82beaed5c6
More touchpoints to convert users into stargazers: the `list` command footer and the MCP server's stderr startup banner (stderr is safe — it never corrupts the JSON-RPC stream on stdout). Claude-Session: https://claude.ai/code/session_016JWn5jRD5tcEFKrubjQ6Px Co-authored-by: Claude <noreply@anthropic.com>
MCP Server
A zero-dependency Model Context Protocol server that exposes this skill library to any MCP client (Claude Desktop, Cline, etc.). Instead of installing 172 files, your assistant can search and pull skills on demand.
Tools
| Tool | What it does |
|---|---|
list_skills |
List every skill (name, tier, one-line description). Optional tier filter. |
search_skills |
Keyword search across name, description, and body — returns the best matches. |
get_skill |
Return the full instructions for one skill by name, ready to apply. |
Configure it
Claude Desktop — add to claude_desktop_config.json:
{
"mcpServers": {
"pm-claude-skills": {
"command": "npx",
"args": ["-y", "pm-claude-skills-mcp"]
}
}
}
From a local clone (no npm install):
{
"mcpServers": {
"pm-claude-skills": {
"command": "node",
"args": ["/absolute/path/to/pm-claude-skills/mcp/server.mjs"]
}
}
}
Restart the client. Then ask it to "search the skills for customer churn" or "get the rice-prioritisation skill and apply it to my backlog" — it calls the tools automatically.
How it works
Pure Node standard library, MCP stdio transport (newline-delimited JSON-RPC 2.0). It reads the bundled skills/ at startup and serves them in-memory; all logging goes to stderr so it never corrupts the protocol stream. No network, no data leaves your machine.