unity-behavior

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes Unity Behavior graph assets and blackboard variables, which serves as an indirect prompt injection surface.
  • Ingestion points: Data is read from external .asset files and runtime objects via tools defined in SKILL.md (e.g., behavior_graph_info, behavior_agent_get, behavior_blackboard_list).
  • Boundary markers: Absent; the instructions do not define delimiters or provide specific warnings to the agent to ignore instructions embedded in the graph data.
  • Capability inventory: SKILL.md defines capabilities to create assets (behavior_graph_create), modify variables (behavior_blackboard_set), and attach components (behavior_agent_add).
  • Sanitization: Absent; no sanitization of ingested asset or blackboard data is mentioned.
  • [EXTERNAL_DOWNLOADS]: The skill includes instructions to install the com.unity.behavior package from Unity using the package_install tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 08:54 AM
Security Audit — agent-trust-hub — unity-behavior