cro

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly says "Use this even if the user just shares a URL and asks for feedback" and the included eval prompts contain external URLs (e.g., https://example.com/product), indicating the agent is expected to fetch and interpret arbitrary public web pages—which are untrusted third-party content that could influence its actions.