design-exploration

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill utilize npx to download and run the @agent-native/core package from the NPM registry for connecting to the hosted design service.
  • [REMOTE_CODE_EXECUTION]: The skill provides instructions for executing the @agent-native/core package to establish a connection between the local environment and the Design MCP application at https://design.agent-native.com/mcp.
  • [COMMAND_EXECUTION]: Terminal commands such as npx @agent-native/core@latest connect and npx -y @agent-native/core@latest reconnect are included for session management and authentication.
  • [CREDENTIALS_UNSAFE]: The skill implements best practices for secret management, specifically instructing the agent to avoid storing shared secrets in files and to redact Authorization and token values when inspecting configurations.
  • [PROMPT_INJECTION]: The skill ingests untrusted design briefs and existing HTML/design system data; however, it defines strict structural and stylistic constraints for the output, reducing the risk of accidental instruction obedience.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 06:15 PM
Security Audit — agent-trust-hub — design-exploration