unity-editor-tools
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs shell command execution using the
unity-clitool to inspect and maintain the Unity Editor state. - [DYNAMIC_EXECUTION]: According to
references/runtime-checklist.md, the agent is instructed to usecargo runto compile and execute theunity-clibinary from the local workspace source if no global binary is installed. - [INDIRECT_PROMPT_INJECTION]: Ingestion points: The skill reads Unity console logs via the
unity-cli raw read_consolecommand (SKILL.md). Boundary markers: There are no delimiters or instructions to treat console output as untrusted, which could lead to instructions in logs being processed as agent commands. Capability inventory: The agent has the ability to modify project settings (update_project_settings) and manage packages (package_manager), providing a vector for exploitation if logs contain malicious instructions. Sanitization: No sanitization or filtering is applied to the ingested console logs.
Audit Metadata