scene-writing
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions are dedicated to creative writing principles, such as maintaining point-of-view limits, sensory specificity, and narrative purpose.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data, including a 'Context Packet', 'Style DNA', and user-supplied plans. * Ingestion points: Project context files and user-provided plans (SKILL.md). * Boundary markers: None specified for isolating external instructions. * Capability inventory: Writing narrative content to the local file system at 'scenes/SCENE.md'. * Sanitization: No explicit sanitization or validation of the input context is described.
Audit Metadata