normalize

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions are focused on UI/UX design tasks and do not exhibit any malicious patterns or security vulnerabilities.
  • [DATA_EXFILTRATION]: The skill instructs the agent to search local project files for documentation, which is a standard activity for code auditing and does not involve external network requests or credential access.
  • [PROMPT_INJECTION]: No evidence of prompt injection, safety bypasses, or instructions to ignore system guidelines was found. The instructional language is professional and task-oriented.
  • [COMMAND_EXECUTION]: While the skill suggests using tools like grep for local analysis, it does not include dangerous shell commands, privilege escalation, or persistence mechanisms.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 11:37 PM
Security Audit — agent-trust-hub — normalize