design-agent-customize
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- Command Execution: The skill instructs the user to use
uv runto execute various tasks, such as running tests withpytest, configuring the agent, and interacting with the agent's CLI. These commands are part of the intended development and management workflow. - External Service Integrations: The documentation describes how to connect the agent to external tools and data sources via Model Context Protocol (MCP) servers. It mentions common services like GitHub, Figma, Notion, and Linear as examples of available integrations.
- Dependency Management: The skill mentions "vendored skills" from external sources, specifically citing LangSmith, which is a service provided by the vendor (LangChain). It notes that these are managed via a lock file and should not be edited directly.
- Secret Management Guidance: The skill provides instructions for handling sensitive API keys (such as OpenAI, ElevenLabs, and Groq) using environment variables and configuration files like
.envandmcp.toml. This aligns with standard security practices for local tool configuration.
Audit Metadata