# Test Infra

This directory is the working home for NexusCRM regression infrastructure.

It exists to answer three questions for future agents and maintainers:

1. What kinds of regression assets do we maintain?
2. Where should a newly fixed issue add protection?
3. How do quick CI-facing checks differ from slower UI / staging inspection?

## Read Order

1. `regression-strategy.md`
2. `issue-to-test-policy.md`
3. `regression-runbook.md`
4. the current backlog documents

## Document Map

| File | Purpose |
|---|---|
| `regression-strategy.md` | Defines the two regression tiers: fast API checks vs slower UI/runtime inspection, plus scenario coverage. |
| `issue-to-test-policy.md` | Tells agents what to do after fixing an issue so regression assets keep growing. |
| `regression-runbook.md` | Describes how these assets should run in CI, staging, and future production-style inspection loops. |
| `api-regression-backlog-2026-04-23.md` | Fast API / contract / integration-oriented regression backlog. |
| `ui-regression-backlog-2026-04-23.md` | UI/runtime/scenario-oriented regression backlog. |

## Core Operating Model

- Fast feedback belongs in API-facing regression assets that can be run often and eventually enforced in CI.
- Slower but higher-fidelity coverage belongs in UI/runtime inspection assets, including multi-step scenarios.
- A completed bug fix is not the end of the loop if it leaves no durable regression protection behind.
- Not every issue needs both tiers, but every high-risk merged issue must be classified against both tiers.