AI-Powered Refactoring Summary
In this section, we completed the full AI-powered refactoring loop using Kiro and vFunction:
| Step | Action | Tool |
|---|---|---|
| 1 | Set refined measurement as baseline | vFunction Platform |
| 2 | Understood TODOs | vFunction Platform |
| 3 | Configured Kiro with vFunction | Kiro IDE |
| 4 | Listed domains and TODOs, explored and fixed a TODO | Kiro Chat + vFunction MCP |
| 5 | Built, deployed, re-learned, and validated resolution | vFunction Platform |
Key Takeaways
- vFunction identifies architectural issues through continuous observation and generates actionable TODO items.
- Kiro provides the agentic AI capability to fix those issues directly in the IDE using architecture-aware prompts from the vFunction MCP server.
- The loop is closed — after fixing code, vFunction re-learns and confirms that the architectural TODO has been resolved.
- No context switching — developers can discover, understand, fix, and verify architectural issues without leaving Kiro.
What You Learned
- How to set a refined measurement as a baseline for tracking architectural drift
- What TODOs are, their types, and how they represent architectural debt
- How to configure Kiro with the vFunction extension, Power, and MCP server
- How to use Kiro Chat to query domains, list TODOs, understand issues, and generate fixes
- How to build, deploy, re-learn, and validate that TODOs are resolved
- How to confirm resolution in the vFunction UI
Wrap Up
This concludes the AI-Powered Refactoring section. You can now repeat this process for additional TODOs to continue reducing architectural debt in the OMS application.
If you are interested to learn about extracting services, continue to the Service Extraction section.