high-tension-shot-design
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is composed of markdown instructions and cinematic reference libraries. It contains no executable scripts, binaries, or commands that interact with the operating system or file system. All provided content is instructional text for AI prompt engineering.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied visual concepts to generate structured cinematic prompts. This creates a theoretical surface for indirect prompt injection where a user might attempt to embed instructions into the final prompt. However, the risk is negligible as the output is intended for creative generation in external tools and the skill does not possess any execution capabilities.
- Ingestion points: User-provided visual ideas and descriptions processed by the workflow in
SKILL.md. - Boundary markers: The skill uses rigid output templates (e.g., Tension Mechanism Menu) to structure the generated prompts, though it lacks explicit instructions to ignore embedded user commands within those inputs.
- Capability inventory: The skill is restricted to text generation; it has no access to network operations, file writing, or dynamic code execution.
- Sanitization: No sanitization of user-provided concepts is performed before they are integrated into the output templates.
Audit Metadata