← Back to Skills Marketplace
240
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install self-skill-evolver
Description
Analyze and improve OpenClaw agent skills by tracking usage, checking skill health, scanning code quality, and generating actionable improvement suggestions....
Usage Guidance
This package is broadly coherent with its described purpose (static analysis, monitoring, reporting). Before installing: 1) Inspect the code yourself — there are several code-quality/runtime issues (truncated/buggy functions) that could cause errors; for example, database.models appears to be truncated and references or returns undefined symbols (e.g., add_feedback returns 'c'), and CLI/reporter call methods like get_all_skills and get_feedback_stats which aren't visible in the included model code. 2) Be aware it will read all files under the skills directory (~/.openclaw/skills by default) and store analysis and usage logs in ~/.openclaw/skill-evolver/skill_evolver.db — do not run it if those directories contain secrets you don't want duplicated. 3) If you plan to install, prefer reviewing and running the package in a disposable environment (virtualenv or container), and consider auditing or fixing the apparent bugs (missing functions, inconsistent data handling) before using it in production. If you want, I can point out the specific locations of the suspicious code snippets and list the functions referenced but missing so you can inspect or patch them.
Capability Analysis
Type: OpenClaw Skill
Name: self-skill-evolver
Version: 0.1.2
The self-skill-evolver bundle is a legitimate utility designed to monitor, analyze, and report on the performance and code quality of OpenClaw agent skills. It uses a local SQLite database to store usage logs and health scores, and its analysis logic (in analyzer.py) uses safe methods like ast.parse for code inspection. While a potential SQL injection vulnerability exists in database/models.py due to string formatting in the clear_old_data and get_usage_stats methods, this appears to be an unintentional implementation flaw rather than a malicious backdoor, especially as the CLI enforces integer types for the affected parameters.
Capability Assessment
Purpose & Capability
The skill's name/description match its behavior: it reads SKILL.md, package.json and Python files under a skills directory and records usage/health metrics. However, the registry metadata claims "instruction-only" (no install spec) while the package actually contains multiple Python modules and a CLI; that's an inconsistency worth checking. The tool also writes a SQLite DB under the user's home (~/.openclaw/skill-evolver) which is coherent with monitoring functionality but should be noted.
Instruction Scope
SKILL.md and the code instruct the agent to analyze skill directories, parse files, compute scores, and write reports. The analyzer only reads files and uses ast.parse (static parsing) — it does not appear to execute inspected code. It will read arbitrary SKILL.md, package.json and .py files from the configured skills directory (default ~/.openclaw/skills) and store metadata and logs; this is in-scope for the stated purpose.
Install Mechanism
No install spec is provided (instruction-only in registry), so nothing will be fetched automatically by the platform. The repo includes a requirements.txt and package.json indicating Python dependencies (PyYAML, GitPython, etc.); installing manually would require standard Python package installation. No remote download or opaque installer URLs are present in the provided files.
Credentials
The skill does not request environment variables, external credentials, or special binaries. It reads/writes to the user's home directory and SQLite DB, which is proportionate to monitoring and reporting functionality.
Persistence & Privilege
The skill persists data locally (SQLite DB under ~/.openclaw/skill-evolver) and will create directories there. always: false and no elevated privileges are requested. Autonomous invocation is enabled by default (platform default) but is not combined with other concerning flags here. Still, persistence means it will keep logs and reports on disk — be aware of what data may be recorded.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install self-skill-evolver - After installation, invoke the skill by name or use
/self-skill-evolver - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.2
- Initial public release of self-skill-evolver.
- Provides a Python CLI tool to analyze and improve OpenClaw agent skills.
- Includes usage monitoring, skill health scoring, and code/SKILL.md analysis.
- Generates automated improvement suggestions and supports JSON/Markdown/HTML reports.
- Offers commands for analyzing, reporting, logging, feedback, and clearing history.
- Requires Python 3.8+, SQLite3, PyYAML, and optionally GitPython.
Metadata
Frequently Asked Questions
What is Self Skill Evolver?
Analyze and improve OpenClaw agent skills by tracking usage, checking skill health, scanning code quality, and generating actionable improvement suggestions.... It is an AI Agent Skill for Claude Code / OpenClaw, with 240 downloads so far.
How do I install Self Skill Evolver?
Run "/install self-skill-evolver" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Self Skill Evolver free?
Yes, Self Skill Evolver is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Self Skill Evolver support?
Self Skill Evolver is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Self Skill Evolver?
It is built and maintained by haidong (@harrylabsj); the current version is v0.1.2.
More Skills