muapi-youtube-shorts
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local script located at library/social/youtube-shorts/scripts/run-youtube-shorts.sh to perform API calls and file processing.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external video content which is transcribed server-side into natural language metadata such as hooks and titles. This creates a surface where instructions hidden in the audio could be processed by the agent.
- [INGESTION_POINTS]: External URLs or local video files provided as source inputs.
- [CAPABILITY_INVENTORY]: Shell script execution and interaction with the muapi-cli tool.
- [BOUNDARY_MARKERS]: No explicit delimiters are defined for the generated transcript data.
- [SANITIZATION]: The instructions do not detail any sanitization for transcribed text before it is returned to the agent.
Audit Metadata