Skip to content
spaget

Claude Code guide

Claude Code uses CLAUDE.md files to configure AI assistants. spaget makes it easy to build powerful Claude Code agents visually.

How Claude Code works

Section titled “How Claude Code works”

Claude Code reads a CLAUDE.md file in your project root (or in ~/.claude/ for global instructions). This file contains instructions, rules, and context that shape how Claude behaves in your project.

Building a Claude Code agent in spaget

Section titled “Building a Claude Code agent in spaget”
  1. Open the visual builder
  2. Add skills relevant to your project (e.g., “TypeScript conventions”, “Git commit style”)
  3. Set rules for code style, testing, and documentation
  4. Export as Claude Code format
  5. Save the downloaded CLAUDE.md to your project root

Tips for effective Claude Code agents

Section titled “Tips for effective Claude Code agents”

Be specific about your stack

Section titled “Be specific about your stack”

Instead of “use TypeScript”, try “use TypeScript with strict mode, prefer interface over type for object shapes, and use explicit return types on exported functions.”

Set coding conventions

Section titled “Set coding conventions”
  • File naming patterns (kebab-case, PascalCase)
  • Import ordering preferences
  • Error handling patterns
  • Testing expectations

Include project context

Section titled “Include project context”

Tell Claude about your project structure, key dependencies, and any patterns it should follow. The more context, the better the results.

Example configuration

Section titled “Example configuration”

A typical Claude Code agent might include:

  • Language & framework rules — TypeScript, React, Next.js conventions
  • Testing rules — “Write tests for all new functions using Vitest”
  • Git rules — “Use conventional commit messages”
  • Style rules — “Keep functions under 30 lines, prefer composition over inheritance”

Next steps

Section titled “Next steps”