llm-router

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

• Third-party content exposure detected (medium risk: 0.30). In scripts/llm_router_client.py, the CLI/user-provided --message/--messages/--system/--prompt and image_url are inserted directly into the LLM request payload and sent to https://api.aisa.one/v1/chat/completions (or /v1/models/{model}:generateContent), so outsider-authored free text is ingested at runtime.