ux-evaluator
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- PROMPT_INJECTION (SAFE): No instructions to bypass safety filters or override system prompts were found. The language is purely instructional for UX evaluation.
- DATA_EXFILTRATION (SAFE): No commands for accessing sensitive files (e.g., .ssh, .aws) or making network requests (e.g., curl, wget) are present.
- REMOTE_CODE_EXECUTION (SAFE): No remote downloads or execution patterns (e.g., curl | bash) were detected.
- OBFUSCATION (SAFE): The content is clear markdown. No Base64, zero-width characters, or encoded strings were found.
- PRIVILEGE_ESCALATION (SAFE): No use of sudo, chmod, or administrative commands.
- INDIRECT_PROMPT_INJECTION (LOW): While the skill processes user-provided UI descriptions and external references, it lacks executable capabilities (file write, network, shell access), limiting the risk to purely informational context influence.
Audit Metadata