youtube-transcript

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly fetches transcripts and metadata from public YouTube URLs (see SKILL.md "Run the script to fetch the transcript" and scripts/fetch_transcript.py calling YouTube oembed and youtube_transcript_api) and instructs the agent to read and use that untrusted, user-provided content for summarization/decision-making, exposing it to indirect prompt injection risk.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill fetches external YouTube content at runtime (e.g., the video URL https://www.youtube.com/watch?v=XXXXXXXXXXX and the metadata endpoint https://www.youtube.com/oembed?url=... ) and injects the retrieved transcript text verbatim into model context for downstream summarization/QA, meaning remote content can directly influence prompts/instructions.