Run automated checks

Before choosing a tool or writing more code, make the requirement concrete. The pipeline catches missing labels and invalid relationships, while manual testing judges focus order and error clarity.

Automation is excellent at repeated machine-testable rules and poor at deciding whether a task is understandable or an alternative is meaningful. The useful target is not encyclopedic coverage. Make one deliberate choice, observe its consequences, and know which requirement would make you revise it.

A common failure is to treat tool output as complete coverage or ignore it because it cannot catch everything. The happy path may still work, which makes the mistake easy to miss. automate reliable rules, review each result in context, and record which important checks remain manual. Keep the first version small enough that every important input and output remains visible.

This supports the module goal: Turn the repaired page into a repeatable accessibility practice rather than a one-time cleanup. Capture the request, command, trace, screenshot, test result, or other evidence that proves the result.

Run two tools, deduplicate findings, verify false positives, and map uncovered tasks to manual tests. Change one condition on purpose, predict the result, and compare the prediction with what actually happened.