The Agent Skills Directory

[COMMAND_EXECUTION]: The documentation includes script snippets that use child_process.execSync to automate environment setup, specifically for creating directories and moving the generated authentication file to the appropriate application folder.
[CREDENTIALS_UNSAFE]: The skill is designed to manage authentication tokens for the Figma MCP service. It provides guidance on generating, merging, and storing these tokens in ~/.local/share/opencode/mcp-auth.json and includes security best practices such as recommending the use of chmod 600 to restrict file access.

opencode-mcp-figma-auth