game-dev-2d-save-progression
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions emphasize security best practices, such as preventing the execution of saved text as code and avoiding the instantiation of arbitrary classes from untrusted fields. The workflow is designed for data integrity, using temporary and backup paths to ensure write safety.
- [INDIRECT_PROMPT_INJECTION]: The skill provides logic for processing external, user-controllable files (game saves), which constitutes an ingestion surface for untrusted data. The skill mitigates this risk by establishing a strict persistence contract and a multi-step implementation pipeline.
- Ingestion points: Raw save data and storage adapters defined in
SKILL.mdand engine-specific references. - Boundary markers: The instructions mandate parsing and validation steps before data affects the runtime state.
- Capability inventory: File system write operations via platform adapters and optional cloud synchronization capabilities.
- Sanitization: Stepwise migration and normalization of defaults are required components of the implementation pipeline to ensure data validity.
Audit Metadata