youtube-transcribe
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides shell commands for executing
yt-dlpandsed. It specifically suggests using the--cookies-from-browser chromeflag to access local browser cookies for authenticated sessions, which involves accessing sensitive user data to bypass age restrictions. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources (YouTube transcripts and video metadata) and processes it using shell commands, creating a potential attack surface.
- Ingestion points: Transcript segments from the YouTube DOM and video metadata/subtitles fetched via
yt-dlp(SKILL.md). - Boundary markers: None identified; external transcript content is processed as plain text without delimiters to differentiate it from instructions.
- Capability inventory: Subprocess execution for
yt-dlpandsed(SKILL.md). - Sanitization: None identified; the skill does not provide instructions for sanitizing video titles used in filenames or transcript content before processing.
Audit Metadata