This PR adds comprehensive instruction files for different AI assistants (Claude, Gemini, GitHub Copilot, and others) working on the Onchain Test Kit project. These files provide tailored guidance that leverages each AI assistant's unique strengths while maintaining consistency in technical standards and development practices.

Overview

The instruction files help AI assistants understand the project's architecture, coding patterns, testing requirements, and development workflow. Each file is specifically tailored to different AI assistants' capabilities while maintaining consistent technical guidance about the Onchain Test Kit's builder patterns, wallet automation, and Playwright integration.

Files Added

Root Level Instructions

• AGENTS.md - Universal guidance for all AI assistants with comprehensive project overview, architecture details, and development best practices
• CLAUDE.md - Claude-specific instructions emphasizing analytical approach, systematic problem-solving, and thorough code analysis
• GEMINI.md - Gemini-focused instructions leveraging multimodal capabilities, creative problem-solving, and holistic system understanding

Directory Structure

• .github/instructions/ - New directory containing assistant-specific instruction files:
  • claude.instructions.md - Concise Claude development guidelines
  • gemini.instructions.md - Focused Gemini development instructions
  • general.instructions.md - Universal AI assistant guidelines
  • README.md - Directory documentation

Key Features

Consistent Technical Foundation

All instruction files maintain consistency with the existing .github/copilot-instructions.md and reference the same:

• TypeScript and Playwright patterns
• Builder pattern for configuration APIs
• Wallet automation architecture
• Testing strategies and fork mode functionality

Assistant-Specific Guidance

• Claude: Emphasizes systematic analysis, architectural understanding, and comprehensive testing
• Gemini: Leverages multimodal processing, pattern recognition, and creative problem-solving
• General: Provides universal guidelines applicable to any AI assistant

Practical Examples

Each file includes relevant code examples demonstrating:

// Configuration pattern
const config = configure()
  .withMetaMask()
  .withSeedPhrase({ seedPhrase: '...', password: '...' })
  .withNetwork({ name: '...', rpcUrl: '...', chainId: 1 })
  .build();

// Testing pattern
test('wallet interaction', async ({ page, metamask }) => {
  await metamask.handleAction('connect', { shouldApprove: true });
});

Benefits

1. Improved Code Quality: AI assistants will better understand project conventions and architectural patterns
2. Consistent Development: All assistants follow the same technical standards while leveraging their unique strengths
3. Better Testing: Clear guidance on testing requirements, including edge cases and error handling
4. Enhanced Documentation: Instructions emphasize keeping documentation up-to-date with code changes
5. Future Maintenance: Structured approach for adding support for new AI assistants

No Breaking Changes

This PR only adds documentation files and does not modify any existing functionality:

• All existing tests pass
• Project builds successfully
• No changes to public APIs
• No impact on runtime behavior

The instruction files will help AI assistants contribute more effectively to the Onchain Test Kit while maintaining the high standards of reliability and type safety that make this toolkit valuable for blockchain application testing.

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.