youtube-transcript
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches transcripts from external YouTube videos, which represents an untrusted data source. This creates a surface for indirect prompt injection, where malicious instructions hidden in a video's captions could influence the AI agent's subsequent analysis or actions.
- Ingestion points: transcript.js (via the fetchTranscript method).
- Boundary markers: Absent. The output is printed as plain text with simple timestamp prefixes.
- Capability inventory: The script itself is limited to network reading (fetching transcripts) and printing to stdout. It does not perform file writes, subprocess execution, or other high-privilege operations.
- Sanitization: Absent. The raw text from the transcript is printed directly to the console.
Audit Metadata