unreal-mod-frameworks
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions the agent to resolve game builds, generated SDKs, and mapping files to implement mod hooks and assets. This ingestion of untrusted external data represents a potential attack surface.
- Ingestion points: Instructions in
references/OVERVIEW.mdand game-specific reference files direct the agent to read installation layouts, SDKs, and external community repositories (e.g., UE4SS, PalSchema). - Boundary markers: The skill establishes "Non-negotiable laws" and mandatory version snapshotting to constrain agent behavior and prevent the invention of API names or hooks, which serves as a manual boundary.
- Capability inventory: The skill is designed for development tasks that involve file modification and interaction with game frameworks.
- Sanitization: There are no explicit mechanisms provided for sanitizing or filtering instructions that might be embedded in game assets or community-provided documentation.
- [SAFE]: The skill provides detailed references to established community documentation and official repositories for modding frameworks. The focus on version control snapshots (
1.0.0,1.0.1, etc.) and rollback capabilities demonstrates a best practice for project stability.
Audit Metadata