reels
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill makes extensive use of system utilities including
ffmpeg,ffprobe, andwhisper-clito perform complex video editing, audio cross-correlation, and transcription tasks. Technical analysis shows that thescripts/make_reel.pyscript executes these utilities usingsubprocess.runwith argument lists (shell=False), which is a secure practice to prevent shell injection. - [EXTERNAL_DOWNLOADS]: The skill uses
yt-dlpto download video content and subtitles from YouTube. It also facilitates the download of Whisper models for local transcription via thewhisper-clitool. These operations are aligned with the skill's primary purpose of video repurposing. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (YouTube transcripts and video content) which could potentially contain adversarial instructions.
- Ingestion points: YouTube URLs and transcripts fetched via Apify or yt-dlp.
- Boundary markers: The skill does not explicitly define markers to segregate untrusted transcript data from system instructions.
- Capability inventory: The skill possesses the ability to write to the local filesystem (reels.json, repurpose-plan.md) and execute media processing commands.
- Sanitization: The rendering engine uses strict JSON schemas for plans and handles subprocess arguments as list items, effectively neutralizing typical command injection vectors from processed data.
Audit Metadata