unreal-mcp

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill implements a feature to load 'Agent Skills'—custom instruction sets stored within an Unreal project—using the AgentSkillToolset.GetSkills tool.\n
  • Ingestion points: External instructions are retrieved from the project environment via AgentSkillToolset (SKILL.md).\n
  • Boundary markers: Project-provided instructions are given precedence over the agent's default guidelines.\n
  • Capability inventory: The toolset allows for mutation of live editor state, asset modification, and triggering of system build processes.\n
  • Sanitization: The skill does not describe specific validation or escaping mechanisms for the instructions loaded from the project.\n- [COMMAND_EXECUTION]: The skill facilitates the execution of system-level compilation tools through the LiveCodingToolset.CompileLiveCoding interface to rebuild C++ code (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 07:55 AM
Security Audit — agent-trust-hub — unreal-mcp