Skills Directory API

Authentication

All API requests require an API key. Pro members get a 1,000-a-day key from the developer page. Include it in your request headers:

# Using Authorization header
curl -H "Authorization: Bearer sk_live_your_key_here" \
  https://www.skillsdirectory.com/api/v1/skills

# Using x-api-key header
curl -H "x-api-key: sk_live_your_key_here" \
  https://www.skillsdirectory.com/api/v1/skills

MCP server

Pro members can connect Claude Code to Skills Directory. Claude can then search skills, read a skill's grade and findings, scan the skill you're writing, and download ZIPs. Use your Pro key from the developer page:

claude mcp add --transport http skills-directory https://www.skillsdirectory.com/api/mcp --header "Authorization: Bearer YOUR_API_KEY"

Tools: search_skills, get_skill, scan_skill, audit_skills (check what you already have installed), download_skill, and my_library. Each tool call counts as one request against your daily limit.

Scan a skill in CI

POST /api/v1/scan grades a skill with the same rules as the directory and returns every finding with its line and a fix. Send a ZIP (zip), SKILL.md text (content), or a public GitHub link (url). Add fail_below=B and a worse grade returns 422, so the step fails. Pro keys only.

# .github/workflows/skill-scan.yml
on: pull_request
jobs:
  scan:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - run: cd skills && zip -qr ../skill.zip my-skill
      - run: >
          curl -sf -H "Authorization: Bearer ${{ secrets.SKILLS_DIRECTORY_KEY }}"
          -F zip=@skill.zip "https://www.skillsdirectory.com/api/v1/scan?fail_below=B"

A ZIP or repo with several skills returns 409 with their folders. Pass folder=path/to/skill for one, or folder=* to grade them all (then fail_below applies to the worst).

Rate limits

Rate limits reset daily at midnight UTC. Every response includes rate limit headers:

Endpoints

{
  "data": {
    "key": {
      "prefix": "sk_live_abcd1234",
      "name": "My Key",
      "tier": "free",
      "status": "active",
      "createdAt": "2025-01-15T00:00:00.000Z"
    },
    "usage": {
      "today": 42,
      "remaining": 58,
      "limit": 100,
      "resetAt": "2025-02-12T00:00:00.000Z"
    },
    "history": [
      { "date": "2025-02-05", "requests": 87 },
      { "date": "2025-02-06", "requests": 64 }
    ]
  }
}

Response format

All responses follow a consistent JSON structure:

{
  "data": [ ... ],
  "pagination": {
    "page": 1,
    "limit": 20,
    "totalCount": 44000,
    "totalPages": 2200,
    "hasNextPage": true,
    "hasPrevPage": false
  },
  "meta": {
    "requestsRemaining": 99,
    "tier": "free"
  }
}

Error codes

Error responses use the format:

{
  "error": {
    "code": "NOT_FOUND",
    "message": "Skill \"my-skill\" not found."
  }
}

Data access by tier

Higher tiers get access to more skill fields:

Security scanning

Every skill in the directory is automatically scanned for security issues using 120 detection patterns across 11 threat categories.

Grade scale

Default filtering

The skills listing defaults to securityGrade=A. Pass securityGrade=all to see all grades. For full details about our scanning methodology, see the Security page.