04-script-video
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions in Vietnamese designed for content creation workflows.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by reading local context files (
.agents/product-marketing-context.md,.agents/personal-brand-context.md) and analyzing user-provided video links. - Ingestion points: Context files and user-provided reference links in the 'Reverse-engineer' section.
- Boundary markers: Not explicitly defined in the instructions.
- Capability inventory: None. The skill does not use any shell tools, network utilities, or file-writing capabilities; it only generates text output.
- Sanitization: No sanitization is performed on ingested data, but the lack of dangerous capabilities mitigates the risk.
Audit Metadata