
Accessibility Bug Hunter for AI Coding Agents
Finds accessibility problems in UI code and turns them into prioritized fixes, WCAG-aware checklists, test plans, remediation tickets, and safe AI coding prompts.
- Audit React/Vue components for WCAG-aware accessibility issues.
- Generate Cursor prompts to fix focus traps in modals and dropdowns.
- Create manual QA checklists for keyboard and screen reader testing.
Secure checkout via Stripe
Included in download
- Audit React/Vue components for WCAG-aware accessibility issues.
- Generate Cursor prompts to fix focus traps in modals and dropdowns.
- browser, terminal, file_write automation included
- Ready for Compatible with ChatGPT Custom GPTs
Sample Output
A real example of what this skill produces.
=== ACCESSIBILITY BUG HUNTER REPORT === Original request: Audit this React checkout form for accessibility issues and create a Cursor prompt to fix it. Component/page: React checkout form UI pattern: Form, checkout flow, payment-adjacent workflow Interpreted accessibility goal: Identify and fix accessibility barriers that could prevent users from completing checkout using keyboard, screen reader, or assistive technology. Confirmed facts: - The target is a React checkout form. - The focus areas are labels, form errors, keyboard navigation, focus states, screen reader behavior, ARIA usage, loading states, and validation behavior. - Existing checkout logic must be preserved. - Unrelated files should not be rewritten. Assumptions: - The form contains required fields and validation. - The form may include address, contact, shipping, or payment-related fields. - The form may display loading and error states. Unknowns to inspect: - whether every input has a connected label - whether error text is associated with invalid fields - whether focus moves after failed submission - whether disabled and loading states are announced - whether all controls are keyboard operable - whether ARIA is correct or unnecessary Accessibility issue summary: Checkout forms are high-impact accessibility targets because inaccessible labels, validation errors, focus behavior, or keyboard support can prevent users from completing purchases or account actions. Prioritized issues: 1. Missing or disconnected form labels Severity: High Affected users: Screen reader users, voice control users, users with cognitive disabilities Why it matters: Users may not know what each field requires. Recommended fix: Use explicit label elements connected to inputs with htmlFor/id. Do not rely on placeholder text as the only label. Acceptance criteria: Every form control has a programmatically associated accessible name. Testing method: Navigate with a screen reader or inspect the accessibility tree to confirm each field name is announced. 2. Error messages not associated with invalid fields Severity: High Affected users: Screen reader users, keyboard users, users with cognitive disabilities Why it matters: Users may not know which field failed validation or how to fix it. Recommended fix: Use aria-invalid="true" on invalid fields and connect error text with aria-describedby. Consider moving focus to the first invalid field or an error summary after failed submission. Acceptance criteria: Each invalid field exposes its error message programmatically. Testing method: Submit the form with empty required fields and verify that errors are announced or discoverable. 3. Weak focus visibility or unpredictable focus order Severity: High Affected users: Keyboard-only users and low-vision users Why it matters: Users may lose track of their position in the form. Recommended fix: Preserve or improve visible focus indicators and ensure tab order follows the visual and logical order. Acceptance criteria: Every interactive element has a visible focus state and can be reached in logical order. Testing method: Complete the form using only Tab, Shift+Tab, Enter, and Space. Semantic HTML recommendations: - Use form, label, input, select, textarea, button, fieldset, and legend where appropriate. - Use button type="submit" for the submit action. - Avoid clickable divs or spans for form actions. - Use fieldset and legend for grouped radio buttons or checkboxes. ARIA review: - Use aria-invalid only when fields are invalid. - Use aria-describedby for helper and error text. - Use aria-live carefully for form-level success or error messages. - Avoid redundant roles on native controls. - Avoid aria-hidden on focusable elements. Keyboard and focus review: - All controls must be reachable by keyboard. - Submit must work by keyboard. - Focus should not be lost after validation. - Focus should move predictably to the first invalid field or error summary. - Loading state should not create a keyboard trap. Form accessibility review: - Required fields should be communicated in text. - Helper text should be associated with relevant fields. - Error text should be specific and actionable. - Autocomplete attributes should be considered for name, email, address, postal code, and payment-adjacent fields when appropriate. Dynamic content review: - Loading state after submit should be communicated. - Server errors should be visible and announced. - Success state should be clear. - Validation changes should not be overly noisy. Remediation plan: 1. Inspect all form controls and connect labels. 2. Add IDs for helper text and error text. 3. Add aria-describedby where needed. 4. Add aria-invalid for invalid fields only. 5. Improve focus handling after failed submission. 6. Confirm submit/loading/disabled states are communicated. 7. Verify keyboard-only completion. 8. Add or recommend tests for labels, errors, focus, and submission behavior. AI coding agent prompt: Inspect the React checkout form for accessibility issues before editing. Preserve existing checkout business logic, validation behavior, payment behavior, API calls, and styling conventions. Do not rewrite unrelated files. Prefer semantic HTML before ARIA. Ensure every form control has a connected label, invalid fields use aria-invalid only when invalid, helper and error text are connected with aria-describedby, and validation errors are visible and programmatically associated. Ensure keyboard users can reach and operate every control, visible focus states are preserved or improved, and focus moves predictably after failed submission. Communicate loading, disabled, success, and error states accessibly. Avoid unnecessary ARIA and do not expose sensitive payment or personal data. Add or recommend tests for labels, validation errors, keyboard submission, and focus behavior. Return files inspected, issues fixed, tests added or run, manual QA steps, and remaining risks. Accessibility test plan: - Complete the form using keyboard only. - Confirm every input has an announced label. - Submit with empty required fields. - Confirm errors are associated with fields. - Confirm focus moves predictably. - Confirm focus indicator is visible. - Confirm loading and disabled states are understandable. - Confirm no console errors or broken validation behavior. - Run existing tests if available. - Perform a screen reader smoke test if possible. Verification checklist: - all form controls have accessible names - errors are associated with invalid fields - required fields are communicated - focus is visible - tab order is logical - submit works by keyboard - loading and error states are understandable - ARIA is valid and not excessive - existing checkout logic is unchanged Compliance caution: This is a WCAG-aware remediation plan, not a legal compliance guarantee. Formal compliance requires manual testing, assistive technology testing, expert review, and jurisdiction-specific legal guidance.

Accessibility Bug Hunter for AI Coding Agents
Finds accessibility problems in UI code and turns them into prioritized fixes, WCAG-aware checklists, test plans, remediation tickets, and safe AI coding prompts.
Secure checkout via Stripe
Included in download
- Audit React/Vue components for WCAG-aware accessibility issues.
- Generate Cursor prompts to fix focus traps in modals and dropdowns.
- browser, terminal, file_write automation included
- Ready for Compatible with ChatGPT Custom GPTs
- Instant install
Sample Output
A real example of what this skill produces.
=== ACCESSIBILITY BUG HUNTER REPORT === Original request: Audit this React checkout form for accessibility issues and create a Cursor prompt to fix it. Component/page: React checkout form UI pattern: Form, checkout flow, payment-adjacent workflow Interpreted accessibility goal: Identify and fix accessibility barriers that could prevent users from completing checkout using keyboard, screen reader, or assistive technology. Confirmed facts: - The target is a React checkout form. - The focus areas are labels, form errors, keyboard navigation, focus states, screen reader behavior, ARIA usage, loading states, and validation behavior. - Existing checkout logic must be preserved. - Unrelated files should not be rewritten. Assumptions: - The form contains required fields and validation. - The form may include address, contact, shipping, or payment-related fields. - The form may display loading and error states. Unknowns to inspect: - whether every input has a connected label - whether error text is associated with invalid fields - whether focus moves after failed submission - whether disabled and loading states are announced - whether all controls are keyboard operable - whether ARIA is correct or unnecessary Accessibility issue summary: Checkout forms are high-impact accessibility targets because inaccessible labels, validation errors, focus behavior, or keyboard support can prevent users from completing purchases or account actions. Prioritized issues: 1. Missing or disconnected form labels Severity: High Affected users: Screen reader users, voice control users, users with cognitive disabilities Why it matters: Users may not know what each field requires. Recommended fix: Use explicit label elements connected to inputs with htmlFor/id. Do not rely on placeholder text as the only label. Acceptance criteria: Every form control has a programmatically associated accessible name. Testing method: Navigate with a screen reader or inspect the accessibility tree to confirm each field name is announced. 2. Error messages not associated with invalid fields Severity: High Affected users: Screen reader users, keyboard users, users with cognitive disabilities Why it matters: Users may not know which field failed validation or how to fix it. Recommended fix: Use aria-invalid="true" on invalid fields and connect error text with aria-describedby. Consider moving focus to the first invalid field or an error summary after failed submission. Acceptance criteria: Each invalid field exposes its error message programmatically. Testing method: Submit the form with empty required fields and verify that errors are announced or discoverable. 3. Weak focus visibility or unpredictable focus order Severity: High Affected users: Keyboard-only users and low-vision users Why it matters: Users may lose track of their position in the form. Recommended fix: Preserve or improve visible focus indicators and ensure tab order follows the visual and logical order. Acceptance criteria: Every interactive element has a visible focus state and can be reached in logical order. Testing method: Complete the form using only Tab, Shift+Tab, Enter, and Space. Semantic HTML recommendations: - Use form, label, input, select, textarea, button, fieldset, and legend where appropriate. - Use button type="submit" for the submit action. - Avoid clickable divs or spans for form actions. - Use fieldset and legend for grouped radio buttons or checkboxes. ARIA review: - Use aria-invalid only when fields are invalid. - Use aria-describedby for helper and error text. - Use aria-live carefully for form-level success or error messages. - Avoid redundant roles on native controls. - Avoid aria-hidden on focusable elements. Keyboard and focus review: - All controls must be reachable by keyboard. - Submit must work by keyboard. - Focus should not be lost after validation. - Focus should move predictably to the first invalid field or error summary. - Loading state should not create a keyboard trap. Form accessibility review: - Required fields should be communicated in text. - Helper text should be associated with relevant fields. - Error text should be specific and actionable. - Autocomplete attributes should be considered for name, email, address, postal code, and payment-adjacent fields when appropriate. Dynamic content review: - Loading state after submit should be communicated. - Server errors should be visible and announced. - Success state should be clear. - Validation changes should not be overly noisy. Remediation plan: 1. Inspect all form controls and connect labels. 2. Add IDs for helper text and error text. 3. Add aria-describedby where needed. 4. Add aria-invalid for invalid fields only. 5. Improve focus handling after failed submission. 6. Confirm submit/loading/disabled states are communicated. 7. Verify keyboard-only completion. 8. Add or recommend tests for labels, errors, focus, and submission behavior. AI coding agent prompt: Inspect the React checkout form for accessibility issues before editing. Preserve existing checkout business logic, validation behavior, payment behavior, API calls, and styling conventions. Do not rewrite unrelated files. Prefer semantic HTML before ARIA. Ensure every form control has a connected label, invalid fields use aria-invalid only when invalid, helper and error text are connected with aria-describedby, and validation errors are visible and programmatically associated. Ensure keyboard users can reach and operate every control, visible focus states are preserved or improved, and focus moves predictably after failed submission. Communicate loading, disabled, success, and error states accessibly. Avoid unnecessary ARIA and do not expose sensitive payment or personal data. Add or recommend tests for labels, validation errors, keyboard submission, and focus behavior. Return files inspected, issues fixed, tests added or run, manual QA steps, and remaining risks. Accessibility test plan: - Complete the form using keyboard only. - Confirm every input has an announced label. - Submit with empty required fields. - Confirm errors are associated with fields. - Confirm focus moves predictably. - Confirm focus indicator is visible. - Confirm loading and disabled states are understandable. - Confirm no console errors or broken validation behavior. - Run existing tests if available. - Perform a screen reader smoke test if possible. Verification checklist: - all form controls have accessible names - errors are associated with invalid fields - required fields are communicated - focus is visible - tab order is logical - submit works by keyboard - loading and error states are understandable - ARIA is valid and not excessive - existing checkout logic is unchanged Compliance caution: This is a WCAG-aware remediation plan, not a legal compliance guarantee. Formal compliance requires manual testing, assistive technology testing, expert review, and jurisdiction-specific legal guidance.
About This Skill
Accessibility Bug Hunter helps AI coding agents, frontend developers, QA teams, product teams, SaaS builders, design-system maintainers, and accessibility-minded creators identify and fix practical UI accessibility issues. It audits components and pages for semantic HTML, keyboard navigation, focus management, visible focus states, accessible names, form labels, error associations, ARIA misuse, dynamic content announcements, contrast risks, modal behavior, dropdowns, tabs, tables, and screen reader risks. The skill creates prioritized accessibility reports, remediation plans, QA checklists, engineering tickets, test strategies, and paste-ready prompts for Cursor, Claude Code, Codex CLI, OpenCode, Replit, and ChatGPT Agents. It is ideal for improving React, Vue, Angular, Svelte, HTML/CSS/JS, SaaS dashboards, forms, checkout flows, admin panels, and design-system components.
📖 Learn more: Best Frontend & Design Skills for Claude Code →
Use Cases
- Audit React/Vue components for WCAG-aware accessibility issues.
- Generate Cursor prompts to fix focus traps in modals and dropdowns.
- Create manual QA checklists for keyboard and screen reader testing.
- Convert accessibility debt into prioritized engineering tickets.
- Ensure form validation errors are programmatically announced.
- Plan accessibility regression tests
Known Limitations
This skill creates WCAG-aware accessibility audits, remediation plans, checklists, and AI coding prompts, but it does not guarantee legal compliance, full WCAG conformance, ADA compliance, Section 508 compliance, EN 301 549 compliance, or complete assistive technology compatibility. Automated and AI-assisted review cannot replace manual keyboard testing, screen reader testing, user testing, expert accessibility review, or jurisdiction-specific legal guidance. Visual contrast, dynamic behavior, focus management, screen reader output, and real user experience may require direct testing in the browser and assistive technologies.
How to Install
mkdir -p ~/.claude/skills && curl -sL https://www.agensi.io/api/install/accessibility-bug-hunter-for-ai-coding-agents | tar xz -C ~/.claude/skills/Free skills install directly. Paid skills require purchase - use the download button above after buying.
Reviews
No reviews yet - be the first to share your experience.
Only users who have downloaded or purchased this skill can leave a review.
Early access skill
Be the first to review this skill.
Only users who have downloaded or purchased this skill can leave a review.
Security Scanned
Passed automated security review
Permissions
Allowed Hosts
File Scopes
This skill uses file access to read user-provided UI code, component files, HTML, CSS, JavaScript, TypeScript, design-system notes, test files, README files, accessibility reports, and UI specifications. It uses write access to create structured Markdown/text outputs such as accessibility audits, remediation plans, QA checklists, engineering tickets, test plans, AI coding prompts, verification checklists, documentation notes, and SKILL.md files. Terminal access is optional and should only be enabled when the agent is expected to run tests or accessibility tooling. Browser or network access is only needed for external documentation research. Environment variable access is not required.
Tags
Compatible with ChatGPT Custom GPTs, ChatGPT Agents, Cursor, Claude Code, Codex CLI, OpenCode, Replit, GitHub Copilot-style workflows, and other AI coding assistants that support structured Markdown instruction files such as SKILL.md. It can also be used manually in any AI chat by pasting the instructions.