youtube-transcript

Warn

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill workflow and setup instructions involve managing system network interfaces using commands such as 'wg-quick' and 'ip rule'. These operations typically require elevated privileges (sudo) and modify the host system's routing table and network configuration.
  • [PROMPT_INJECTION]: The skill processes untrusted YouTube transcripts fetched from the internet, creating a surface for indirect prompt injection. Malicious instructions embedded in video captions could potentially influence the agent's summarization behavior or subsequent actions. Ingestion points: Transcript text fetched in SKILL.md; Boundary markers: Absent; Capability inventory: Shell command execution for script and VPN management; Sanitization: None mentioned.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 16, 2026, 11:09 AM
Security Audit — agent-trust-hub — youtube-transcript