Skip to content

Your First Task

Learn the complete Nova AI workflow by building a simple feature.

Time: 10 minutes

What You'll Learn

  • How to describe tasks in natural language
  • The question-answer workflow
  • How quality gates work
  • Reading the output

Step 1: Start a Task

Open your terminal in any project and run:

/novaai add a health check endpoint that returns server status

Step 2: Answer Questions

Nova AI asks clarifying questions to understand your requirements:

? What framework are you using?
  > FastAPI
  > Flask
  > Django
  > Express

? What should the endpoint return?
  > Simple "ok" status
  > Detailed system info (memory, uptime)
  > Database connection status

Answer based on your project needs.

Step 3: Review the Plan

Nova AI presents an implementation plan:

📋 Implementation Plan

1. Create health endpoint in routes/health.py
2. Add status response model
3. Write unit tests
4. Update API documentation

Agents: @implementer → @code-reviewer → @validator

Approve? [Run Now / Run in Background / Reject]

Press Enter to approve, or type feedback to refine.

Step 4: Watch It Work

Nova AI executes the plan:

⏳ @implementer creating routes/health.py...
✅ @implementer completed (3 files modified)

⏳ @code-reviewer checking security...
✅ @code-reviewer approved

⏳ @validator running tests...
✅ @validator passed (4/4 tests)

Ready to commit!

Step 5: Review Changes

Check what was created:

git diff --stat

Output: 

routes/health.py     | 25 +++++++++++++
models/health.py     | 12 +++++++
tests/test_health.py | 35 +++++++++++++++++++
3 files changed, 72 insertions(+)

Understanding the Output

Agent Status Icons

Icon Meaning
⏳ Agent working
✅ Agent completed successfully
❌ Agent found issues (will retry)
🔄 Agent retrying after fixes

Quality Gates

Every task passes through:

  1. Code Review - Security and correctness
  2. Validation - Tests must pass
  3. Linting - Code style checks

What's Next?