← 返回 Skills 市场
359
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install s1kasl-askdjna
功能描述
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
安全使用建议
This skill appears to do what it says, but review and control what gets written and shared: 1) Check the GitHub repo URL before cloning and inspect any hook code that will be placed in ~/.openclaw/hooks. 2) Avoid logging secrets, credentials, or full transcripts — redact sensitive content before appending to LEARNINGS.md/ERRORS.md. 3) Be cautious when promoting learnings into project or .github files (those may be committed/pushed and become public). 4) If you enable the hook, test it in a sandbox or with a copy of your workspace first so you can verify its behavior and disable it if undesired.
功能分析
Type: OpenClaw Skill
Name: s1kasl-askdjna
Version: 1.0.0
The skill 'self-improvement-www' (SKILL.md) instructs the agent to establish a self-modification loop by logging data and 'promoting' it to behavioral guideline files like SOUL.md, which functions as a persistent prompt-injection vector. It also directs the agent to modify environment configurations (e.g., .claude/settings.json) to execute external shell scripts (scripts/activator.sh) not included in the bundle, posing a risk of unauthorized code execution. While the stated purpose is benign, these mechanisms provide a framework for persistence and behavioral manipulation via the referenced repository (https://github.com/peterskoett/self-improving-agent.git).
能力评估
Purpose & Capability
The skill's declared aim (capturing learnings/errors and promoting them into workspace/project files) matches the instructions: creating ~/.openclaw/.learnings or project .learnings, appending markdown entries, and optionally promoting to AGENTS.md/CLAUDE.md/.github files. No unrelated credentials, binaries, or heavyweight installs are requested.
Instruction Scope
Instructions direct the agent to create and write to local files (e.g., ~/.openclaw/workspace/.learnings and project .learnings) and to use OpenClaw session utilities (sessions_list, sessions_history, sessions_send). This is reasonable for a logging/promote workflow, but it also means session transcripts or user corrections could be recorded and shared across sessions if used—review what gets logged and avoid including secrets or sensitive transcripts.
Install Mechanism
No automated install spec; manual install suggestions point to a public GitHub repo (git clone). That's a normal, low-risk mechanism for an instruction-only skill. No downloads from obscure URLs or archives are present.
Credentials
The skill requests no environment variables, no credentials, and no config paths beyond writing to the user's OpenClaw workspace and project files. The requested access is proportional to its stated purpose of local logging and promotion.
Persistence & Privilege
The skill suggests copying a hook into ~/.openclaw/hooks and enabling it, which modifies the user's environment and causes recurring reminders. It is not marked always:true and does not require elevated credentials, but enabling hooks will give it a persistent presence in sessions—ensure you want automatic reminders and understand what the hook will do.
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 s1kasl-askdjna - 安装完成后，直接呼叫该 Skill 的名称或使用
/s1kasl-askdjna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 self-improvement-www skill for OpenClaw and generic agent workspaces.
- Provides structured logging for learnings, errors, and feature requests in Markdown files.
- Outlines clear triggers for logging (errors, corrections, feature requests, knowledge gaps, best practices).
- Includes detailed templates for LEARNINGS.md, ERRORS.md, and FEATURE_REQUESTS.md entries.
- Supports workflow for promoting important learnings to project memory (AGENTS.md, SOUL.md, TOOLS.md, etc).
- Offers setup instructions for OpenClaw and manual integration with other agent environments.
元数据
常见问题
ksdsl-skilll 是什么？
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 359 次。
如何安装 ksdsl-skilll？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install s1kasl-askdjna」即可一键安装，无需额外配置。
ksdsl-skilll 是免费的吗？
是的，ksdsl-skilll 完全免费（开源免费），可自由下载、安装和使用。
ksdsl-skilll 支持哪些平台？
ksdsl-skilll 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 ksdsl-skilll？
由 f1zzyw（@f1zzyw）开发并维护，当前版本 v1.0.0。
推荐 Skills