{
  "markdown": "# SkillProof MCP\n\nAsk whether a Claude Code skill actually works — **before** you install it.\n\nGitHub has tens of thousands of `SKILL.md` files. Almost none have been run by anyone but their author.\n[SkillProof](https://skillproof.dev) installs them from their repo, triggers them, and runs them on a real\ntask against a no-skill baseline. This MCP server puts those verdicts in your agent's hands.\n\n```\n> is there a tested skill for converting markdown to Confluence?\n\n## Confluence — DIDN'T PASS — scored below the no-skill baseline\nWhat our test found: Ran the bundled convert_markdown_to_wiki.py on a real sample doc: it silently\nturns **bold** text into wiki _italic_ (a genuine regex bug), and leaves standard GitHub-style tables\ncompletely unconverted, despite SKILL.md listing \"tables\" among the elements it handles.\n```\n\nThat is the whole point. A directory that only lists winners tells you nothing.\n\n## Tools\n\n| Tool | What it answers |\n| --- | --- |\n| `find_skill` | \"Is there a tested skill for *X*?\" — ranked matches with verdict, score, test notes, install command |\n| `check_skill` | \"Someone recommended *X* — is it any good?\" — the verdict for one skill by name, slug, or repo |\n\nEvery answer carries one of three verdicts:\n\n- **pass** — installed, triggered, and beat the no-skill baseline on a real task.\n- **setup** — works, but needs a manual step first (the notes say which).\n- **didn't pass** — scored below the no-skill baseline: it either couldn't run, or left you worse off\n  than not installing it.\n\nIf nothing has been tested for your job, the server says so instead of guessing. \"Not tested\" is a real answer.\n\n## Install\n\nClaude Code:\n\n```bash\nclaude mcp add skillproof -- npx -y skillproof-mcp\n```\n\nOr add it to your MCP config by hand:\n\n```json\n{\n  \"mcpServers\": {\n    \"skillproof\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"skillproof-mcp\"]\n    }\n  }\n}\n```\n\nWorks in any MCP client (Claude Code, Claude Desktop, Cursor, Windsurf, Zed). Node 18+. No API key, no account.\n\n## Where the data comes from\n\nThe server reads the live catalog at [`skillproof.dev/api/skills.json`](https://skillproof.dev/api/skills.json)\nand caches it for 15 minutes. Nothing is bundled, so verdicts are never stale. The scoring rubric —\ninstall /5, triggering /5, output-vs-baseline /10, docs /5 — is published at\n[skillproof.dev/methodology](https://skillproof.dev/methodology).\n\nCatalog data is CC BY 4.0: use it, cite `skillproof.dev`.\n\n## License\n\nMIT\n",
  "bytes": 2472,
  "sha": "62c0d47092307774aa86684ac0b31f9d606c6bcbb6e349b627a5c7a3886b6798",
  "repo_slug": "skillproofdev/skillproof-mcp",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_skillproofdev_skillproof_2720d106/readme"
}