3d-animation-short-generator
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user-supplied story ideas and premises in Step 0, which are subsequently used to drive the generation of character cards, scene cards, and video clips. This workflow creates a surface where malicious instructions embedded in the story description could potentially influence the behavior of the image and video generation tools.
- Ingestion points: User-provided story idea and production format preferences in SKILL.md Step 0.
- Boundary markers: The skill employs structured canvas artifacts such as the Project Brief, Story Outline, and a six-column Shot Information Table, along with specific double-binding labels (e.g., [char:name]) to organize and isolate data.
- Capability inventory: The skill utilizes the hub_generate_image and hub_generate_video tools as specified in references/shot-table-spec.md and references/model-selection.md.
- Sanitization: The skill instructions do not specify any sanitization, filtering, or explicit 'ignore instructions' warnings for the raw user input before it is interpolated into the production workflow.
Audit Metadata