youtube
Warn
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: MEDIUMPRIVILEGE_ESCALATIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill instructs the agent to use 'sudo' to install system packages ('sudo apt install -y yt-dlp'), which grants elevated permissions to the process and presents a privilege escalation risk.
- [DYNAMIC_EXECUTION]: The skill utilizes inline Python execution ('python3 -c') to dynamically process and deduplicate transcript text at runtime.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from YouTube transcripts and audio, which could contain malicious instructions.
- Ingestion points: Transcript files and audio data fetched from user-provided YouTube URLs.
- Boundary markers: Absent; the final transcript is saved as plain text without delimiters or warnings to indicate the content is untrusted.
- Capability inventory: The skill uses 'Bash' to execute package managers, downloader tools, and file operations, alongside 'Write' for file creation.
- Sanitization: Basic regex is used to remove subtitle tags, but the content is not sanitized for potential embedded instructions that could influence the agent.
Audit Metadata