search-context

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

• Third-party content exposure detected (medium risk: 0.30). The workflow reads arbitrary GitHub repositories and performs code searches, which can ingest outsider-authored free text from untrusted external codebases.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill searches and clones arbitrary public GitHub repositories dynamically at runtime based on user queries, resulting in unverified third-party code access under weak/unknown provenance.