Documentation

API Endpoint

POST /api/scan

Submit raw SKILL.md content to receive an instant security analysis.

Request Body

{
  "content": "string (max 100,000 chars)"
}

Response

{
  "score": 0-100,
  "risk": "low" | "medium" | "high" | "critical",
  "findings": [
    {
      "severity": "critical" | "high" | "medium" | "low",
      "type": "prompt_injection" | "data_exfiltration" | ...,
      "detail": "..."
    }
  ]
}

Detection Categories

  • • Prompt Injection — role override, jailbreak, DAN mode
  • • Data Exfiltration — suspicious external network calls
  • • Base64 Obfuscation — encoded payloads
  • • Hardcoded Secrets — API keys, tokens, credentials
  • • Code Execution — eval, exec, subprocess, shell
  • • File System Access — .env, /etc/passwd, SSH keys