build-game-monster-system
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill architecture involves processing external data such as monster definitions and asset metadata, creating a potential surface for indirect prompt injection if source data is untrusted.
- Ingestion points: The agent processes external MonsterDefinition objects and asset sources (GLB, hybrid) as defined in references/monster-contract.md.
- Boundary markers: The instructions lack explicit delimiters or warnings to ignore embedded instructions within processed data.
- Capability inventory: The skill utilizes an in-app browser for asset rendering and generates runtime adapters for Three.js.
- Sanitization: No explicit content sanitization or validation steps for external strings or metadata are provided in the instructions.
Audit Metadata