hooks — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited hooks (Hook) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Convert MCP servers to Claude Skills with progressive disclosure, reducing context usage by 90%+.
stdio, http, and sse MCP server typesInstall the plugin using Claude Code's --plugin-dir option:
# Clone or download the plugin
git clone https://github.com/your-org/mcp-skills-plugin.git
# Run Claude Code with the plugin
claude --plugin-dir /path/to/mcp-skills-pluginOr add to your Claude Code configuration for persistent use.
pip install mcpOnce the plugin is installed, Claude can use the mcp-to-skill-converter skill.
#### List Available Servers
python ${PLUGIN_DIR}/skills/mcp-to-skill-converter/mcp_to_skill.py --listOutput:
📄 Servers in: /path/to/.mcp.json
Name Type Compatible Command
--------------------------------------------------------------------------------
github stdio ✅ Yes npx
context7 stdio ✅ Yes npx
brave-search stdio ✅ Yes npx
📊 Total: 10 servers, 8 compatible#### Convert a Single Server
python ${PLUGIN_DIR}/skills/mcp-to-skill-converter/mcp_to_skill.py --name githubThis will:
.claude/skills/mcp-skills/github/.mcp.json to prevent duplicate loading#### Convert All Compatible Servers
python ${PLUGIN_DIR}/skills/mcp-to-skill-converter/mcp_to_skill.py --allAfter conversion, Claude can invoke the MCP tools through the generated skill:
# List tools in a skill
python .claude/skills/mcp-skills/executor.py --skill github --list
# Get tool schema
python .claude/skills/mcp-skills/executor.py --skill github --describe create_issue
# Call a tool
python .claude/skills/mcp-skills/executor.py --skill github --call '{"tool": "create_issue", "arguments": {"title": "Bug fix", "body": "Details..."}}'The plugin automatically loads the MCP skills registry at the start of each Claude Code session. This provides:
| Scenario | Native MCP | Skills Plugin | Savings |
|---|---|---|---|
| Idle (no tools used) | 30-100k tokens | ~150 tokens | 99%+ |
| Using 1 skill | 30-100k tokens | ~5k tokens | 90%+ |
| Tool execution | 30-100k tokens | 0 tokens | 100% |
mcp-skills-plugin/
├── .claude-plugin/
│ └── plugin.json # Plugin manifest
├── skills/
│ └── mcp-to-skill-converter/
│ ├── SKILL.md # Skill documentation
│ ├── mcp_to_skill.py # Converter script
│ └── templates/
│ ├── index.json
│ └── registry-SKILL.md
├── hooks/
│ ├── hooks.json # Hook configuration
│ └── load-mcp-skills.sh # Session start hook
├── README.md
└── LICENSE| Type | Compatible | Notes |
|---|---|---|
stdio | ✅ Yes | Standard input/output protocol |
http | ✅ Yes | Streamable HTTP protocol |
sse | ✅ Yes | Server-Sent Events protocol |
⚠️ Important Security Considerations:
mcp-config.json files..mcp.json and writes to the skills directory. Ensure appropriate file permissions.pip install mcpSpecify the path explicitly:
python mcp_to_skill.py --mcp-json /path/to/.mcp.json --name githubRun --list to see available servers in your .mcp.json.
Ensure the hook script is executable:
chmod +x hooks/load-mcp-skills.shContributions are welcome! Please read the contributing guidelines before submitting PRs.
MIT License - see LICENSE for details.
Reduce your Claude Code context usage by 90%+ with MCP Skills.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.