Files
TinyTorch/docs/development/AGENT_SETUP_PROMPT.md
Vijay Janapa Reddi 8cccf322b5 Add progressive demo system with repository reorganization
Implements comprehensive demo system showing AI capabilities unlocked by each module export:
- 8 progressive demos from tensor math to language generation
- Complete tito demo CLI integration with capability matrix
- Real AI demonstrations including XOR solving, computer vision, attention mechanisms
- Educational explanations connecting implementations to production ML systems

Repository reorganization:
- demos/ directory with all demo files and comprehensive README
- docs/ organized by category (development, nbgrader, user guides)
- scripts/ for utility and testing scripts
- Clean root directory with only essential files

Students can now run 'tito demo' after each module export to see their framework's
growing intelligence through hands-on demonstrations.
2025-09-18 17:36:32 -04:00

3.8 KiB

Universal Agent Setup Prompt for Any Project

Copy and paste this prompt to set up the same agent orchestration system in your other projects:


🤖 Multi-Agent Team Setup Request

I want to set up a professional multi-agent workflow system for this project. Please create a CLAUDE.md file in the project root that establishes the following:

1. Agent Team Structure

Create these specialized agents with clear responsibilities:

Workflow Coordinator (Team Lead)
    ├── Architect (Strategy & Design)
    ├── Developer (Implementation)
    ├── Quality Assurance (Testing & Validation)
    └── Documentation (Communication & Docs)

2. Mandatory QA Testing Protocol

  • QA Agent MUST test ALL code changes before ANY commit
  • QA has veto power - can block commits if tests fail
  • Developer CANNOT mark tasks complete without QA approval
  • Every code change triggers automatic QA review
  • Create comprehensive test suites that check:
    • Code imports without errors
    • Functions/classes work correctly
    • No syntax errors present
    • All tests pass successfully

3. Standard Workflow Pattern

Enforce this sequence for EVERY code update:

  1. Planning Phase (Coordinator + Architect)
    • Define objectives and structure
  2. Implementation Phase (Developer)
    • Write code following specifications
    • MUST notify QA when done
  3. Testing Phase (QA) - MANDATORY
    • Run comprehensive tests
    • Block progress if tests fail
  4. Documentation Phase (Documentation)
    • Add clear documentation
  5. Review Phase (Coordinator)
    • Verify all agents completed tasks
    • Ensure QA tests passed

4. Agent Communication Rules

  • Structured handoffs between agents
  • Each handoff includes checklist of completed/pending items
  • Test results must accompany all code changes
  • Clear accountability for each agent

5. Git Workflow Standards

  • Always use feature branches (never commit to main)
  • Test before committing
  • Descriptive commit messages
  • Clean commit history

6. Enforcement Mechanisms

  • File should be named CLAUDE.md in project root
  • Claude reads this automatically at session start
  • Rules override default behavior
  • QA testing cannot be skipped

7. Project-Specific Customization

Adapt the agents for this specific project:

  • What type of project is this? [Web app, ML system, API, etc.]
  • What testing frameworks should QA use?
  • What documentation standards apply?
  • Any special requirements?

Please create the CLAUDE.md file with these specifications, ensuring that:

  1. Every future Claude session will automatically follow these rules
  2. QA testing is mandatory and cannot be bypassed
  3. The agent team works as a cohesive unit
  4. Code quality is maintained through enforced testing gates

Also create any supporting test files or scripts needed for the QA agent to properly validate code changes.


📋 Quick Setup Instructions

  1. Copy the above prompt
  2. Paste it in your other Claude project
  3. Claude will create the CLAUDE.md file
  4. The system activates automatically in future sessions

🎯 Benefits You'll Get

  • Automatic QA testing before commits
  • Structured agent teamwork
  • Consistent code quality
  • Clear accountability
  • Professional development workflow
  • No broken code in your repository

💡 Pro Tips

  • Customize agent names for your project type
  • Add project-specific testing requirements
  • Include your preferred git workflow
  • Specify documentation standards
  • Add any unique project rules

🔄 Verification

After setup, test by asking Claude:

  • "What are the agent responsibilities?"
  • "Can you commit without QA testing?"
  • "Show me the workflow for updating code"

Claude should respond following the established protocols.