yt-dlp

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

• Third-party content exposure detected (medium risk: 0.30). Outsider text flows into yt-dlp probe <url>/transcript <url>/download <url> dispatch and is then ingested at runtime by scripts/probe_url.py via yt-dlp --dump-json --no-download <url>, which reads and returns captions/subtitles text from the referenced outsider-authored video/page.