video-review
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill safely executes ffmpeg and ffprobe binaries for video analysis. Evidence:
scripts/review.pyinvokes these tools viasubprocess.runusing list-based arguments, avoiding shell injection. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata and filenames, exposing a surface for potential prompt injection. 1. Ingestion points:
ffprobemetadata from target videos and file listings from theartifacts/directory. 2. Boundary markers: The JSON output does not delimit untrusted strings or include instructions to ignore embedded commands. 3. Capability inventory: The skill can read from and write to the local filesystem and execute video processing commands. 4. Sanitization: Metadata and filenames are not sanitized before being returned to the agent.
Audit Metadata