marketing-community

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs the agent to audit and use live community URLs (e.g., "Live community / URL audit mode" and the Response Protocol: "proceed from the repo, live community URL...") which requires fetching and interpreting untrusted, user-generated content from public forums, Discord, Reddit, websites, etc., and using it to drive recommendations and actions.