audit-skill

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

• Third-party content exposure detected (medium risk: 0.65). The workflow resolves <skill-source> from either a local repository path or a GitHub fallback, and then runs npx skills add <skill-source> --list / npx skills add <skill-source> ... which will fetch and ingest outsider-authored SKILL.md content from the GitHub repository at runtime when the local repo is absent.

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

• Potentially malicious external URL detected (high risk: 0.90). The workflow runs npx (e.g., "npx skills add ") and specifies a GitHub fallback "onfloukimson/devkimson-skills", which at runtime would fetch and install remote package code from that repository (executing remote code), so this external repo is a runtime dependency controlling behavior.