swift-private-bundle

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to read and search third‑party GitHub content (the local ~/GitHub/knowledge mirror and remote hocgin repositories via commands like gh repo view, gh search code, and gh api .../contents/Package.swift), and to base integration decisions and recommended actions on those README/Package.swift results, which clearly exposes it to untrusted, user-generated repository content that could influence behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill runs runtime git/gh commands to fetch content from github.com/hocgin/knowledge and target repos (e.g., gh api repos/hocgin/some-repo/contents/Package.swift and git@github.com:hocgin/some-repo.git), and the fetched files are ingested to directly inform the agent's responses, so these URLs are runtime dependencies that can control prompts.