video-content-analysis

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text can enter the LLM context via the user-supplied /<analyze-video <path>> video file’s extracted transcript/metadata (e.g., participant speech transcribed into summary.md/transcript.txt and then printed inline in Step 4), which is not authored by the operating user.