create-product-stopmotion-video
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
references/example-build.shexecutesffmpegandffprobeto process image plates and assemble video files. It also contains hardcoded absolute paths pointing to the developer's local user directory (e.g.,/Users/shivsakhuja/...), which limits portability and exposes local environment metadata. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it incorporates user-supplied briefs into a workflow with shell execution capabilities without sanitization.
- Ingestion points: User briefs and asset paths defined in
SKILL.md. - Boundary markers: None present; there are no delimiters separating user input from the agent's instructions.
- Capability inventory: Shell command execution (
ffmpeg) and local file system writes. - Sanitization: No input validation or escaping mechanisms are described.
Audit Metadata