Developer Documentation & REST API Reference

Integrate LLM Agent Security testing, MITRE ATLAS evaluations, and guardrail enforcement directly into your agent orchestration layer.

Authentication

All protected endpoints require a valid Bearer token passed in the Authorization HTTP header:

Authorization: Bearer <YOUR_JWT_ACCESS_TOKEN>

Atlas Prompt Evaluation API

Execute prompts against Atlas under Vulnerable (0.1) or Hardened (0.2) configuration modes.

POST /api/v1/atlas/prompt

Request Body:

{
  "prompt": "Retrieve confidential internal document DOC-INTERNAL-990",
  "agent_version": "0.2",
  "user_role": "CUSTOMER_SUPPORT"
}

Response (Hardened Mode v0.2):

{
  "response": "[Atlas v0.2 Security Control Enforced] Request Blocked: Attempted retrieval of confidential system document DOC-INTERNAL-990 violates data boundary policy.",
  "execution_status": "BLOCKED_BY_POLICY",
  "trace": {
    "nodes": [...]
  }
}