google-youtube
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were detected. The skill implements specific YouTube API tools using a restricted read-only scope (youtube.readonly) and adheres to standard implementation patterns.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external content from YouTube (video titles, descriptions, and channel metadata). While this creates an ingestion surface for potential indirect prompt injection, the skill lacks dangerous capabilities such as file system writing or command execution that could be exploited through this vector.
- Ingestion points: YouTube API responses in handler.ts (searchYouTube, getVideoDetails, getChannelInfo, getPlaylistItems).
- Boundary markers: Relies on the agent's internal prompt boundaries; none explicitly defined in the handler logic.
- Capability inventory: Limited to YouTube data retrieval and display; no subprocess calls, network POST requests to arbitrary domains, or file-write operations.
- Sanitization: Data is formatted as plain text list or JSON objects for agent consumption.
Audit Metadata