openrouter-provider-ranking

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The runtime path scripts/probe_endpoints.py reads user-supplied prompt text from --prompt-file (arbitrary local file content) and sends it to OpenRouter for candidate verification.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill makes runtime HTTP calls to OpenRouter (e.g., the Endpoints API and chat completions endpoint) — "https://openrouter.ai/api/v1/models/" and "https://openrouter.ai/api/v1/chat/completions" — and uses the returned endpoint/response data to build provider routing and request fragments that directly control runtime routing/requests.