← 返回 Skills 市场
getmem.ai Memory
作者
NimbleV2023
· GitHub ↗
· v1.0.3
· MIT-0
128
总下载
0
收藏
0
当前安装
3
版本数
在 OpenClaw 中安装
/install getmem-openclaw
功能描述
Persistent memory for every user via getmem.ai. Automatically ingests conversations and injects relevant context before each agent reply. Two API calls. $20...
使用说明 (SKILL.md)
getmem.ai Memory Plugin for OpenClaw
Gives your OpenClaw agent persistent memory via getmem.ai.
Install
openclaw plugins install clawhub:@getmem/openclaw-getmem
openclaw config set plugins.openclaw-getmem.apiKey gm_live_...
openclaw gateway restart
Get your API key at platform.getmem.ai — $20 free credit on signup.
How it works
- Before each reply — fetches memory context for the user and injects it into the agent prompt
- After each reply — ingests the exchange into getmem for future retrieval
Memory accumulates automatically. No code changes needed.
安全使用建议
This plugin appears to be what it claims: it sends user messages and assistant replies (plus user identifiers) to getmem.ai and injects returned memory into the agent prompt. Before installing: 1) Be aware all conversation content forwarded to getmem.ai will be stored/used by that external service — review getmem.ai's privacy and security policies and the API key provisioning. 2) Do not enable this for channels or users that exchange sensitive or regulated data unless you have accepted the external storage risk. 3) Verify the plugin config (apiKey and baseUrl) points to the official getmem endpoint; changing baseUrl would redirect data to whatever endpoint you configure. 4) The repository contains source — you can audit the included files (dist/src) yourself; the code shows no hidden endpoints or unrelated credential access. 5) If unsure, test in an isolated environment or disable (enabled: false) until you’ve confirmed compliance with your data-handling policies.
功能分析
Type: OpenClaw Skill
Name: getmem-openclaw
Version: 1.0.3
The plugin provides persistent memory for OpenClaw agents by integrating with the getmem.ai service. It uses the 'message:received' hook to fetch relevant context and the 'message:sent' hook to store conversation history at https://memory.getmem.ai. All code in src/index.ts and dist/index.js is consistent with the stated purpose, and no evidence of unauthorized data exfiltration, malicious execution, or obfuscation was found.
能力标签
能力评估
Purpose & Capability
Name/description, README, SKILL.md, package.json, and the compiled/source code all describe a memory plugin that calls getmem.ai's /v1/memory/get and /v1/memory/ingest endpoints. The plugin requires a getmem API key via plugin config (not an environment variable), which is appropriate for this function. Minor inconsistencies (small variations in IDs/names across dist/src files and an odd SKILL.md frontmatter 'install' entry referencing 'kind: pip' for an npm package) appear to be documentation/metadata quirks rather than functional mismatches.
Instruction Scope
Runtime hooks only read message context/session keys and perform two external API calls per exchange (fetch memory before reply and ingest the exchange after reply). They do not read unrelated files, env vars, or system config. This behavior is within the stated scope but it does mean all user messages and assistant replies (and user identifiers) are transmitted to the external service — a privacy-sensitive action that is expected for a memory plugin.
Install Mechanism
Registry shows this as an instruction-only plugin (no platform install spec), but the package includes full source, package.json, and a package-lock that references an npm package 'getmem-ai'. There is no high-risk remote download URL in the manifest. The SKILL.md frontmatter has an inconsistent 'install' block (id: npm-getmem-ai, kind: pip) which is likely a documentation mistake. Overall install mechanism is standard for OpenClaw plugins and not unusual.
Credentials
The plugin asks for a single getmem API key (via plugin config), which is proportional to its function. However, the plugin will transmit user IDs and full message content to an external service (default baseUrl https://memory.getmem.ai). That data flow is expected but has privacy and compliance implications — there are no unrelated credentials requested.
Persistence & Privilege
The plugin does not set always:true and does not request elevated platform privileges. It registers hooks (message:received and message:sent) within normal plugin behavior and does not modify other plugins or global agent settings beyond its own config.
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 getmem-openclaw - 安装完成后，直接呼叫该 Skill 的名称或使用
/getmem-openclaw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
v1.0.3
Fix: openclaw.plugin.json id aligned to getmem-openclaw (was openclaw-getmem causing config mismatch on restart)
v1.0.2
Fix: correctly strip channel prefix from user IDs (telegram:ID → ID); ingest now reliably stores every conversation turn
v1.0.1
Fix hook registration: add required name option, fix npm version, add startup self-test
元数据
常见问题
getmem.ai Memory 是什么？
Persistent memory for every user via getmem.ai. Automatically ingests conversations and injects relevant context before each agent reply. Two API calls. $20... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 128 次。
如何安装 getmem.ai Memory？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install getmem-openclaw」即可一键安装，无需额外配置。
getmem.ai Memory 是免费的吗？
是的，getmem.ai Memory 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
getmem.ai Memory 支持哪些平台？
getmem.ai 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 getmem.ai Memory？
由 NimbleV2023（@nimblev2023）开发并维护，当前版本 v1.0.3。
推荐 Skills