ai-friendly-testing

AI-Friendly Testing Patterns

Bun Test with AI Integration

Always use AI-friendly test output when working with AI assistants:

# Dedicated AI command (preferred)
bun run test:ai

# Environment variable approach
CLAUDECODE=1 bun test
CLAUDECODE=1 bun test tests/unit
CLAUDECODE=1 bun test --coverage

Auto-Detection

CI scripts automatically detect AI environments:

• CLAUDECODE=1 - Claude Code
• REPL_ID=1 - Replit AI
• AGENT=1 - Generic AI agent

Commands that auto-enable quiet mode:

bun run ci:local    # ✅ Auto-detects AI environment
bun run ci:bun      # ✅ Auto-detects AI environment

Benefits

• ✅ Shows only test failures (not passing tests)
• ✅ Reduces output verbosity for AI context windows
• ✅ Preserves failure details and summaries
• ✅ Improves readability in AI sessions

Test Mock Patterns

For D1 database mocks, handle both regular and cost cap queries:

mockEnv = {
  ANALYTICS: {
    prepare: (query: string) => {
      // Handle cost cap queries (dbstat)
      if (query.includes('dbstat') || query.includes('SUM(pgsize)')) {
        return {
          first: async () => ({ size: 1000000, rows: 1000 }),
          bind: () => ({ first: async () => ({ size: 1000000, rows: 1000 }) }),
          run: async () => ({ success: true }),
          all: async () => []
        };
      }
      
      // Handle regular queries
      return {
        bind: (...params: any[]) => ({
          first: async () => ({ /* your data */ })
        }),
        first: async () => null,
        run: async () => ({ success: true }),
        all: async () => []
      };
    }
  }
};

Environment Variables

Set any of these to enable quiet mode:

Variable	Use Case
CLAUDECODE=1	Claude Code sessions
REPL_ID=1	Replit AI environments
AGENT=1	Generic AI agent flag

Package Scripts

Available in package.json:

{
  "scripts": {
    "test:ai": "CLAUDECODE=1 bun test",
    "test": "bun test",
    "test:unit": "bun test tests/unit",
    "test:integration": "bun test tests/integration"
  }
}