github-automation

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

• Third-party content exposure detected (medium risk: 0.30). The required GitHub workflows involve reading and managing repository issues and pull requests, which expose the agent to outsider-authored free text via issue bodies, comments, and PR descriptions.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill references https://rube.app/mcp as an MCP endpoint for server configuration, but the URL points to an unverified third-party service lacking strong or official provenance for the dependency.