← 返回 Skills 市场

getmem.ai Memory

Name: getmem.ai Memory
Author: nimblev2023

作者 NimbleV2023 · GitHub ↗ · v1.0.2 · MIT-0

cross-platform ⚠ suspicious

总下载

当前安装

版本数

在 OpenClaw 中安装

/install getmem

功能描述

Persistent memory for AI agents via getmem.ai. Call mem.get() before each LLM call to inject context, and mem.ingest() after each turn to save the conversation.

使用说明 (SKILL.md)

getmem.ai Memory Skill

Persistent memory for your AI agent via getmem.ai.

Setup

Set your API key in the environment:

export GETMEM_API_KEY=gm_live_YOUR_KEY_HERE

Get your key at https://platform.getmem.ai — $20 free credit on signup.

Usage

import getmem_ai as getmem, os

mem = getmem.init(os.environ["GETMEM_API_KEY"])

# Before each LLM call — get relevant memory context
result = mem.get(user_id, query=user_message)
context = result["context"]  # inject into system prompt

# After each turn — save both user + assistant messages
mem.ingest(user_id, messages=[
    {"role": "user", "content": user_message},
    {"role": "assistant", "content": reply},
])

How it works

mem.get() fetches only the relevant memories for the current query (semantic search)
Context is injected into your system prompt — typically 200-800 tokens
mem.ingest() saves the full conversation exchange asynchronously
Memory persists indefinitely — no TTL, no purge

Token savings

Standard approach: full conversation history every turn = 10,000-40,000 tokens With getmem: only relevant context = 200-800 tokens. Save up to 95% on context tokens.

Links

Website: https://getmem.ai
Platform: https://platform.getmem.ai
PyPI: https://pypi.org/project/getmem-ai/
npm: https://npmjs.com/package/getmem
OpenClaw plugin: clawhub:getmem-openclaw

安全使用建议

This skill appears to be a legitimate getmem.ai integration, but there are important mismatches and privacy implications you should consider before installing: - The SKILL.md requires a GETMEM_API_KEY and includes a pip install for getmem-ai, but the registry metadata does not declare that secret or the install step. Ask the publisher to correct the metadata so the required secret is explicit. - Using the skill will send user and assistant messages to getmem.ai and the documentation says memory persists indefinitely. Do not enable this skill for agents that will handle sensitive, regulated, or confidential data unless you confirm retention, encryption, access controls, and deletion policies with getmem.ai. - Verify the PyPI package (getmem-ai) is the official SDK (check maintainer, code repository, and recent releases) before installing. Prefer installing in an isolated/sandboxed environment first. - If you need memory but cannot risk sending data externally, consider a local or self-hosted memory solution instead. If you want to proceed, request that the publisher update the registry metadata to declare GETMEM_API_KEY as a required secret and to provide a homepage/source repository and clear privacy/retention documentation.

功能分析

Type: OpenClaw Skill Name: getmem Version: 1.0.2 The getmem skill is a legitimate integration for the getmem.ai persistent memory service. It facilitates storing and retrieving conversation context via the 'getmem-ai' PyPI package to optimize token usage. The skill's behavior is transparent, requires a user-provided API key, and contains no evidence of malicious intent, obfuscation, or unauthorized data access beyond its stated purpose.

能力标签

requires-sensitive-credentials

能力评估

ℹ Purpose & Capability

The SKILL.md and install metadata describe installing the getmem-ai package and calling mem.get()/mem.ingest(), which is coherent with a persistent-memory skill. However, the registry metadata claims no required env vars or install spec while the SKILL.md explicitly requires GETMEM_API_KEY and includes a pip install entry — a clear metadata/instruction mismatch.

⚠ Instruction Scope

Runtime instructions tell the agent to send user and assistant messages to getmem.ai (mem.ingest) and to fetch contextual memory (mem.get). This is expected for a memory service, but it means full conversation content is transmitted to an external service and stored indefinitely (no TTL/purge). The instructions also reference os.environ['GETMEM_API_KEY'] even though the registry metadata doesn't declare it. Transmitting potentially sensitive data off-host with long retention is a material privacy/security risk.

ℹ Install Mechanism

The SKILL.md includes a pip install (getmem-ai on PyPI), which is a common and expected distribution method (moderate risk). The registry metadata earlier reported no install spec — another inconsistency between registry and SKILL.md. No downloads from arbitrary URLs are used, which reduces installation risk.

⚠ Credentials

The skill requires an API key (GETMEM_API_KEY) according to SKILL.md, but the registry metadata lists no required env vars or primary credential. The missing declaration of a secret-level environment variable is an incoherence and increases the chance a user won't realize they're providing a long-lived secret that allows an external service to store conversation data.

✓ Persistence & Privilege

The skill does not request always:true and does not claim elevated system privileges. Autonomous invocation is allowed by default (normal). Note: combining autonomous invocation with remote indefinite storage increases blast radius for data exfiltration, but that is a behavioral/privacy concern rather than an OS-privilege escalation.

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
安装完成后，直接呼叫该 Skill 的名称或使用 /getmem 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

v1.0.2

Fix: remove JSON metadata block that triggered SUSPICIOUS flag; use clean YAML frontmatter; update OpenClaw plugin reference to clawhub:getmem-openclaw

v1.0.1

No user-facing changes in this version. - No file changes detected. - Functionality and documentation remain the same.

v1.0.0

- Initial release of the getmem skill for persistent AI memory using getmem.ai. - Remembers users across sessions; fetch relevant memory context before LLM calls with mem.get(). - Ingest and persist conversations after each turn with mem.ingest(). - Significantly reduces token usage by retrieving only relevant context (200–800 tokens). - Easy setup via API key and simple usage with clear Python code examples. - Supports both Python and npm packages and integration with OpenClaw.

元数据

Slug getmem

版本 1.0.2

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 3

常见问题

getmem.ai Memory 是什么？

Persistent memory for AI agents via getmem.ai. Call mem.get() before each LLM call to inject context, and mem.ingest() after each turn to save the conversation. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 85 次。

如何安装 getmem.ai Memory？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install getmem」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 v1.0.2。