diagnose

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

• Third-party content exposure detected (medium risk: 0.30). The skill fetches an external knowledge base and executes an external script from GitHub via curl without user selection of specific items, allowing potential third-party content injection.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill fetches and executes an arbitrary remote audit script and pulls instructions from an unofficial GitHub repository at runtime, presenting a weak/unknown third-party operational dependency.