baoyu-youtube-transcript

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

• Third-party content exposure detected (high risk: 0.80). The skill fetches public, user-generated YouTube content (transcripts and video descriptions via scripts/youtube.ts: fetchHtml/fetchTranscriptSnippets and yt-dlp fallback) and then explicitly passes the raw transcript and description into an AI sub-agent for speaker identification per SKILL.md and prompts/speaker-transcript.md, so untrusted third-party text is read and can materially change downstream processing/outputs.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill runs yt-dlp at runtime with the --remote-components "ejs:github" argument (used in fetchYtDlpInfo), which causes yt-dlp to fetch and execute remote components from GitHub during the fallback path, enabling remote code to be pulled and run.