video-generation
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is composed of instructional markdown and YAML configuration files without any embedded scripts, binary executables, or network-capable tools. It does not access sensitive system paths or environment variables.
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a text-processing engine for user-supplied creative content, which presents a surface for indirect prompt injection.
- Ingestion points: The skill ingests untrusted data from users in the form of script beats, storyboard descriptions, and reference tags as defined in the SKILL.md workflow.
- Boundary markers: The skill requires the agent to encapsulate its output within a single code block and follow a rigid internal structure (SCENE CONTEXT through POSITIVE LOCKS), which serves as a structural delimiter.
- Capability inventory: The skill is restricted to text generation and has no access to subprocess execution, file system modification, or network requests.
- Sanitization: The instructions enforce strict formatting protocols, such as mandatory units for speed and atmosphere (km/h, %), and specific physical performance constraints (lips closed) which limit the potential for external content to hijack the agent's output logic.
Audit Metadata