="center">
  3. <img src="docs/logo.png" alt="AIVectorMemory Logo" width="200">
  4. </p>
  5. <p align="center">
  6. <img src="docs/image.png" alt="AI Vector Memory Architecture" width="100%">
  7. </p>
  8. <h1 align="center">AIVectorMemory</h1>
  9. <p align="center">
  10. <strong>Give your AI coding assistant a memory — Cross-session persistent memory MCP Server</strong>
  11. </p>
  12. <p align="center">
  13. <a href="https://pypi.org/project/aivectormemory/"><img src="https://img.shields.io/pypi/v/aivectormemory?color=blue&label=PyPI" alt="PyPI"></a>
  14. <a href="https://pypi.org/project/aivectormemory/"><img src="https://img.shields.io/pypi/pyversions/aivectormemory" alt="Python"></a>
  15. <a href="https://github.com/Edlineas/aivectormemory/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-Apache_2.0-green" alt="License"></a>
  16. <a href="https://modelcontextprotocol.io"><img src="https://img.shields.io/badge/MCP-compatible-purple" alt="MCP"></a>
  17. </p>
  18. ---
  19. > **Still using CLAUDE.md / MEMORY.md as memory?** This Markdown-file memory approach has fatal flaws: the file keeps growing, injecting everything into every session and burning massive tokens; content only supports keyword matching — search "database timeout" and you won't find "MySQL connection pool pitfall"; sharing one file across projects causes cross-contamination; there's no task tracking, so dev progress lives entirely in your head; not to mention the 200-line truncation, manual maintenance, and inability to deduplicate or merge.
  20. >
  21. > **AIVectorMemory is a fundamentally different approach.** Local vector database storage with semantic search for precise recall (matches even when wording differs), on-demand retrieval that loads only relevant memories (token usage drops 50%+), automatic multi-project isolation with zero interference, and built-in issue tracking + task management that lets AI fully automate your dev workflow. All data is permanently stored on your machine — zero cloud dependency, never lost when switching sessions or IDEs.
  22. ## ✨ Core Features
  23. | Feature | Description |
  24. |---------|-------------|
  25. | 🧠 **Cross-Session Memory** | Your AI finally remembers your project — pitfalls, decisions, conventions all persist across sessions |
  26. | 🔍 **Hybrid Smart Search** | FTS5 full-text + vector semantic dual-path search, RRF fusion ranking + composite scoring (recency × frequency × importance), far more precise than pure vector search |
  27. | 🐛 **Issue Tracking** | Built-in Issue Tracker — discover → investigate → fix → archive, full lifecycle. AI manages bugs automatically |
  28. | 📋 **Task Management** | Spec → task breakdown → nested subtasks → status sync → linked archival. AI drives the complete dev workflow |
  29. | 🚦 **Session State** | Blocking management + breakpoint resume + progress tracking, seamless handoff across sessions and context compaction |
  30. | 🪝 **Hooks + Steering** | Auto-inject workflow rules + behavior guard hooks, consistent AI behavior guaranteed — no need to repeat instructions |
  31. | 🧬 **Memory Evolution** | Contradiction detection auto-supersedes stale knowledge + short-term → long-term auto-promotion + 90-day auto-archive, self-evolving memory |
  32. | 📊 **Desktop App + Web Dashboard** | Native desktop app (macOS/Windows/Linux) + Web dashboard, 3D vector network reveals knowledge connections at a glance |
  33. | 💰 **Save 50%+ Tokens** | Stop copy-pasting project context every conversation. Semantic retrieval on demand, no more bulk injection |
  34. | 🏠 **Fully Local** | Zero cloud dependency. ONNX local inference, no API Key, data never leaves your machine |
  35. | 🔌 **11 IDEs Covered** | Cursor / Kiro / Claude Code / Windsurf / VSCode / Copilot / OpenCode / Trae / Codex / Antigravity / OpenClaw — one-click install & uninstall |
  36. | 📁 **Multi-Project Isolation** | One DB for all projects, auto-isolated with zero interference, seamless project switching |
  37. | 🔄 **Smart Dedup** | Similarity > 0.95 auto-merges updates, keeping your memory store clean — never gets messy over time |
  38. | 🌐 **7 Languages** | 简体中文 / 繁體中文 / English / Español / Deutsch / Français / 日本語, full-stack i18n for dashboard + Steering rules |
  39. <p align="center">
  40. QQ群:1085682431 &nbsp;|&nbsp; 微信:changhuibiz<br>
  41. 共同参与项目开发加QQ群或微信交流
  42. </p>
  43. <p align="center">
  44. <img src="docs/003.png" alt="Login" width="100%">
  45. <br>
  46. <em>Login</em>
  47. </p>
  48. <p align="center">
  49. <img src="docs/001.png" alt="Project Selection" width="100%">
  50. <br>
  51. <em>Project Selection</em>
  52. </p>
  53. <p align="center">
  54. <img src="docs/002.png" alt="Overview & Vector Network" width="100%">
  55. <br>
  56. <em>Overview & Vector Network</em>
  57. </p>
  58. ## 🏗️ Architecture
  59. ```
  60. ┌─────────────────────────────────────────────────┐
  61. │ AI IDE │
  62. │ OpenCode / Codex / Claude Code / Cursor / ... │
  63. └──────────────────────┬──────────────────────────┘
  64. │ MCP Protocol (stdio)
  65. ┌──────────────────────▼──────────────────────────┐
  66. │ AIVectorMemory Server │
  67. │ │
  68. │ ┌──────────┐ ┌──────────┐ ┌──────────────────┐ │
  69. │ │ remember │ │ recall │ │ auto_save │ │
  70. │ │ forget │ │ task │ │ status/track │ │
  71. │ └────┬─────┘ └────┬─────┘ └───────┬──────────┘ │
  72. │ │ │ │ │
  73. │ ┌────▼────────────▼───────────────▼──────────┐ │
  74. │ │ Embedding Engine (ONNX) │ │
  75. │ │ intfloat/multilingual-e5-small │ │
  76. │ └────────────────────┬───────────────────────┘ │
  77. │ │ │
  78. │ ┌────────────────────▼───────────────────────┐ │
  79. │ │ SQLite + sqlite-vec (Vector Index) │ │
  80. │ │ ~/.aivectormemory/memory.db │ │
  81. │ └────────────────────────────────────────────┘ │
  82. └──────────────────────────────────────────────────┘
  83. ```
  84. ## 🚀 Quick Start
  85. ### Option 1: pip install (Recommended)
  86. ```bash
  87. # Install
  88. pip install aivectormemory
  89. # Upgrade to latest version
  90. pip install --upgrade aivectormemory
  91. # Navigate to your project directory, one-click IDE setup
  92. cd /path/to/your/project
  93. run install
  94. ```
  95. `run install` interactively guides you to select your IDE, auto-generating MCP config, Steering rules, and Hooks — no manual setup needed.
  96. > **macOS users note**:
  97. > - If you get `externally-managed-environment` error, add `--break-system-packages`
  98. > - If you get `enable_load_extension` error, your Python doesn't support SQLite extension loading (macOS built-in Python and python.org installers don't support it). Use Homebrew Python instead:
  99. > ```bash
  100. > brew install python
  101. > /opt/homebrew/bin/python3 -m pip install aivectormemory
  102. > ```
  103. ### Option 2: uvx (zero install)
  104. No `pip install` needed, run directly:
  105. ```bash
  106. cd /path/to/your/project
  107. uvx aivectormemory install
  108. ```
  109. > Requires [uv](https://docs.astral.sh/uv/getting-started/installation/) to be installed. `uvx` auto-downloads and runs the package — no manual installation needed.
  110. ### Option 3: Manual configuration
  111. ```json
  112. {
  113. "mcpServers": {
  114. "aivectormemory": {
  115. "command": "run",
  116. "args": ["--project-dir", "/path/to/your/project"]
  117. }
  118. }
  119. }
  120. ```
  121. <details>
  122. <summary>📍 IDE Configuration File Locations</summary>
  123. | IDE | Config Path |
  124. |-----|------------|
  125. | Kiro | `.kiro/settings/mcp.json` |
  126. | Cursor | `.cursor/mcp.json` |
  127. | Claude Code | `.mcp.json` |
  128. | Windsurf | `.windsurf/mcp.json` |
  129. | VSCode | `.vscode/mcp.json` |
  130. | Trae | `.trae/mcp.json` |
  131. | OpenCode | `opencode.json` |
  132. | Codex | `.codex/config.toml` |
  133. </details>
  134. For Codex, use project-scoped TOML instead of JSON:
  135. ```toml
  136. [mcp_servers.aivectormemory]
  137. command = "run"
  138. args = ["--project-dir", "/path/to/your/project"]
  139. ```
  140. > Codex only loads project-scoped `.codex/config.toml` after the repository is marked as a trusted project.
  141. ## 🛠️ 9 MCP Tools
  142. ### `remember` — Store a memory
  143. ```
  144. content (string, required) Memory content in Markdown format
  145. tags (string[], required) Tags, e.g. ["pitfall", "python"]
  146. scope (string) "project" (default) / "user" (cross-project)
  147. ```
  148. Similarity > 0.95 auto-updates existing memory, no duplicates.
  149. ### `recall` — Semantic search
  150. ```
  151. query (string) Semantic search keywords
  152. tags (string[]) Exact tag filter
  153. scope (string) "project" / "user" / "all"
  154. top_k (integer) Number of results, default 5
  155. ```
  156. Vector similarity matching — finds related memories even with different wording.
  157. ### `forget` — Delete memories
  158. ```
  159. memory_id (string) Single ID
  160. memory_ids (string[]) Batch IDs
  161. ```
  162. ### `status` — Session state
  163. ```
  164. state (object, optional) Omit to read, pass to update
  165. is_blocked, block_reason, current_task,
  166. next_step, progress[], recent_changes[], pending[]
  167. ```
  168. Maintains work progress across sessions, auto-restores context in new sessions.
  169. ### `track` — Issue tracking
  170. ```
  171. action (string) "create" / "update" / "archive" / "list"
  172. title (string) Issue title
  173. issue_id (integer) Issue ID
  174. status (string) "pending" / "in_progress" / "completed"
  175. content (string) Investigation content
  176. ```
  177. ### `task` — Task management
  178. ```
  179. action (string, required) "batch_create" / "update" / "list" / "delete" / "archive"
  180. feature_id (string) Linked feature identifier (required for list)
  181. tasks (array) Task list (batch_create, supports subtasks)
  182. task_id (integer) Task ID (update)
  183. status (string) "pending" / "in_progress" / "completed" / "skipped"
  184. ```
  185. Links to spec docs via feature_id. Update auto-syncs tasks.md checkboxes and linked issue status.
  186. ### `readme` — README generation
  187. ```
  188. action (string) "generate" (default) / "diff" (compare differences)
  189. lang (string) Language: en / zh-TW / ja / de / fr / es
  190. sections (string[]) Specify sections: header / tools / deps
  191. ```
  192. Auto-generates README content from TOOL_DEFINITIONS / pyproject.toml, multi-language support.
  193. ### `auto_save` — Auto save preferences
  194. ```
  195. preferences (string[]) User-expressed technical preferences (fixed scope=user, cross-project)
  196. extra_tags (string[]) Additional tags
  197. ```
  198. Auto-extracts and stores user preferences at end of each conversation, smart dedup.
  199. ### `graph` — Code knowledge graph
  200. ```
  201. action (string, required) "query" / "trace" / "batch" / "add_node" / "add_edge" / "remove" / "refresh"
  202. name (string) Entity name (add_node/query)
  203. entity_type (string) Entity type: function/class/module/api/table/config (add_node/query)
  204. file_path (string) File path, auto-converts to relative (add_node/query/refresh)
  205. source (string) Source node name or ID (add_edge)
  206. target (string) Target node name or ID (add_edge)
  207. relation (string) Relation type: calls/imports/inherits/uses/depends_on/contains (add_edge/trace)
  208. start (string) Start node name or ID (trace)
  209. direction (string) Traversal direction: "up" / "down" / "both" (trace)
  210. max_depth (integer) Max traversal depth, default 3 (trace)
  211. ```
  212. Manages function call chains, data flows, and dependency relationships. Trace upstream/downstream impact before code changes.
  213. ## 📊 Web Dashboard
  214. ```bash
  215. run web --port 9080
  216. run web --port 9080 --quiet # Suppress request logs
  217. run web --port 9080 --quiet --daemon # Run in background (macOS/Linux)
  218. ```
  219. Visit `http://localhost:9080` in your browser. Default username `admin`, password `admin123` (can be changed in settings after first login).
  220. - Multi-project switching, memory browse/search/edit/delete/export/import
  221. - Semantic search (vector similarity matching)
  222. - One-click project data deletion
  223. - Session status, issue tracking
  224. - Tag management (rename, merge, batch delete)
  225. - Token authentication protection
  226. - 3D vector memory network visualization
  227. - 🌐 Multi-language support (简体中文 / 繁體中文 / English / Español / Deutsch / Français / 日本語)
  228. <p align="center">
  229. <img src="docs/20260306234753_6_1635.jpg" alt="WeChat Group" width="280">
  230. &nbsp;&nbsp;&nbsp;&nbsp;
  231. <img src="docs/8_1635.jpg" alt="QQ Group: 1085682431" width="280">
  232. <br>
  233. <em>Scan to join WeChat group &nbsp;|&nbsp; Scan to join QQ group</em>
  234. </p>
  235. ## ⚡ Pairing with Steering Rules
  236. AIVectorMemory is the storage layer. Use Steering rules to tell AI **when and how** to call these tools.
  237. Running `run install` auto-generates Steering rules and Hooks config — no manual setup needed.
  238. | IDE | Steering Location | Hooks |
  239. |-----|------------------|-------|
  240. | Kiro | `.kiro/steering/aivectormemory.md` | `.kiro/hooks/*.hook` |
  241. | Cursor | `.cursor/rules/aivectormemory.md` | `.cursor/hooks.json` |
  242. | Claude Code | `CLAUDE.md` (appended) | `.claude/settings.json` |
  243. | Windsurf | `.windsurf/rules/aivectormemory.md` | `.windsurf/hooks.json` |
  244. | VSCode | `.github/copilot-instructions.md` (appended) | `.claude/settings.json` |
  245. | Trae | `.trae/rules/aivectormemory.md` | — |
  246. | OpenCode | `AGENTS.md` (appended) | `.opencode/plugins/*.js` |
  247. | Codex | `AGENTS.md` (appended) | — |
  248. <details>
  249. <summary>📋 Steering Rules Example (auto-generated)</summary>
  250. ```markdown
  251. # AIVectorMemory - Workflow Rules
  252. ## 1. New Session Startup (execute in order)
  253. 1. `recall` (tags: ["project-knowledge"], scope: "project", top_k: 100) load project knowledge
  254. 2. `recall` (tags: ["preference"], scope: "user", top_k: 20) load user preferences
  255. 3. `status` (no state param) read session state
  256. 4. Blocked → report and wait; Not blocked → enter processing flow
  257. ## 2. Message Processing Flow
  258. - Step A: `status` read state, wait if blocked
  259. - Step B: Classify message type (chat/correction/preference/code issue)
  260. - Step C: `track create` record issue
  261. - Step D: Investigate (`recall` pitfalls + read code + find root cause)
  262. - Step E: Present plan to user, set blocked awaiting confirmation
  263. - Step F: Modify code (`recall` pitfalls before changes)
  264. - Step G: Run tests to verify
  265. - Step H: Set blocked awaiting user verification
  266. - Step I: User confirms → `track archive` + clear block
  267. ## 3. Blocking Rules
  268. Must `status({ is_blocked: true })` when proposing plans or awaiting verification.
  269. Only clear after explicit user confirmation. Never self-clear.
  270. ## 4-9. Issue Tracking / Code Checks / Spec Task Mgmt / Memory Quality / Tool Reference / Dev Standards
  271. (Full rules auto-generated by `run install`)
  272. ```
  273. </details>
  274. <details>
  275. <summary>🔗 Hooks Config Example (Kiro only, auto-generated)</summary>
  276. Auto-save on session end removed. Dev workflow check (`.kiro/hooks/dev-workflow-check.kiro.hook`):
  277. ```json
  278. {
  279. "enabled": true,
  280. "name": "Dev Workflow Check",
  281. "version": "1",
  282. "when": { "type": "promptSubmit" },
  283. "then": {
  284. "type": "askAgent",
  285. "prompt": "Core principles: verify before acting, no blind testing, only mark done after tests pass"
  286. }
  287. }
  288. ```
  289. </details>
  290. ## 🇨🇳 Users in China
  291. The embedding model (~200MB) is auto-downloaded on first run. If slow:
  292. ```bash
  293. export HF_ENDPOINT=https://hf-mirror.com
  294. ```
  295. Or add env to MCP config:
  296. ```json
  297. {
  298. "env": { "HF_ENDPOINT": "https://hf-mirror.com" }
  299. }
  300. ```
  301. ## 📦 Tech Stack
  302. | Component | Technology |
  303. |-----------|-----------|
  304. | Runtime | Python >= 3.10 |
  305. | Vector DB | SQLite + sqlite-vec |
  306. | Embedding | ONNX Runtime + intfloat/multilingual-e5-small |
  307. | Tokenizer | HuggingFace Tokenizers |
  308. | Protocol | Model Context Protocol (MCP) |
  309. | Web | Native HTTPServer + Vanilla JS |
  310. ## 📋 Changelog
  311. ### v2.4.5
  312. **Patch: Hard Constraints Against Opus 4.7 Default Tendencies**
  313. - 🚫 §1 added **No Clarification-Style Follow-ups**: forbid re-asking "phased or one-shot / full or partial / should I do X / do A first or B first" for imperative commands; under ambiguity, execute the most complete scope
  314. - 🚫 §1 added **No Defensive Reporting**: forbid wording like "kept per instruction / marked pending / non-critical path / unnecessary sub-tests / for later iteration" as excuse for unexecuted items
  315. - 📋 §1 added **Report Format**: forbid Phase A/B/C/D list + "Final Status" + "Not Done (Kept Per Instruction)" three-section format; when user says "do all", no "Not Done/Kept" section allowed
  316. - 🎯 Root cause: counteract Opus 4.7's default "defensive reporting", "clarification follow-ups", and "structured checklist" tendencies compared to 4.6
  317. - 🔄 7-language rule files (STEERING_CONTENT + DEV_WORKFLOW_PROMPT) fully synced with CLAUDE.md v2.4.5 updates
  318. ### v2.4.4
  319. **Patch: Full A-I Message Processing Flow Alignment**
  320. - 🧩 CLAUDE.md §4 message processing B routes fully expanded: all 4 branches (casual/correction/preference/other) unified to terminate at I(user confirm & archive), eliminating incomplete "stop at F" flows
  321. - ⚙️ inject-workflow-rules.sh message type judgment section fully aligned with §4 B: 4 routes with consistent granularity
  322. - 🔧 Fixed 3 conflicts: route granularity inconsistency (2 vs 4 routes) / B and E responsibility mixing ("solution+block" misplaced) / G/H/I flow missing
  323. - 📝 Unified violation clause: "Proceeding to C/D/E/F steps without outputting judgment result = violation"
  324. - 🔄 7-language rule files (STEERING_CONTENT + DEV_WORKFLOW_PROMPT) fully synced with CLAUDE.md v2.4.4 updates
  325. ### v2.4.3
  326. **Patch: Rule Enforcement & Graph Visualization**
  327. - 🧠 §4.B: Two-step mandatory structure (understand message → determine type), skipping = violation
  328. - 📋 §8: Review + block embedded in each Spec step, skipping = violation
  329. - 🧬 DEV_WORKFLOW_PROMPT: Added graph trace/batch/add_node rules for investigation and code modification
  330. - 📊 Graph dashboard: Dynamic force layout scaling, edge labels hidden by default (hover to show), node label collision detection
  331. - 🔧 Removed redundant "frequent violation reminders" section from DEV_WORKFLOW_PROMPT
  332. - 📝 Unified §1 IDENTITY & TONE across all 7 languages (no translation)
  333. ### v2.4.1
  334. **Patch: i18n Rules Sync**
  335. - 🔄 Synced all 7 language rule files (STEERING_CONTENT + DEV_WORKFLOW_PROMPT) with CLAUDE.md v2.4.0 updates
  336. - 🧬 Added `graph` tool references to all steering rules (trace/batch/add_node/add_edge/remove)
  337. - ✏️ Updated authority role from "Lead Architect" to "Project Owner" across all languages
  338. - 📝 Added 3 new violation rules and 2 new forbidden items to all languages
  339. ### v2.4.0
  340. **New: Code Knowledge Graph**
  341. - 🧬 `graph` tool — manage function call chains, data flows, and dependency relationships as a structured knowledge graph
  342. - 🔍 `trace` action — traverse upstream/downstream call chains from any entity, assess impact scope before code changes
  343. - 📊 Web dashboard graph visualization page — browse nodes, edges, and call relationships in the knowledge graph
  344. - 🗃️ DB migration v15 — new `graph_nodes` and `graph_edges` tables for graph storage
  345. - 🌐 All 7 language README files updated in sync
  346. ### v2.3.1
  347. **Enhancement: Rule System Overhaul + OpenClaw Support**
  348. - 🧠 Fixed 5 missing memory system calls in AI rules: recall pitfalls before investigation (Step D), before dangerous ops (§7), before Spec writing (§8), before subtask execution (§8), and remember pitfalls after fix (Step I)
  349. - 🦞 Added OpenClaw IDE support — now 11 IDEs total (MCP config merges into ~/.openclaw/openclaw.json, steering appends to AGENTS.md)
  350. - 🎭 Playwright self-test rules strengthened — added ToolSearch deferred tools loading requirement, banned `open` command workaround
  351. - 🔧 Merged v2.2.0–v2.2.6 features: hooks system (bash_guard + stop_guard + check_track), scoring engine improvements, recall optimizations, web dashboard bulk delete, desktop memory delete modal
  352. - ⚠️ DEV_WORKFLOW_PROMPT: added 2 new violation reminders (recall before code change, remember after fix)
  353. - 🌐 All 7 language rule files updated in sync
  354. ### v2.1.1
  355. **Enhancement: AI Rule System Upgrade**
  356. - 📋 CLAUDE.md completion: added Identity & Tone (§1), 7 Core Principles (§3), message type judgment examples, expanded IDE safety and self-test sections
  357. - ⚠️ Hook added Common Violations Reminder: ❌ negative examples reinforcing the 4 most frequently missed rules (self-test, recall, track create, IDE safety)
  358. - 🌐 All 7 language rule files updated in sync (zh-CN/zh-TW/en/ja/es/de/fr)
  359. - 🔢 CLAUDE.md sections renumbered to §1–§11, cross-references updated accordingly
  360. ### v2.1.0
  361. **New: Smart Memory Engine + Uninstall**
  362. - 🧠 FTS5 full-text search with Chinese tokenization (jieba) — keyword search now actually works for CJK content
  363. - 🔀 Hybrid retrieval: vector + FTS5 dual-path with RRF (Reciprocal Rank Fusion) merging
  364. - 📊 Composite scoring: results ranked by similarity × 0.5 + recency × 0.3 + frequency × 0.2, weighted by importance
  365. - ⚡ Conflict detection: similar memories (0.85–0.95) auto-superseded, old facts fade automatically
  366. - 📦 Memory tiers: frequently accessed memories auto-promote to long_term and get searched first
  367. - 🗑️ Auto-archive: stale short_term memories (90 days inactive + low importance) cleaned up automatically
  368. - 🔗 Relation expansion: tag overlap ≥ 2 builds related links, 1-hop expansion surfaces connected memories
  369. - 📝 Auto-summary: long memories (>500 chars) get summaries, brief mode returns summaries to save tokens
  370. - 🧹 Code cleanup: removed 15 dead code items, refactored 7 duplicate patterns into shared utilities
  371. - ❌ `run uninstall` — cleanly removes all IDE configurations (MCP, steering, hooks, permissions) while preserving memory data
  372. ### v2.0.9
  373. **Enhancement: Security & Rule Optimization**
  374. - 🔒 Fixed SQL injection, command injection, and path traversal vulnerabilities
  375. - 🛡️ Added transaction protection for data integrity (archive, insert, update operations)
  376. - 🧠 Unified similarity formula across all search paths
  377. - 📏 Compressed AI workflow rules by 38% (219→136 lines) with zero process removal
  378. - 🧹 v12 migration cleans up legacy garbage memories automatically
  379. - 🌐 All 7 languages synchronized
  380. ### v2.0.8
  381. **New: Playwright Browser Testing Built-in**
  382. - 🎭 `run install` now automatically configures Playwright browser testing — AI can open a real browser to verify frontend changes instead of guessing
  383. - 🎭 Uses a dedicated test browser (Chrome for Testing) that won't interfere with your personal browser tabs
  384. - 🔑 Simplified permission setup — no more manual permission popups for common tools
  385. - 📏 Updated AI rules across all 7 languages to enforce proper browser testing behavior
  386. ### v2.0.7
  387. **Enhancement: More IDE Support**
  388. - 🖥️ Added support for Antigravity and GitHub Copilot IDEs
  389. - 🔑 `run install` now auto-configures tool permissions, reducing manual setup
  390. - 📏 Streamlined AI self-testing rules
  391. ### v2.0.6
  392. **Enhancement: Faster Startup**
  393. - ⚡ Optimized memory loading on session start — loads faster with less context usage
  394. - 🔑 Auto-configures Claude Code permissions during installation
  395. - 🌐 All 7 languages synchronized
  396. ### v2.0.5
  397. **Enhancement: Simpler Rules**
  398. - 📏 AI workflow rules restructured for clarity and reduced token usage
  399. - 💾 AI now automatically saves your preferences at the end of each session
  400. - 🌐 All 7 languages synchronized
  401. ### v2.0.4
  402. **Fix: Tool Reliability**
  403. - 🔧 Comprehensive audit and fix of all MCP tool parameters — improved reliability across all IDEs
  404. ### v2.0.3
  405. **Enhancement: Better Search & Safety**
  406. - 🔍 Memory search now combines semantic and keyword matching for more accurate recall
  407. - 🛡️ Added cross-project protection — AI won't accidentally modify files in other projects
  408. ### v2.0.2
  409. **Enhancement: Rule Generalization & Desktop Version Fix**
  410. - 📏 Added "recall before asking user" rule — AI must query memory system before asking user for project information (server address, passwords, deploy config, etc.)
  411. - 📏 Generalized pre-operation check rule — removed specific examples to apply to all operation scenarios
  412. - 🖥️ Fixed desktop app settings page showing hardcoded version "1.0.0" instead of actual app version
  413. - 🌐 All 7 language i18n steering rules and workflow prompts synchronized
  414. ### v2.0.1
  415. **Fix: Hook Cross-Project Compatibility**
  416. - 🔧 `check_track.sh` now derives project path from script location instead of `$(pwd)`, fixing track detection failure when Claude Code runs hooks from non-root working directory
  417. - 🔧 `compact-recovery.sh` now uses relative path derivation instead of hardcoded absolute paths, ensuring correct behavior when installed to any project
  418. - 🔧 Removed redundant CLAUDE.md re-injection from compact-recovery (already auto-loaded by Claude Code)
  419. - 🔧 `install.py` template synchronized with all hook fixes
  420. - 🌐 All 7 language i18n compact-recovery hints updated
  421. ### v2.0
  422. **Performance: ONNX INT8 Quantization**
  423. - ⚡ Embedding model auto-quantized from FP32 to INT8 on first load, model file from 448MB down to 113MB
  424. - ⚡ MCP Server memory usage reduced from ~1.6GB to ~768MB (50%+ reduction)
  425. - ⚡ Quantization is transparent to users — automatic on first use, cached for subsequent loads, falls back to FP32 on failure
  426. **New: Remember Password**
  427. - 🔐 Login page on both desktop and web dashboard now has a "Remember password" checkbox
  428. - 🔐 When checked, credentials are saved to localStorage and auto-filled on next login; when unchecked, saved credentials are cleared
  429. - 🔐 Checkbox is hidden in registration mode
  430. **Enhancement: Steering Rules**
  431. - 📝 IDENTITY & TONE section strengthened with more specific constraints (no pleasantries, no translating user messages, etc.)
  432. - 📝 Self-testing requirements now distinguish between backend-only, MCP Server, and frontend-visible changes (Playwright required for frontend)
  433. - 📝 Development rules now mandate self-testing after completing development
  434. - 📝 All 7 language versions synchronized
  435. ### v1.0.11
  436. - 🐛 Desktop app version comparison switched to semantic versioning, fixing false upgrade prompts when local version is higher
  437. - 🐛 Health check page field names aligned with backend, fixing consistency status always showing Mismatch
  438. - 🔧 check_track.sh hook adds Python fallback, resolving silent hook failure when system sqlite3 is unavailable (#4)
  439. ### v1.0.10
  440. - 🖥️ Desktop app one-click install + upgrade detection
  441. - 🖥️ Auto-detect Python and aivectormemory installation status on startup
  442. - 🖥️ Show one-click install button when not installed, check PyPI and desktop new versions when installed
  443. - 🐛 Installation detection switched to importlib.metadata.version() for accurate package version
  444. ### v1.0.8
  445. - 🔧 Fix PyPI package size anomaly (sdist from 32MB down to 230KB), excluded accidentally packaged dev files
  446. ### v1.0.6
  447. **New: Native Desktop App**
  448. - 🖥️ Native desktop client supporting macOS (ARM64), Windows (x64), Linux (x64)
  449. - 🖥️ Desktop app shares the same database as Web dashboard, fully feature-equivalent
  450. - 🖥️ Dark/light theme switching, Glass frosted visual style
  451. - 🖥️ Login auth, project selection, stats overview, memory management, issue tracking, task management, tag management, settings, data maintenance — full feature coverage
  452. - 📦 Auto-published installers via GitHub Releases, download and use
  453. **New: CI/CD Auto Build**
  454. - 🔄 GitHub Actions auto-builds desktop installers for all 3 platforms
  455. - 🔄 Push a tag to trigger the full compile, package, and release pipeline
  456. **Fixes**
  457. - 🐛 Windows platform compatibility fixes
  458. - 🐛 sqlite-vec extension download URL fix
  459. ### v1.0.5
  460. **Optimization: Token Usage Reduction**
  461. - ⚡ Steering rules changed from per-message dynamic injection to static loading, reducing repeated token consumption
  462. - ⚡ Greatest impact for Claude Code users — ~2K fewer tokens per message
  463. ### v1.0.4
  464. **New: Full-Stack i18n (7 Languages)**
  465. - 🌐 Web dashboard + desktop UI fully supports 7 languages: 简体中文 / 繁體中文 / English / Español / Deutsch / Français / 日本語
  466. - 🌐 One-click language switch in settings page, takes effect immediately
  467. - 🌐 MCP tool responses follow language setting, AI replies automatically use the corresponding language
  468. - 🌐 Switching language auto-regenerates steering rules for all installed projects
  469. **New: Web Dashboard Settings Page**
  470. - ⚙️ Language switch, theme settings, system info display
  471. - ⚙️ Database health check, repair, backup and other maintenance tools
  472. ### v1.0.3
  473. **Optimization: Memory Search**
  474. - 🔍 `recall` search supports OR/AND tag matching modes, fixing missed results with multi-tag searches
  475. - 🔍 Semantic search + tag filter defaults to OR matching (broader), tags-only browsing keeps AND matching (more precise)
  476. <details>
  477. <summary>📋 v0.2.x and earlier changelog</summary>
  478. See [CHANGELOG-archive.md](docs/CHANGELOG-archive.md)
  479. </details>
  480. ## License
  481. Apache-2.0

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.