← 返回 Skills 市场
Self Improving Agent CN
作者
zhengxinjipai
· GitHub ↗
· v1.0.0
24950
总下载
29
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install self-improving-agent-cn
功能描述
AI自我改进与记忆系统 - 解决'同类错误反复犯、用户纠正不长记性'的痛点。自动捕获错误、用户纠正、最佳实践，并转化为长期记忆。
安全使用建议
Review this skill carefully before installing. It is not showing remote execution or hidden exfiltration, but it can persist user text, command failures, and suggested fixes into long-term memory and agent instruction files. Only use it if you are comfortable with automatic memory writes, and avoid letting it save secrets, raw command lines, or sudo-based fixes without explicit review.
能力评估
Credentials
The skill writes global memory and instructs synchronization to current-project .learnings, AGENTS.md, and MEMORY.md. That is high-impact for an agent-behavior skill because project files may be version-controlled or reused across sessions.
Install Mechanism
Installation is simple local file usage with no remote payloads or package dependencies observed. It creates a predictable global memory directory under ~/.openclaw/memory/self-improving but does not define restrictive permissions or a permissions boundary.
Instruction Scope
Automatic triggers such as broad correction phrases and command failures are under-scoped, and the retrieved records are presented as recommendations or corrections without labeling them as untrusted data or requiring confirmation before reuse.
Persistence & Privilege
Persistent plaintext storage is central to the skill, but it has no enforced redaction, retention, approval, or deletion controls. It also includes examples where a future command may automatically switch to sudo after a prior permission failure.
Purpose & Capability
The stated purpose matches the scripts: logging errors, corrections, and best practices, then checking them later. The concern is that the skill also tells the agent to automatically promote memories into AGENTS.md and MEMORY.md, turning free-form user or tool text into persistent agent guidance.
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 self-improving-agent-cn - 安装完成后，直接呼叫该 Skill 的名称或使用
/self-improving-agent-cn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 AI自我改进与记忆系统:
- 自动捕获、记录命令错误、用户纠正与最佳实践并转为长期记忆
- 支持跨项目同步与全局+项目级记忆管理
- 执行前智能检查相关记忆，规避重复犯错
- 触发机制自动化，无需手动干预
- 重要经验自动更新至 AGENTS.md、MEMORY.md 等文档
- 增强与原有 self-improvement skill 的对比与兼容
元数据
常见问题
Self Improving Agent CN 是什么？
AI自我改进与记忆系统 - 解决'同类错误反复犯、用户纠正不长记性'的痛点。自动捕获错误、用户纠正、最佳实践，并转化为长期记忆。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 24950 次。
如何安装 Self Improving Agent CN？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install self-improving-agent-cn」即可一键安装，无需额外配置。
Self Improving Agent CN 是免费的吗？
是的，Self Improving Agent CN 完全免费（开源免费），可自由下载、安装和使用。
Self Improving Agent CN 支持哪些平台？
Self Improving Agent CN 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 Self Improving Agent CN？
由 zhengxinjipai（@zhengxinjipai）开发并维护，当前版本 v1.0.0。
推荐 Skills