freeride

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

• Third-party content exposure detected (high risk: 0.80). The skill directly fetches untrusted model metadata from the public OpenRouter API (see main.py's fetch_all_models calling https://openrouter.ai/api/v1/models and the README/SKILL.md stating it "Finds the 30+ free models on OpenRouter"), then ranks and uses that external data to choose and configure primary/fallback models and drive automatic watcher rotations, so third-party content can materially influence tool behavior.