youtube-claims
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Python scripts (
fetch.py,doctor.py,cache.py,segments.py) to handle data processing, transcript slicing, and caching. These operations are limited to the skill's specific logic. - [EXTERNAL_DOWNLOADS]: Fetches video metadata, transcripts, and thumbnail images from YouTube and Google's infrastructure using well-known libraries like
yt-dlpandyoutube-transcript-api. - [PROMPT_INJECTION]: The skill processes untrusted external data in the form of YouTube transcripts, creating an indirect prompt injection surface.
- Ingestion points: External transcripts are ingested through
scripts/fetch.pyand passed to the LLM. - Boundary markers: Prompts in the
prompts/directory use structured output requirements (JSON/Markdown) and explicit framing to constrain the model's behavior. - Capability inventory: The skill has the ability to make network requests (via
yt-dlp) and write files to the~/youtube-reports/directory. - Sanitization: YouTube video IDs are strictly validated using regex in
scripts/fetch.pybefore being used in commands.
Audit Metadata