game-build
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use AppleScript to automate browser window and tab management (e.g., 'AppleScript over Chrome windows/tabs matching the EXACT dev URL').
- [COMMAND_EXECUTION]: Employs
javascript_toolandbrowser_batchfor runtime tuning and capturing visual proofs within the browser environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of game design documents (GDD).
- Ingestion points: Reads
DESIGN_READYGDD files andVISUAL_READYtargets from the local filesystem. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the GDD are present.
- Capability inventory: The skill possesses the ability to execute code via
javascript_tool, control system applications via AppleScript, and write files to the local repository. - Sanitization: The instructions do not specify validation or sanitization of the content extracted from the GDD before use in logic or generation steps.
Audit Metadata