github

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly reads public GitHub content — issues, pull requests, repository files and code via tools like github_get_issue, github_get_pull, github_get_file, and github_search_code — which is untrusted, user-generated third-party content the agent is instructed to read and that can influence its decisions and actions.