youtube-clip-extractor
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
yt-dlpandffmpegto process video files. The commands involve shell interpolation of variables such asYOUTUBE_URLand{video_id}. If the agent does not properly sanitize these inputs, it could lead to command injection attacks. - [EXTERNAL_DOWNLOADS]: The skill performs network requests to YouTube servers via
yt-dlpto fetch video and subtitle files. - [INDIRECT_PROMPT_INJECTION]: The skill analyzes transcript files (
.vtt) in Phase 2 and Phase 4. This introduces a surface for indirect prompt injection where malicious instructions embedded in the video's subtitles could attempt to influence the agent's analysis or output generation. - Ingestion points: Transcript files located at
transcripts/{video_id}.en.vttare read during the analysis phase. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present when processing the transcript content.
- Capability inventory: The skill utilizes
yt-dlpfor network/file operations andffmpegfor file processing. - Sanitization: There is no evidence of transcript content sanitization before it is processed by the agent.
Audit Metadata