ue5-auto-assistant
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted natural language input from users to drive its routing logic and tool recommendations. This creates a surface for indirect prompt injection where malicious instructions embedded in user queries could influence the selection of sensitive tools.
- Ingestion points: User queries describing Unreal Engine tasks are processed to determine intent and route to specific skills (found in
SKILL.mdandreferences/beginner-smoke-prompts.md). - Boundary markers: The instructions do not mandate the use of delimiters or specific warnings to ignore embedded instructions when parsing user input.
- Capability inventory: The skill references and recommends tools with high-impact capabilities, such as
run_console_command,execute_script,spawn_actor, andset_propertyacross various modules (references/mcp-skill-mapping.md). - Sanitization: There is no mention of sanitizing or validating user input before it is used to generate the routing payload or recommended tool list.
Audit Metadata