據 AlphaMoat 統計,截至 2026-10,Smart Memory 累計下載 13,033 次、獲 3 個星標,按累計下載量在本次參與排名的 10,914 個 AI 技能中排名第 2,128 位,在「AI 智慧體」分類(1,632 個)中排名第 285 位。
列表預設按累計下載量排序,也可按星標排序;詳情頁排名按累計下載量計算。
| 平台 | 累計下載 | 星標 | 評論 |
|---|---|---|---|
| ClawHub ↗ | 8353 | 3 | - |
| SkillHub ↗ | 1.3萬 | 3 | — |
各平台的計數口徑不同,統計時間不一致。
技能以逐條轉錄為基礎做記憶攝入與檢索,支援偏好、身份與任務狀態等型別的長期記憶,並用帶證據的修訂鏈管理記憶更替。它還能從轉錄歷史確定性重建,並把提示片語裝限制在 token 預算內、附帶追溯後設資料。 不得將憑證或金鑰存入記憶;定期檢視持久化內容並移除敏感或過期資料。
資料釋出於 2026-10-29
以下是根據技能文件生成的示例,不代表實際執行結果。
2026-09 起統計方法調整:同名技能按作者分別統計,並覆蓋全部技能。圖中虛線為調整前的月份,與之後的資料不直接比較,也不計算跨越調整的增長。 該技能有同名的其他版本,調整前的資料歸屬當前下載量最大的這一個。
| 月份 | 累計下載 | 環比 | 星標 |
|---|---|---|---|
| 2026-10 | 1.3萬 | — | 3 |
| 2026-09 | 9246 | — | 3 |
| 2026-08 | 8396 | +2.8% | 3 |
| 2026-07 | 8166 | +4.4% | 3 |
| 2026-06 | 7825 | +4.0% | 3 |
| 2026-05 | 7527 | — | 3 |
| 2026-03 | 5836 | — | 3 |
以下是作者釋出的原始檔案,按純文字顯示,本站未做修改或稽核;其中的連結不可點選,命令請核實後再執行。抓取於 2026-10-23。
--- name: smart-memory description: Persistent local transcript-first memory for OpenClaw via a Node adapter and FastAPI engine. --- # Smart Memory v3.1 Skill Smart Memory v3.1 is a local transcript-first cognitive memory runtime with revision-aware derivation, pinned context lanes, entity-aware retrieval, and bounded prompt composition. Core runtime:
- Node adapter: `smart-memory/index.js`
- Local API: `server.py`
- System facade: `cognitive_memory_system.py`
- Canonical store: `storage/sqlite_memory_store.py` plus `transcripts/`
## Core Capabilities
- transcript-first ingest and per-message transcript logging
- typed long-term memory including `preference`, `identity`, and `task_state`
- evidence-backed revision lifecycle decisions and supersession chains
- explicit core and working memory lanes
- entity-aware retrieval with lightweight relationship hints
- deterministic rebuild from transcript history
- hot-memory compatibility projection for working context
- strict token-bounded prompt composition with trace metadata
- inspection endpoints for transcripts, evidence, history, lanes, and eval runs
## OpenClaw Integration
Use the native wrapper package in `skills/smart-memory-openclaw/`.
Primary exports:
- `createSmartMemorySkill(options)`
- `createOpenClawHooks({ skill, agentIdentity, summarizeWithLLM })`
The wrapper remains stable while the backend is now transcript-first under the hood.
### Tool Interface
1. `memory_search`
- purpose: query relevant memory through `/retrieve`
- supports `query`, `type`, `limit`, `min_relevance`, and optional `conversation_history`
- health-checks the backend before execution
2. `memory_commit`
- purpose: persist important facts, decisions, beliefs, goals, or session summaries
- health-checks the backend before execution
- serializes commits to protect local embedding throughput
- queues failed commits in `.memory_retry_queue.json`
3. `memory_insights`
- purpose: surface pending background insights
- health-checks the backend before execution
- calls `/insights/pending`
## API Endpoints
Core endpoints:
- `GET /health`
- `POST /ingest`
- `POST /retrieve`
- `POST /compose`
- `POST /run_background`
- `GET /memories`
- `GET /memory/{memory_id}`
- `GET /insights/pending`
Transcript and inspection endpoints:
- `POST /transcripts/message`
- `GET /transcripts/{session_id}`
- `GET /transcript/message/{message_id}`
- `GET /memory/{memory_id}/evidence`
- `POST /revise`
- `GET /memory/{memory_id}/history`
- `GET /memory/{memory_id}/active`
- `GET /memory/{memory_id}/chain`
- `GET /lanes/{lane_name}`
- `POST /lanes/{lane_name}/{memory_id}`
- `DELETE /lanes/{lane_name}/{memory_id}`
- `POST /rebuild`
- `POST /rebuild/{session_id}`
- `GET /eval/suite/{suite_name}`
- `GET /eval/case/{case_id}`
## Operating guidance
- query memory before speaking when continuity matters
- do not claim prior context unless retrieval actually supports it
- transcripts are canonical, memories are derived
- treat SQLite as canonical runtime storage
- treat JSON as offline export or backup only
- keep CPU-only PyTorch policy intact
## Deprecated
Legacy vector-memory CLI artifacts remain deprecated and should not be revived.
將開發過程中的經驗、錯誤與修正記錄為 Markdown 檔案,供日後重用並提升為工作區記憶。
用自然語言描述需求,在本機與遠端技能市場尋找技能並取得推薦。
給 AI agent 用的瀏覽器自動化命令列技能。
將使用者明確糾正與自我檢視整理成分層記憶,並依使用情況升級或封存。
規劃 AI 代理主動跟進、工作記憶與回顧流程的工作區設定指南。
指導如何建立與更新有效 Claude 技能的寫作指南。
在安裝第三方技能之前,按清單審查來源、許可權與危險模式的技能。
把盛唐詩人李白的詩學觀、意象系統與口吻,蒸餾成 AI 可呼叫的角色包。
資料月份:2026-10 · 下載/星標數聚合自公開技能註冊庫(ClawHub、SkillHub),按月更新。口徑說明見資料方法。