research

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

• Third-party content exposure detected (high risk: 0.90). The skill's Phase 2 "Systematic Information Gathering" explicitly instructs using the gemini bash command or the WebSearch tool and to use the docs-seeker skill to read GitHub repository content, which fetches and ingests open/public third-party (user-generated) material that the agent must read and use to drive research decisions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs the agent at runtime to fetch and read GitHub repository content via the docs-seeker (i.e., external URLs such as https://github.com/) and inject those README/docs into the agent context to drive analysis, so external repo content can directly control prompts.