independent-review

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

• Third-party content exposure detected (medium risk: 0.30). The workflow consumes an external email body (or similar external submission) as free text into a monitored queue/feed.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill executes independent_review.sh, which contains remote curl commands to fetch installation scripts and installers from external domains (chatgpt.com, antigravity.google, ollama.com), constituting a weak/unknown/suspicious operational dependency chain.