character-portrait-scene-planner
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is limited to scene planning and does not contain any executable scripts, network requests, or credential handling. It strictly enforces a 'planning-only' boundary, ensuring that both runtime and artifact statuses remain in 'Not Run' and 'Not Produced' states respectively.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input which could theoretically contain instructions intended to influence the agent's behavior during the planning phase.
- Ingestion points: Inputs include 'Subject identity', 'Reference images', and 'Wardrobe / prop list' as specified in the input schema in
SKILL.md. - Boundary markers: Absent; there are no specific delimiters or warnings to ignore instructions embedded within the user-provided character details.
- Capability inventory: The skill lacks capabilities for command execution, file system modification, or network access, restricting its operations to text-based planning and coordination with other planning skills.
- Sanitization: No sanitization or validation of the content within the user-provided inputs is described.
Audit Metadata