unity-game-director
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell scripts provided in the package (
detect_unity_project.shandprobe_asset_credentials.sh) to detect project settings and verify environment configurations. These scripts use standard system utilities (find, grep, awk) and perform deterministic checks without network exfiltration. - [PROMPT_INJECTION]: The skill ingests data from local project files (e.g.,
ProjectVersion.txt,manifest.json,art-spec.yaml) to identify the Unity version and pipeline state. While this represents a surface for indirect prompt injection if these files contain untrusted content, the risk is minimal and typical for automated development tools. - [COMMAND_EXECUTION]: The
probe_asset_credentials.shscript sources user shell profiles (~/.bashrc,~/.zshrc, etc.) to verify if necessary API keys for asset generation (e.g., TRIPO_API_KEY, GEMINI_API_KEY) are set in the environment. The script only outputs the status (SET/MISSING) and does not print or transmit the actual secret values. - [SAFE]: The skill references well-known external services for game assets (Tripo, Gemini, PixelLab) and utilizes a vendor-specific MCP bridge (
com.coplaydev.unity-mcp) for Editor automation, which is consistent with its stated purpose.
Audit Metadata