Skip to content

Quick Start

Get your first AI-powered code review or refactoring done in 5 minutes.

1. Choose an Agent

Select the right agent for your task:

Task Agent
Develop a Python service python-expert
Plan architecture lead-architect
Perform a code review code-reviewer
Write tests test-strategist
Create a CI/CD pipeline devops-agent
Build a frontend frontend-expert

2. Open Chat and Invoke the Agent

Open Copilot Chat and formulate your request:

Example: Refactor a Python Service

@workspace Use the Python Expert agent.
Refactor the UserService in src/services/user_service.py.
Apply the Repository Pattern and add Dependency Injection.

Example: Architecture Review

@workspace Use the Architecture Reviewer agent.
Review the current microservice architecture and identify
improvement opportunities regarding coupling and cohesion.

Example: Create a Test Strategy

@workspace Use the Test Strategist agent.
Create a test strategy for the Payment Service.
Include Unit Tests, Integration Tests, and Contract Tests.

3. Add Skills

For deeper expertise, you can add skills:

@workspace Use the Python Expert with the following skills:
- Clean Code
- SOLID Principles
- Design Patterns (Repository, Factory)

Refactor the database access layer in src/repositories/.

4. Work Iteratively

Agents work best in dialogue. Refine the results:

graph LR
    A[Submit request] --> B[Review agent response]
    B --> C{Satisfied?}
    C -->|Yes| D[Accept code]
    C -->|No| E[Give feedback]
    E --> B

Best Practice

Give the agent specific feedback: "Change the error handling to use Custom Exceptions" instead of "Make it better".

5. Validate Results

  • Review code changes in the diff
  • Run tests
  • Check linting/formatting
  • Create a PR with agent assistance
@workspace Use the PR Crafting skill.
Create a meaningful PR description for the
changes in the UserService.

Next Steps