← 返回 Skills 市场
mar-elite-longterm-memory
作者
MarjorieBroad
· GitHub ↗
· v1.0.0
· MIT-0
67
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install mar-elite-longterm-memory
功能描述
Ultimate AI agent memory system for Cursor, Claude, ChatGPT & Copilot. WAL protocol + vector search + git-notes + cloud backup. Never lose context again. Vib...
安全使用建议
This skill appears to do what it claims: manage local memory files and optionally back them up / auto-extract facts via SkillBoss. The main risk is privacy/exfiltration: if you provide SKILLBOSS_API_KEY the skill's documented workflows will send conversation text, extracted facts, and memory items to a third-party endpoint (api.heybossai.com). Before installing: 1) Decide whether you trust the SkillBoss service and its privacy policy; avoid sending passwords, secrets, or sensitive PII. 2) Consider using an ephemeral or restricted API key, or omit SKILLBOSS_API_KEY to keep everything local. 3) Review and run the included bin/elite-memory.js in a safe test workspace to confirm behavior. 4) If you need stronger guarantees, prefer self-hosted vector/backup providers or remove/disable the auto-extraction/cloud steps. 5) Ask the author for precise retention/ deletion policies for data stored on SkillBoss and for an explicit toggle to opt out of automatic uploads. If you cannot confirm those controls, treat the SkillBoss integration as a potential data-exfiltration channel and proceed cautiously.
能力标签
能力评估
Purpose & Capability
Name, description, and files all describe a multi-layer memory system. The single required env var (SKILLBOSS_API_KEY) matches the documented SkillBoss cloud backup and auto-extraction features. The included CLI only creates/reads markdown files and checks for a LanceDB path — those actions align with the stated purpose.
Instruction Scope
SKILL.md and README explicitly show code that sends conversation data, embeddings, and stored memory to api.heybossai.com using SKILLBOSS_API_KEY. The documentation describes 'auto-extraction' and 'auto-recall' as automatic processes; this gives the agent broad discretion to extract and transmit facts and context. That is functionally necessary for cloud backup/embedding, but it also creates a real risk of sending sensitive user data to a third party. The CLI itself does not perform network calls, but the runtime instructions instruct the agent to do so.
Install Mechanism
No install spec is provided (instruction-only skill); the packaged code is a small, non-obfuscated CLI that only writes local files. There are no downloads from untrusted URLs and no extract/install steps that would write arbitrary code to system locations. package.json contains an optional dependency (mem0ai) but nothing automatically installs arbitrary remote archives.
Credentials
The skill requests a single environment variable (SKILLBOSS_API_KEY), which is proportionate for the documented SkillBoss integration. However, the SKILL.md instructs storing conversation content and memory to the SkillBoss KV/chat/embedding endpoints — granting that API key allows remote storage and analysis of potentially sensitive conversations. The skill also instructs modifying or referencing user config (~/.clawdbot/clawdbot.json) but doesn't declare config-path requirements; that's a minor transparency gap.
Persistence & Privilege
always:false (no force-installed privilege). The skill writes files in the current workspace (SESSION-STATE.md, MEMORY.md, memory/) which is expected for a memory tool. It does not request system-wide privileges or modify other skills' configs in the provided files.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install mar-elite-longterm-memory - 安装完成后，直接呼叫该 Skill 的名称或使用
/mar-elite-longterm-memory触发 - 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
v1.0.0
- Initial release of Elite Longterm Memory providing a comprehensive AI agent memory system.
- Integrates WAL protocol, vector search (LanceDB), git-notes knowledge graph, and cloud backup.
- Supports memory persistence across sessions, curated human-readable archives, and optional SkillBoss API hub cloud sync.
- Introduces automatic fact extraction and semantic search across all stored memories.
- Ready for use with Cursor, Claude, ChatGPT & Copilot—enabling context-aware, "vibe-coding" workflows.
元数据
常见问题
mar-elite-longterm-memory 是什么？
Ultimate AI agent memory system for Cursor, Claude, ChatGPT & Copilot. WAL protocol + vector search + git-notes + cloud backup. Never lose context again. Vib... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 67 次。
如何安装 mar-elite-longterm-memory？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install mar-elite-longterm-memory」即可一键安装，无需额外配置。
mar-elite-longterm-memory 是免费的吗？
是的，mar-elite-longterm-memory 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
mar-elite-longterm-memory 支持哪些平台？
mar-elite-longterm-memory 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 mar-elite-longterm-memory？
由 MarjorieBroad（@marjoriebroad）开发并维护，当前版本 v1.0.0。
推荐 Skills