unity-skills-index

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides the editor_execute_menu function, which allows the AI agent to execute arbitrary menu commands within the Unity Editor environment.
  • [REMOTE_CODE_EXECUTION]: The script_create, script_append, and script_replace modules enable the generation and modification of C# scripts. These scripts are subsequently compiled and executed by the Unity Editor's internal systems.
  • [EXTERNAL_DOWNLOADS]: The package_install skill facilitates the installation of Unity packages from the official Unity Package Manager registry.
  • [PROMPT_INJECTION]: The skill processes untrusted project data, creating a surface for indirect prompt injection.
  • Ingestion points: Ingests project source code and scene hierarchy through script_read, script_analyze, and scene_export_report.
  • Boundary markers: Ingested data is structured as JSON before being returned to the agent context.
  • Capability inventory: Possesses extensive file system access (asset_delete, script_replace) and editor command execution capabilities.
  • Sanitization: The skill descriptions do not specify sanitization or comment-stripping during the parsing of C# scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 11:50 AM