claude-design

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: A comprehensive analysis of the skill's instructions and metadata reveals no malicious patterns. The content is strictly focused on UI/UX design workflows, artifact generation, and adherence to design principles.
  • [PROMPT_INJECTION]: The skill contains instructions to disregard platform-specific tooling from hosted environments (e.g., 'ignore references to hosted-only tools'). These are functional requirements for adapting web-based behaviors to a CLI/API context and do not constitute an attempt to bypass agent safety guidelines.
  • [DATA_EXFILTRATION]: No suspicious network operations, credential harvesting, or sensitive file access patterns were identified. Instructions to read project-specific files (e.g., theme files, tokens, layouts) are within the scope of the stated design purpose and do not target sensitive system directories.
  • [REMOTE_CODE_EXECUTION]: No remote code execution patterns or unsafe script downloads were found. The skill provides best-practice advice for using external libraries (e.g., React from CDNs), such as pinning exact versions to ensure stability and safety.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 11:30 PM
Security Audit — agent-trust-hub — claude-design