grill-autopilot

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

• Third-party content exposure detected (low risk: 0.10). The skill orchestrates multi-agent design reviews by consulting external peer AI models (such as codex or grok) and performing web research, which can ingest outsider-authored web pages or package metadata.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill references mattpocock/skills and adonis0123/adonis-skills via npx skills add during setup, which constitute runtime external dependency locators.