Guides
Fix accessibility issues with a coding agent
Find and fix applicable keyboard, semantic, contrast, labeling, motion, and responsive accessibility problems.
Answers: use a coding agent to fix accessibility issues · Updated
An accessibility shift combines automated checks with interaction-level review. Fix observable barriers, verify keyboard and screen-reader semantics where relevant, and rescan until no new in-scope findings remain or time expires.
Inspect interaction, not only markup
Automated rules catch important failures but not whether a workflow makes sense. Walk the primary paths with a keyboard and inspect names, roles, focus order, state, and error recovery.
- Semantic headings and landmarks
- Keyboard reachability and visible focus
- Accessible names and form errors
- Contrast and non-color cues
- Reduced-motion behavior
- Responsive zoom and reflow
Fix at the right layer
Prefer native semantics and shared components over page-specific patches. Add focused tests when the project has an established accessibility test path.
- Do not hide focus outlines.
- Do not add ARIA where HTML already expresses the behavior.
- Keep touch targets usable on small screens.
Leave evidence
Record the path tested, the barrier repaired, the verification used, and any remaining trade-off that needs owner review.
Start with a bounded prompt
This prompt names the outcome and preserves Nightshift’s review boundary. Paste it into the supported coding host from the project you want to change.
Run the accessibility repair shift using this project's existing accessibility checks. Fix objective violations, preserve the design system, and park findings that require human judgment.What this workflow does not claim
- The shift requires an existing configured accessibility check and does not silently add one.
- Automated checks cannot establish WCAG or legal compliance.
Evidence and sources
These links support the released behavior, public outcomes, or problem language described on this page.