video-editing

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

• Third-party content exposure detected (high risk: 0.70). The required workflow here is the transcript-based editing flow that delegates filler-word removal to the ai-video-tools agent, which (by design) analyzes transcript text and proposes edits; transcript content is outsider-authored (recordings/transcripts come from other people), so that free-form text can be fed into the agent’s LLM context via the agent delegation path.