Health checks & recommendations
Get an AI health report on your project with concrete, actionable recommendations.
Open a project's Health tab and Evonx scores the codebase two ways: a deterministic health check of things the platform can measure, and an AI Health Check that reads the same project and returns risks with recommended next actions.
What you see
- Health Check — a score plus a list of concrete items (buildability, tests, configuration gaps). Each item passes or fails for a reason you can inspect.
- AI Health Check — a summary, a second score, and a list of risks with severity and reasoning. High-severity items are the ones worth acting on first.
Acting on a finding
-
Open the Health tab on the project you care about.
-
Read the AI summary, then the high-severity risks. Each risk names a problem in the codebase — not a generic lecture.
-
Turn a risk into work: send it to chat (the panel offers this) so an Evolution Thread can plan the fix. Verify the result in a preview before merging.
This is most useful on imported repositories, where Evonx didn't write the first version and the scan is how you learn what you're starting from. Greenfield projects still get a score; treat it as a baseline, not a verdict.
Related
- Import from GitHub — where health checks earn their keep on existing code.
- Monitoring & publish metrics — traffic and errors once the app is live, which is a different question than codebase health.
Last updated: September 4, 2026