API Reference

Scan a SKILL.md for security issues using the SISCO (Skill Security Scanner). Runs static analysis including YARA pattern detection, prompt injection checks, and code safety analysis.

Request Body

{
  "content": "---\nname: my-skill\ndescription: What it does and when to use it\n---\n# My Skill\nInstructions..."
}

Response

{
  "status": "PASSED",        // PASSED | WARNED | BLOCKED
  "findings": [
    {
      "rule_id": "prompt_injection_safety_override",
      "severity": "HIGH",
      "title": "Prompt injection detected",
      "description": "Skill contains instructions to ignore safety rules",
      "remediation": "Remove the instruction override",
      "line_number": 12,
      "file_path": "SKILL.md"
    }
  ],
  "findings_count": 1,
  "is_safe": false,
  "scan_duration_ms": 342.1,
  "usage": { "remaining": 4 }
}

Example

curl -X POST https://skillsea.droidtech.ai/api/skills/scan \
  -H "Authorization: Bearer sk_live_YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"content": "---\nname: my-skill\ndescription: Test skill\n---\n# Test"}'

Improve a SKILL.md using Claude AI. Analyzes the skill against Anthropic's official skill-creator best practices and returns an improved version with a diff of changes.

Request Body

{
  "content": "---\nname: my-skill\ndescription: Does things\n---\n# My Skill\nSome instructions"
}

Response

{
  "original": "---\nname: my-skill\n...",
  "improved": "---\nname: my-skill\ndescription: Comprehensive description of what the skill does and when to use it...\n---\n# My Skill\n...",
  "diff": [
    "- description: Does things",
    "+ description: Comprehensive description..."
  ],
  "changes": [
    "Expanded description to include trigger scenarios",
    "Added structured sections with examples"
  ],
  "issues_found": [
    "Description too vague - doesn't explain when to invoke",
    "Missing concrete examples"
  ],
  "usage": { "remaining": 4 }
}

Example

curl -X POST https://skillsea.droidtech.ai/api/skills/rewrite \
  -H "Authorization: Bearer sk_live_YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"content": "---\nname: my-skill\ndescription: Does things\n---\n# My Skill\nInstructions"}'

Create a new API token. Max 5 active tokens per user.

List your active tokens with usage counts.

Revoke a token. Cannot be undone.

Reads an active skill's SKILL.md, sends it to the rewriter API, and shows the diff.

Sends an active skill's SKILL.md to the scanner API and displays findings.