video-cut
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
subprocess.runto callffmpegandffprobefor media analysis and rendering. All calls use argument lists rather than shell strings, and no user-controlled data is passed to the shell environment, mitigating command injection risks. The usage is consistent with the skill's primary purpose of video editing. - [INDIRECT_PROMPT_INJECTION]: The skill processes a
clip_plan.jsonfile generated by the agent. It implements comprehensive validation, including JSON schema checks, type casting for numeric values (e.g., timestamps), and sanitization of file paths. The ingested data is used for calculation and metadata storage rather than being interpreted as executable instructions. - [DATA_EXPOSURE]: The skill operates entirely on local files within a specified workspace. It does not perform network requests and does not access sensitive system files or credentials.
Audit Metadata