← Back to Skills Marketplace
97
Downloads
1
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install skill-check
Description
审查和优化现有技能（Skill）结构质量。当用户请求「优化技能」「审查技能」或需要检查技能规范性时触发。
Usage Guidance
这是一个结构审查与自动修复工具，整体内部一致且没有网络外联或凭据请求。但在安装/运行前请注意：
- review-loop.py 会调用 analyze.py 和 fix.py：fix.py 会在你指定的 skill 目录下创建 scripts、references、assets 子目录（会修改文件系统）。仅对你明确允许的目录运行自动修复。建议先在技能副本上运行 analyze.py 以查看报告，再决定是否运行 fix.py 或 review-loop.py。
- SKILL.md 提到的 `todo_write` 并未在 manifest 中声明；如果你的 agent/environment 提供该工具，审查结果可能会被发送到任务系统。确认 todo_write 的实现与权限（避免在不受信的外部服务上自动发布敏感内容）。
- 建议在受限环境（或对目标技能做备份）下试运行；如需更高信心，可打开并人工审阅 scripts/*.py 源码（它们是小脚本且易于审计）。
Capability Analysis
Package: SKILL审查 (xpi)
Version:
Description: 审查和优化SKILL目录结构和逻辑完整性
The package is a utility toolset designed to analyze and optimize the directory structure and documentation of 'SKILL' modules. The Python scripts (analyze.py, fix.py, and review-loop.py) perform local file system operations such as reading Markdown files, checking for specific directory structures, and creating missing folders (scripts, references, assets). It uses subprocess.run to execute its own internal scripts but does not perform any network requests, exfiltrate sensitive data, or download external payloads. The logic is focused entirely on static analysis and local maintenance of the project structure.
Capability Assessment
Purpose & Capability
名称/描述与代码和说明一致：脚本 analyze.py、fix.py、review-loop.py 用于静态审查与自动修复，references 文档提供模式。没有请求与技能目的不相关的凭据、二进制或外部安装。
Instruction Scope
SKILL.md 明确以静态审查为主并指导运行本包内脚本（python3 scripts/analyze.py 等）。注意：工作流中提到会调用 `todo_write` 工具来创建待办项，但该工具未在 manifest/requirements 中声明——这只是外部 agent 工具的调用约定，可能会将审查结果写入外部任务系统；此外 review-loop.py 调用 fix.py，会在目标技能目录中创建 directories（scripts/references/assets），即会对文件系统做出修改，这与自动修复功能一致但需要用户确认目标路径。
Install Mechanism
无安装 spec（仅随技能分发脚本和文档）。脚本均为本地 Python 脚本，无下载、无外部包安装命令、未见远程提取或可疑 URL。
Credentials
不要求任何环境变量或凭据。脚本在本地读取/分析目标技能目录内容并创建目录（fix.py），这些操作与技能目的相符。
Persistence & Privilege
flags: always=false（正常）。技能允许模型调用（默认），但无声明要求长期驻留或修改其他技能的配置；唯一有写操作的是 fix.py（仅对指定目标目录创建缺失目录），这是预期的自动修复行为。
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install skill-check - After installation, invoke the skill by name or use
/skill-check - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Major update: skill purpose now focused on structural and logical review/optimization for skills.
- Added detailed review criteria covering structural, logical, and Agent compatibility dimensions.
- Workflow now runs static script checks, then LLM-based reasoning for issues and actionable solutions.
- Included comprehensive directory/file structure guidelines and review report templates.
- Introduced self-iteration review loop and clear failure handling.
- All content and instructions updated to be more prescriptive and process-oriented.
Metadata
Frequently Asked Questions
What is SKILL审查?
审查和优化现有技能（Skill）结构质量。当用户请求「优化技能」「审查技能」或需要检查技能规范性时触发。 It is an AI Agent Skill for Claude Code / OpenClaw, with 97 downloads so far.
How do I install SKILL审查?
Run "/install skill-check" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is SKILL审查 free?
Yes, SKILL审查 is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does SKILL审查 support?
SKILL审查 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created SKILL审查?
It is built and maintained by 鹿Sir (@linchuncheng); the current version is v1.0.0.
More Skills