Comprehensive evaluation toolkit for analyzing Claude skills across security, quality, utility, and compliance dimensions. This skill should be used when users need to evaluate a skill before installation, review before publishing, or assess overall quality and safety. Performs 5-layer security analysis, validates structure and documentation, checks compliance with skill-creator guidelines, and generates markdown reports with scoring and recommendations.
Version: 1.2.3 Status: Production Ready License: MIT (or your choice)
Comprehensive evaluation toolkit for analyzing Claude skills across security, quality, utility, and compliance dimensions.
# Evaluate a skill
python scripts/evaluate_skill.py /path/to/skill
# Security-focused evaluation
python scripts/evaluate_skill.py --mode security /path/to/skill
# Pre-publication review
python scripts/evaluate_skill.py --mode pre-publish /path/to/skill
# Save report to file
python scripts/evaluate_skill.py -o report.md /path/to/skill
✅ 4-Dimensional Evaluation
✅ Advanced Security Scanning
✅ Professional Reports
✅ Cross-Platform Support
Self-evaluation results for skill-evaluator v1.2.1:
| Dimension | Score | Grade | |-----------|-------|-------| | Security | 100.0/100 | Perfect ✅ | | Quality | 90.3/100 | Excellent ✅ | | Utility | 100.0/100 | Perfect ✅ | | Compliance | 95.0/100 | Excellent ✅ | | Overall | 96.9/100 | HIGHLY RECOMMENDED ✅ |
🔮 Community Trust Score - 5th dimension incorporating:
See VERSION_ROADMAP.md for complete v2.0 plans.
Install PyYAML:
pip install pyyaml
Recommended Method: Install as a plugin
macOS/Linux:
cp -r skill-evaluator ~/.claude/plugins/
Windows:
xcopy /E /I skill-evaluator "%USERPROFILE%\.claude\plugins\skill-evaluator"
skill-evaluator.zipmacOS/Linux:
unzip skill-evaluator.zip -d ~/.claude/plugins/
Windows:
# Using PowerShell
Expand-Archive -Path skill-evaluator.zip -DestinationPath "$env:USERPROFILE\.claude\plugins"
After installation, Claude should recognize the skill. Test it:
"Evaluate the skill at ~/.claude/plugins/skill-evaluator"
Note: Claude on the web does not currently support custom plugins/skills. You can:
git clone <this-repo>
cd skill-evaluator
python scripts/evaluate_skill.py /path/to/skill
references/ as context in web conversationsPlace skill in a shared location and symlink:
macOS/Linux:
ln -s /shared/skills/skill-evaluator ~/.claude/plugins/skill-evaluator
Windows (as Administrator):
mklink /D "%USERPROFILE%\.claude\plugins\skill-evaluator" "C:\Shared\Skills\skill-evaluator"
Set custom plugin location via environment variable:
export CLAUDE_PLUGINS_DIR=/custom/path/to/plugins
Check the plugin directory exists:
ls ~/.claude/plugins/skill-evaluator
Verify SKILL.md exists:
cat ~/.claude/plugins/skill-evaluator/SKILL.md
Restart Claude Code/Desktop
If evaluation fails with import errors:
# Check Python version
python --version # Should be 3.8+
# Install PyYAML
pip install pyyaml
# If using virtual environment
source venv/bin/activate # macOS/Linux
venv\Scripts\activate # Windows
pip install pyyaml
macOS/Linux:
chmod -R 755 ~/.claude/plugins/skill-evaluator
Windows: Right-click folder → Properties → Security → Ensure your user has Read & Execute permissions
Default plugin directories by platform:
| Platform | Path |
|----------|------|
| macOS | ~/.claude/plugins/ |
| Linux | ~/.claude/plugins/ |
| Windows | %USERPROFILE%\.claude\plugins\ |
Full path examples:
/Users/username/.claude/plugins/skill-evaluator/home/username/.claude/plugins/skill-evaluatorC:\Users\username\.claude\plugins\skill-evaluatorpip install pyyaml
python scripts/evaluate_skill.py --mode security ~/downloads/new-skill.zip
python scripts/evaluate_skill.py --mode pre-publish ~/my-skill/
python scripts/evaluate_skill.py \
-o report.md \
--json results.json \
~/skills/my-skill/
⚠️ This evaluation CANNOT determine with certainty that a skill is safe.
This tool provides pattern-based static analysis as ONE input into your security decision. You are responsible for:
See report disclaimers for complete limitations.
Contributions welcome! Please:
MIT License - See LICENSE file for details.
Copyright (c) 2025 skill-evaluator contributors
Developed for the Claude Code community.
Special thanks to:
Evaluate responsibly. Trust, but verify. 🔍
Search for places (restaurants, cafes, etc.) via Google Places API proxy on localhost.
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Start voice calls via the OpenClaw voice-call plugin.
Notion API for creating and managing pages, databases, and blocks.
Gemini CLI for one-shot Q&A, summaries, and generation.
Category:developer