create-skill

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

• Third-party content exposure detected (high risk: 0.75). In from-video mode, the workflow calls watch-video <url> visual, which fetches and converts user-supplied external video content into readable artifacts (transcript.txt, moments.md, summary.md) that are then synthesized into the agent’s LLM context—an outsider-authored free-text source.