recipe-cleanup

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

• Third-party content exposure detected (medium risk: 0.30). The skill ingests outsider-authored free text when the first argument is a http(s) URL or a YouTube watch URL, since it fetches the page/description via WebFetch/Playwright and extracts recipe content (and potentially linked external recipe pages) from that content at runtime.

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

• Potentially malicious external URL detected (high risk: 0.90). This skill fetches external webpages at runtime (e.g., youtube.com/watch?v=VIDEO_ID and https://i.ytimg.com/vi/VIDEO_ID/hq720.jpg) to extract recipe text and creator info which are then injected into the agent's processing, so fetched content directly controls the agent's prompts/output.