fixing-accessibility
by ibelickfixing-accessibility helps audit and repair HTML accessibility issues before release. Use it for buttons, forms, dialogs, tabs, icon-only controls, keyboard flow, focus handling, form errors, contrast, and screen reader labeling. The fixing-accessibility skill is best for targeted UI code fixes, not broad compliance reports.
This skill scores 82/100, which means it is a solid directory listing for users who need a targeted accessibility-fixing workflow. The repository gives enough concrete guidance for agents to trigger it, inspect UI files, and make minimal code-level fixes with less guesswork than a generic prompt.
- Clear triggerability: the description and `/fixing-accessibility` usage pattern make it easy for an agent to know when to apply it.
- Operationally useful workflow: it tells the agent to quote exact violations, explain impact, and suggest concrete fixes instead of rewriting large UI sections.
- Good coverage of real UI failure modes: accessible names, keyboard access, focus/dialogs, forms/errors, announcements, contrast, and motion are all explicitly scoped.
- No install command, scripts, or supporting files, so adoption depends on reading SKILL.md rather than a broader toolchain or examples.
- The skill is strong on rules but lighter on worked examples and edge-case guidance, which may leave some implementation details to agent judgment.
Overview of fixing-accessibility skill
The fixing-accessibility skill helps you audit and repair common HTML accessibility problems before they ship. It is best for designers, frontend engineers, and AI agents working on UI changes that affect keyboard use, screen readers, forms, dialogs, and visual state clarity. If you need fixing-accessibility for UX Audit, this skill turns a rough review into a focused remediation pass instead of a generic accessibility lecture.
What fixing-accessibility actually does
This fixing-accessibility skill looks for concrete issues such as missing accessible names, broken keyboard navigation, weak focus handling, invalid dialog behavior, poor form error messaging, and contrast or state problems. Its main value is not broad WCAG theory; it is giving you a practical review frame that catches the issues most likely to block real users.
Best fit and clear boundaries
Use it when you are adding or changing controls like buttons, inputs, menus, tabs, dropdowns, modals, or icon-only actions. It is less useful for full enterprise accessibility audits, legal compliance review, or visual design critiques that require human policy judgment. The skill is strongest when the task is to fix UI code, not to write a report.
Why this skill is worth installing
The biggest advantage of fixing-accessibility install is that it constrains the agent toward minimal, targeted fixes instead of rewriting the UI. That makes it easier to adopt in live codebases where you need safe changes, specific line-level findings, and quick iteration on accessibility defects.
How to Use fixing-accessibility skill
Install and trigger it in context
Use fixing-accessibility install in your skill manager, then invoke it directly in a UI-focused conversation or review. The canonical pattern in the repo is /fixing-accessibility for applying the guidance to the current discussion, or /fixing-accessibility <file> when you want a file review with specific findings. Keep the request anchored to the component or screen you are changing.
Give the skill a task it can act on
The best fixing-accessibility usage starts with a concrete target, not a vague request like “check accessibility.” Say what changed, where it lives, and what kind of interaction it has. Strong prompts include the UI pattern, expected behavior, and any constraints on edits.
Example prompt:
/fixing-accessibility src/components/Modal.tsx Review for keyboard access, focus trap, aria labeling, and escape handling. Keep fixes minimal and preserve existing design.
Read the right files first
Start with SKILL.md, because it contains the rule priorities and the interaction model. Then inspect the component or page file, plus any nearby form, modal, or shared interaction utilities. Since this repository has no extra rules, references, or scripts, the skill is intentionally lightweight; that means the main gain comes from applying the guidance to your own codebase carefully, not from hunting for hidden helpers.
Workflow that improves output quality
- Identify the interaction type: form, dialog, menu, tabs, icon button, or hidden content.
- Ask for a focused review against the skill’s priority categories.
- Require exact snippets or lines for each violation.
- Ask for a minimal code-level fix, not a redesign.
- Re-run after changes to catch regressions in keyboard flow, labeling, or focus.
fixing-accessibility skill FAQ
Is fixing-accessibility for UX Audit or code repair?
Both, but it is primarily a repair-oriented review skill. For UX Audit work, it helps you find accessibility blockers and turn them into actionable fixes. If you need a narrative audit document with severity scoring, you may need to combine it with a broader review process.
How is this different from a normal accessibility prompt?
A normal prompt often produces a generic checklist. The fixing-accessibility skill is narrower and more operational: it prioritizes accessible names, keyboard access, focus management, semantics, form errors, announcements, and contrast in a fixed order. That structure helps an agent make fewer irrelevant suggestions.
Can beginners use it?
Yes. Beginners usually benefit most because the skill tells them what to inspect first and what to fix minimally. The main limitation is that they still need to provide the specific component or file; the skill cannot guess which UI surface matters.
When should I not use it?
Do not rely on it for legal compliance sign-off, complex accessibility policy decisions, or tests that require specialized assistive technology validation. It is also a poor fit when you want broad redesign suggestions instead of targeted fixes to existing UI code.
How to Improve fixing-accessibility skill
Provide stronger input than “make it accessible”
The best results come from naming the component, interaction, and user risk. For example, “Review DatePicker.tsx for keyboard navigation, focus return, and announced errors after validation” is much better than “fix accessibility.” This gives the skill a clear failure surface and reduces superficial output.
Ask for evidence, not only advice
When you want a useful fixing-accessibility guide output, ask the assistant to quote the exact snippet or line that violates a rule, explain why it matters in one sentence, and suggest the smallest viable fix. That format makes it easier to act on the review and easier to verify that the issue is real.
Iterate on the highest-risk categories first
If the first pass finds many issues, fix them in this order: accessible names, keyboard access, focus and dialogs, then semantics and forms. Those categories usually have the biggest user impact and the clearest code-level resolution. After that, rerun the skill on the updated file to catch regressions in state, contrast, or announcements.
Common failure modes to avoid
The most common mistake is asking for broad accessibility feedback without specifying the UI pattern or file. Another is allowing large rewrites when the skill is designed for targeted edits. A third is ignoring input context, such as whether a modal traps focus or whether a button is icon-only, which can lead to generic advice instead of the exact fix the code needs.