{
  "markdown": "# decimalai-mcp\n\n*Part of [DecimalAI](https://decimal.ai). Most users want the Python SDK → [decimal-labs/decimalai-python](https://github.com/decimal-labs/decimalai-python).*\n\nMCP server for the [DecimalAI](https://decimal.ai) skills registry — the registry that ranks agent skills by **measured effectiveness** (verified A/B benchmarks, live pass rates, AI rater scores), not download counts.\n\n[![PyPI](https://img.shields.io/pypi/v/decimalai-mcp)](https://pypi.org/project/decimalai-mcp/)\n[![Downloads](https://static.pepy.tech/badge/decimalai-mcp/month)](https://pepy.tech/project/decimalai-mcp)\n[![CI](https://img.shields.io/github/actions/workflow/status/decimal-labs/decimalai-mcp/ci.yml?branch=main)](https://github.com/decimal-labs/decimalai-mcp/actions/workflows/ci.yml)\n[![Python](https://img.shields.io/pypi/pyversions/decimalai-mcp)](https://pypi.org/project/decimalai-mcp/)\n[![License](https://img.shields.io/badge/License-MIT-blue.svg)](https://github.com/decimal-labs/decimalai-mcp/blob/main/LICENSE)\n\n<!-- mcp-name: ai.decimal/registry -->\n\nGives any MCP client (Claude Desktop, Claude Code, Cursor, …) three read-only tools:\n\n| Tool | What it does |\n|---|---|\n| `search_skills(query, category?, sort?, limit?)` | Hybrid keyword/semantic search over the public registry |\n| `get_skill(slug)` | Full record: trust & safety-scan status, verified benchmark lift, SkillScore, ratings, SKILL.md body |\n| `get_leaderboard(sort?, category?, window_days?, limit?)` | Ranked leaderboard: `skill_score`, `biggest_improvement`, `efficiency`, `top_rated` |\n\n**No API key required** — all three tools read public registry endpoints. If you set `DECIMAL_API_KEY` (from [app.decimal.ai/settings](https://app.decimal.ai/settings)), the same tools additionally show which skills your org has already installed (`installed_as`).\n\n## Install\n\n```bash\npip install decimalai-mcp\n# or, no install needed at config time:\nuvx decimalai-mcp\n```\n\nRequires Python 3.10+.\n\n## Claude Code\n\n```bash\nclaude mcp add decimalai -- uvx decimalai-mcp\n# with an API key:\nclaude mcp add decimalai -e DECIMAL_API_KEY=dai_sk_... -- uvx decimalai-mcp\n```\n\n## Claude Desktop\n\nAdd to `claude_desktop_config.json` (Settings → Developer → Edit Config):\n\n```json\n{\n  \"mcpServers\": {\n    \"decimalai\": {\n      \"command\": \"uvx\",\n      \"args\": [\"decimalai-mcp\"],\n      \"env\": {\n        \"DECIMAL_API_KEY\": \"dai_sk_optional\"\n      }\n    }\n  }\n}\n```\n\nOmit the `env` block entirely for anonymous read-only access.\n\n## Configuration\n\n| Env var | Default | Purpose |\n|---|---|---|\n| `DECIMAL_API_KEY` | *(unset)* | Optional. Unlocks per-org enrichment (e.g. `installed_as`) on the same public endpoints. |\n| `DECIMAL_API_URL` | `https://api.decimal.ai` | Point at a self-hosted / local backend. |\n\n## Why no `check_manifest_impact` tool?\n\nThe manifest-impact endpoint (`POST /api/v1/regression-check`) is authenticated on the platform — it analyzes **your org's** production traces against a candidate manifest, so there is no public variant to expose. This server is deliberately a read-only, key-optional public-registry surface. If demand shows up, an authed `check_manifest_impact` (requiring `DECIMAL_API_KEY`) is a natural v0.2 addition; the [regression-check GitHub Action](https://github.com/decimal-labs/regression-check) covers the CI use-case today.\n\n## Endpoints used (all public)\n\n- `GET /api/v1/registry/skills` — browse/search\n- `GET /api/v1/registry/skills/{slug}` — detail\n- `GET /api/v1/registry/leaderboard` — ranked leaderboard (`category` filtering falls back to the browse endpoint's documented `view=ranks` mode, because the leaderboard endpoint is uncategorized)\n\n## Development\n\n```bash\npip install -e \".[dev]\"\npytest              # all HTTP mocked; no network\npython -m decimalai_mcp.server   # run over stdio\n```\n\nRun `pytest` yourself before opening a PR. CI also asserts that the pinned `mcp<2` still provides `FastMCP`, which the mocked tests do not cover — run that one too:\n\n```bash\npython -c \"from mcp.server.fastmcp import FastMCP; import decimalai_mcp.server\"\n```\n\nReleases are cut from a published GitHub Release — see\n[`RELEASING.md`](https://github.com/decimal-labs/decimalai-mcp/blob/main/RELEASING.md) for the gates a\nchange has to pass and the version strings that must move together.\n\n## License\n\nMIT\n",
  "bytes": 4304,
  "sha": "b22e596944c503d8742d78eb9f2e8f5c0b72461685c04b882432f27b180a3ac5",
  "repo_slug": "decimal-labs/decimalai-mcp",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_ai_decimal_registry_20cb6426/readme"
}