remix-script
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill relies on standard library Python scripts and does not involve external network calls or untrusted dependencies.
- [COMMAND_EXECUTION]: Scaffolding and validation are performed via local CLI scripts (
new_script.py,validate_script.py). - Evidence: Workflow steps 5 and 8 in
SKILL.md. - [PROMPT_INJECTION]: The skill processes untrusted input from brand research files and source samples, presenting a surface for indirect prompt injection.
- Ingestion points: Brand research documentation and
source-sample.jsonfiles. - Boundary markers: None. Content is ingested directly.
- Capability inventory: The skill writes ad scripts and storyboards used by downstream production tools.
- Sanitization: No escaping or validation is performed on ingested text before generation.
Audit Metadata