youtube-transcript
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes transcripts fetched from YouTube, which constitutes untrusted external data. If a video's captions contain malicious instructions, the agent might inadvertently execute them while summarizing or formatting the text.
- Ingestion points: The output of the
scripts/get_transcript.pyscript is processed by the agent as described inSKILL.md. - Boundary markers: There are no explicit instructions for the agent to wrap the external transcript in delimiters or to ignore instructions embedded within the transcript text to prevent hijacking.
- Capability inventory: The skill executes local Python scripts using
uv runinSKILL.mdand allows the agent to write the resulting transcript to the local file system as described in 'Step 3: Save if requested'. - Sanitization: No sanitization or filtering logic is applied to the transcript content before it is presented to the agent's context.
- [COMMAND_EXECUTION]: The skill requires the agent to execute a local Python script using the
uvpackage manager to perform its primary function. - Evidence:
SKILL.mdinstructs the agent to runuv run .claude/skills/youtube-transcript/scripts/get_transcript.py "VIDEO_URL_OR_ID".
Audit Metadata