Accessibility Audit Report

Accessibility Audit Report

Before you start

Gather the following from the user. If anything is missing, ask before proceeding:

1. Scope — Which pages, flows, or components are being audited? (e.g., "checkout flow", "marketing site", "design system")
2. Target conformance level — WCAG 2.1 AA (most common), WCAG 2.2 AA, or AAA?
3. Testing methods used — Automated tools (axe, Lighthouse), manual keyboard/screen reader testing, user testing with disabled participants?
4. Tech stack — Framework, component library, and any existing a11y tooling in the pipeline
5. Audience context — Is there a legal compliance deadline, a known user complaint, or a proactive audit?

If the user says "just audit everything," push back: "Which user flows are highest traffic or highest risk? Start there — a focused audit produces actionable results, a broad audit produces a backlog nobody reads."

Audit report template

1. Executive Summary

3-5 sentences. State the overall conformance status, the number of issues found by severity, and the single most critical finding. This section is for stakeholders who will not read the full report.

This audit evaluated 12 screens across the checkout flow against WCAG 2.1 AA.
We identified 23 issues: 4 critical, 7 major, 9 minor, and 3 best-practice
recommendations. The most critical finding is that the payment form is entirely
inaccessible to keyboard-only users, blocking approximately 8% of users from
completing purchases.

2. Scope & Methodology

List what was tested, what was not, and how. Be explicit about tools and versions.

Tested: /cart, /checkout/shipping, /checkout/payment, /checkout/confirmation
Not tested: Account settings, admin dashboard
Methods: axe-core 4.9, NVDA 2024.1 + Chrome, VoiceOver + Safari, manual
keyboard navigation, color contrast analyzer

3. Summary of Findings

Provide a scannable table. Every issue in the report must appear here.

| # | Issue | WCAG Criterion | Severity | Component/Page | |---|-------|---------------|----------|----------------| | 1 | Payment form not keyboard accessible | 2.1.1 Keyboard | Critical | /checkout/payment | | 2 | Images missing alt text | 1.1.1 Non-text Content | Major | /cart | | 3 | Insufficient color contrast on helper text | 1.4.3 Contrast (Minimum) | Minor | Global |

4. Detailed Findings

Write one section per issue. Every finding must include all six fields below — no exceptions.

Finding #[N]: [Short description]

• WCAG Criterion: [Number + Name] (e.g., 2.1.1 Keyboard)
• Severity: Critical / Major / Minor / Best Practice
• Affected Element: [CSS selector, component name, or page URL]
• Description: What the problem is, who it affects, and what the user experiences. Be specific — "screen reader users cannot determine the purpose of this button" not "button is inaccessible."
• Remediation: Step-by-step fix. Include the specific ARIA attribute, HTML element, or CSS property needed.
• Code Example:

Pay Now

Pay Now

Severity definitions:

| Severity | Meaning | |----------|---------| | Critical | Blocks a user group from completing a core task. Fix immediately. | | Major | Causes significant difficulty but a workaround exists. Fix within current sprint. | | Minor | Causes inconvenience but does not block functionality. Fix within the quarter. | | Best Practice | Not a WCAG violation but improves the experience. Schedule as capacity allows. |

5. Remediation Priority Matrix

Group findings by effort and impact to help the team sequence work.

| | Low Effort | High Effort | |---|-----------|-------------| | High Impact | Fix first (e.g., missing alt text, missing labels) | Plan next (e.g., rebuild inaccessible custom widget) | | Low Impact | Quick wins (e.g., skip-to-content link) | Backlog (e.g., ARIA live region for non-critical notifications) |

6. Recommended Timeline

Map findings to concrete timeframes tied to severity.

Week 1-2:  All Critical findings (#1, #4)
Week 3-4:  All Major findings (#2, #5, #7, #8, #10, #11, #12)
Week 5-8:  Minor findings, prioritized by the matrix above
Ongoing:   Best-practice recommendations integrated into component library

Include a note on regression prevention: "Add axe-core to CI. Every new component must pass automated a11y checks before merge."

Quality checklist

Before delivering the report, verify:

• [ ] Every finding maps to a specific WCAG success criterion, not just a general principle
• [ ] Severity levels are consistent — same type of issue gets the same severity throughout the report
• [ ] Remediation steps are specific enough for a developer to implement without further research
• [ ] Code examples show before AND after, not just the correct version
• [ ] The summary table matches the detailed findings exactly — no orphaned or missing entries
• [ ] The executive summary includes a concrete number of issues and highlights the worst one
• [ ] The timeline is realistic given the severity distribution and team capacity
• [ ] Testing methodology is documented — another auditor could reproduce the findings

Common mistakes to avoid

• Listing the tool output as the report. axe-core output is raw data, not a report. Every automated finding needs human interpretation: who is affected, how badly, and what to do about it.
• Missing the keyboard test. Automated tools catch roughly 30% of accessibility issues. If you only ran axe and Lighthouse, say so — and note that keyboard and screen reader testing is still needed.
• Vague remediation. "Make this accessible" is not a remediation step. "Add aria-label="Close dialog" to the ` element at .modal-close`" is.
• Inconsistent severity. If a missing alt text on a decorative image is "Critical" but a missing alt text on a product image is "Minor," your severity framework is broken. Define it once, apply it uniformly.
• No prioritization. A flat list of 40 findings paralyzes teams. The priority matrix and timeline exist to prevent this — always include them.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: The-AI-Directory-Company
• Source: The-AI-Directory-Company/agents-and-skills
• License: MIT
• Homepage: https://ai-directory.company

Install and usage instructions live in the source repository linked above.