design

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

• Third-party content exposure detected (high risk: 0.90). This skill's SKILL.md explicitly instructs the agent to fetch and read external sources (e.g., "When the user provides a repository URL or pastes source code..." and "When only a URL is provided: fetching returns extracted text only"), so untrusted public web pages and repos are ingested and used to drive decisions and code generation, creating a clear avenue for indirect prompt injection.