gh-code-search

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

• Third-party content exposure detected (high risk: 0.90). The skill performs live searches and fetches code from public GitHub repositories (scripts/github.ts: searchGitHubCode and fetchCodeFiles fetch repo files and content) and then has the agent (Claude) analyze those fetched, user-authored GitHub files and snippets as part of its orchestration and report generation, exposing it to untrusted third-party content.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill uses Octokit/gh at runtime to fetch raw GitHub files and include their contents and blob URLs (e.g., https://github.com///blob/) in the generated markdown that is fed to Claude, which means externally hosted GitHub content is fetched during runtime and injected into the agent prompt context.