stream-transcript-processor
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches auto-generated captions from well-known streaming services (YouTube and Twitch) using the yt-dlp utility.
- [COMMAND_EXECUTION]: Executes the yt-dlp command using subprocess.run with list-based arguments, ensuring safe parameter handling. It also references other tools like ffmpeg for media processing.
- [PROMPT_INJECTION]: The skill processes untrusted transcript data from external URLs, presenting a surface for indirect prompt injection.
- Ingestion points: Transcript data is fetched from external video URLs via fetch_transcript.py.
- Boundary markers: Absent; the skill does not use specific delimiters to isolate transcript content from agent commands.
- Capability inventory: The skill can execute subprocesses (yt-dlp, ffmpeg) and write files to the local directory.
- Sanitization: Basic removal of VTT markers and HTML-style tags is performed, but instructions in the text are not filtered.
Audit Metadata