← 返回 Skills 市场
paudyyin

Memory (TencentDB)

作者 paudyyin · GitHub ↗ · v0.3.6 · MIT-0
cross-platform ⚠ suspicious
72
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install memory-tencentdb
功能描述
四层记忆系统插件 — 自动捕获对话、提取结构化记忆、归纳场景、生成用户画像。支持 L0→L1→L2→L3 渐进式管线,本地 SQLite + 腾讯云向量数据库双后端。
使用说明 (SKILL.md)

memory-tencentdb 鈥?鍥涘眰璁板繂绯荤粺

涓?OpenClaw AI Agent 鎻愪緵 L0鈫扡1鈫扡2鈫扡3 闀挎湡璁板繂鑳藉姏銆?

鏋舵瀯

瀵硅瘽寮€濮?  鈫?Auto-Recall: 鍚戦噺/娣峰悎鎼滅储鐩稿叧璁板繂 + 鍔犺浇 Persona 鈫?娉ㄥ叆绯荤粺涓婁笅鏂?
瀵硅瘽缁撴潫
  鈫?L0: 褰曞埗瀵硅瘽娑堟伅 鈫?SQLite + JSONL 鍙屽啓
  鈫?Pipeline Scheduler: 杈惧埌 N 杞悗瑙﹀彂
     鈹溾攢鈹€ L1: LLM 鎻愬彇缁撴瀯鍖栬蹇?+ 鍚戦噺鍘婚噸 鈫?JSONL + SQLite
     鈹溾攢鈹€ L2: LLM 褰掔撼鍦烘櫙鍧?鈫?Markdown 鏂囦欢
     鈹斺攢鈹€ L3: LLM 鐢熸垚/鏇存柊鐢ㄦ埛鐢诲儚 鈫?persona.md

鍥涘眰璁板繂

灞傜骇 鍚嶇О 璇存槑
L0 瀵硅瘽褰曞埗 鍘熷瀵硅瘽娑堟伅鎹曡幏
L1 璁板繂鎻愬彇 缁撴瀯鍖栬蹇嗭紙persona/episodic/instruction锛?
L2 鍦烘櫙褰掔撼 鍦烘櫙鍧楋紙Scene Block锛?
L3 鐢ㄦ埛鐢诲儚 鐢ㄦ埛浜烘牸鐢诲儚锛圥ersona锛?

璁板繂绫诲瀷

  • persona锛氱敤鎴风殑绋冲畾灞炴€с€佸亸濂姐€佹妧鑳姐€佷环鍊艰
  • episodic锛氬瑙傚彂鐢熺殑鍔ㄤ綔銆佸喅瀹氥€佽鍒掓垨缁撴灉
  • instruction锛氱敤鎴峰 AI 鐨勯暱鏈熻涓鸿鍒?

瀹夎

clawhub install memory-tencentdb

閰嶇疆

鍦?openclaw.json 涓坊鍔狅細

{
  "plugins": {
    "allow": ["mx", "memory-tencentdb"],
    "load": {
      "paths": ["\x3C鎻掍欢璺緞>"]
    },
    "entries": {
      "memory-tencentdb": {
        "enabled": true,
        "config": {
          "llm": {
            "enabled": true,
            "baseUrl": "https://浣犵殑LLM绔偣/v1",
            "model": "浣犵殑妯″瀷ID"
          },
          "pipeline": {
            "everyNConversations": 3
          }
        }
      }
    }
  }
}

瀛樺偍鍚庣

  • sqlite锛堥粯璁わ級锛氭湰鍦?SQLite + sqlite-vec
  • tcvdb锛氳吘璁簯鍚戦噺鏁版嵁搴?

CLI 鍛戒护

openclaw memory-tdai seed --input conversations.json

鏁呴殰鎺掓煡

濡傛灉 L1 鎻愬彇涓虹┖锛?1. 妫€鏌?plugins.allow 鍖呭惈 "memory-tencentdb" 2. 妫€鏌?plugins.load.paths 鍖呭惈鎻掍欢璺緞 3. 妫€鏌?llm.enabled 涓?true 4. 妫€鏌?LLM 绔偣鍙闂?\r

安全使用建议
Review this before installing in any environment with private or regulated conversations. Confirm who can trigger memory search, whether raw conversation search should be disabled or restricted, set a finite retention period, and use local-only model/storage settings if conversation content must not be sent to third-party services. Also inspect the published package contents for the postinstall patch script before allowing install-time scripts to run.
能力标签
requires-sensitive-credentials
能力评估
Purpose & Capability
The stated purpose matches the advertised capabilities: automatic conversation capture, memory extraction, persona generation, search, and recall into future prompts. These are expected for a long-term memory plugin, but they handle sensitive user data.
Instruction Scope
The README says capture, extraction, persona updates, and recall are enabled by default, L0/L1 retention defaults to no cleanup, and agent-callable tools can search raw conversation history. The docs disclose these functions but do not clearly describe consent, redaction, authorization, or safe scoping boundaries.
Install Mechanism
package.json declares a postinstall command that runs a shell patch script and suppresses errors. The script body is not present in the provided artifact set, and the README does not explain this install-time mutation behavior.
Credentials
The plugin may send conversation-derived content to a configured remote LLM endpoint, embedding service, or TencentDB vector backend. This is coherent with the feature set but needs clearer data-sharing warnings before production use.
Persistence & Privilege
The plugin persists raw conversations, extracted memories, scene summaries, persona files, SQLite/vector data, metadata, and backups. Persistence is purpose-aligned, but default indefinite retention and model-visible recall increase privacy impact.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install memory-tencentdb
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /memory-tencentdb 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.3.6
Four-layer memory system plugin for OpenClaw - L0/L1/L2/L3 pipeline, SQLite + TencentDB backends
v0.3.5
Four-layer memory system: L0 conversation capture → L1 memory extraction → L2 scene induction → L3 persona generation. Supports local SQLite and Tencent Cloud VectorDB backends.
元数据
Slug memory-tencentdb
版本 0.3.6
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 2
常见问题

Memory (TencentDB) 是什么?

四层记忆系统插件 — 自动捕获对话、提取结构化记忆、归纳场景、生成用户画像。支持 L0→L1→L2→L3 渐进式管线,本地 SQLite + 腾讯云向量数据库双后端。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 72 次。

如何安装 Memory (TencentDB)?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install memory-tencentdb」即可一键安装,无需额外配置。

Memory (TencentDB) 是免费的吗?

是的,Memory (TencentDB) 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Memory (TencentDB) 支持哪些平台?

Memory (TencentDB) 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Memory (TencentDB)?

由 paudyyin(@paudyyin)开发并维护,当前版本 v0.3.6。

💬 留言讨论