ui-tree-architecture-review

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No instruction override or safety bypass patterns were detected. The skill uses standard professional language to guide architectural analysis.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: No hardcoded credentials, sensitive file paths, or unauthorized network operations were found. References to documentation target established and trusted domains.
  • [OBFUSCATION]: The content is clear and readable with no base64, hex, or unicode-based obfuscation.
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: No package installations or remote script executions are present. The skill restricts itself to reading and analyzing existing project files.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes project source code to analyze the UI tree. While this is an ingestion of untrusted data, the skill provides a structured template and guidelines that focus on static analysis rather than execution, and it lacks the high-privilege capabilities required for a successful exploit.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 10:27 AM
Security Audit — agent-trust-hub — ui-tree-architecture-review