508-audit

name

508-audit

description

Run a full Section 508 + WCAG 2.2 Level AA accessibility audit on a federal / .gov / public-sector website or ICT product. Section 508 incorporates WCAG 2.0 AA as the binding federal standard, but most modern federal procurements now test against WCAG 2.2 AA (the 2024-incorporated update). Outputs the official compliance trio agencies require: (1) Voluntary Product Accessibility Template (VPAT® 2.5 Rev 508) covering all 508 chapters, (2) Accessibility Conformance Report (ACR) with each criterion marked Supports / Partially Supports / Does Not Support / Not Applicable + explanation, (3) Remediation plan with prioritized fixes (P0 blockers like keyboard traps, P1 perceivability fails, P2 cosmetic). Combines axe-core automated scan + manual keyboard / screen-reader / cognitive walkthrough findings. TRIGGER on "Section 508", "508 audit", "WCAG", "VPAT", "ACR", "federal accessibility", ".gov accessibility", "ADA Title II", "ICT accessibility", "accessibility audit for government", "GSA accessibility".

Section 508 + WCAG 2.2 Accessibility Audit

You run a federal-grade accessibility audit and produce the deliverable trio agencies require (VPAT 2.5 + ACR + remediation plan). Federal procurement is increasingly contingent on these — failure to provide credible 508 documentation eliminates vendors from competition.

Standards landscape (2026):

• Section 508 (Revised 2017): binding for federal procurement. Incorporates WCAG 2.0 AA technical criteria.
• WCAG 2.2 AA (W3C, October 2023): adds 9 new success criteria (focus appearance, dragging movements, target size, etc.). Many agencies test against 2.2 even though 508 floor is still 2.0. Testing higher is fine.
• ADA Title II (April 2024 final rule): public entities (state, local governments) must conform to WCAG 2.1 AA by April 2026 (large entities) or 2027 (small).
• Section 504 (federal grantees): similar requirements via funding agreements.

============================================================ === PRE-FLIGHT ===

Verify:

• Site URL + scope (pages to audit). Federal audit baseline: every "key" template + ≥ 25 sampled pages.
• Site rendering: static HTML / SSR / CSR? Pure CSR can't be audited by lightweight tools without rendering.
• Stack: identify frontend framework + UI library (React + MUI, Vue + Vuetify, etc.). Different defaults.
• Documentation deliverable target: VPAT 2.5 Rev 508, VPAT 2.5 Int (international), or VPAT 2.5 INT/EU (with EN 301 549).
• Test tools available: axe-core, Lighthouse, WAVE, screen reader (NVDA / JAWS / VoiceOver), keyboard.

Recovery:

• If site is CSR-only, render via Playwright before scanning — axe-core needs DOM, not raw HTML.
• If user can't provide screen-reader access, output manual test cases for them to run + capture results.

============================================================ === PHASE 1: AUTOMATED SCAN ===

Run axe-core (@axe-core/cli or via Playwright @axe-core/playwright) against every priority URL:

npx @axe-core/cli --rules wcag22aa,best-practice https://example.gov \
  --save axe-report.json \
  --tags wcag2a,wcag2aa,wcag21a,wcag21aa,wcag22a,wcag22aa

Also run Lighthouse accessibility category for each URL. Persist both as automated_findings.json.

Caveats explicit to the user:

• Automated tools catch ~30-40% of real WCAG issues. The remaining 60-70% require manual testing.
• Common false positives: color-contrast on text in images (axe can't read), aria misuse in WAI-ARIA patterns.
• Auto-scan is the FLOOR, not the ceiling.

VALIDATION: Automated scan completes on every priority URL with no exit-1. Findings deduplicated.

============================================================ === PHASE 2: KEYBOARD AUDIT ===

Manual keyboard walkthrough. For each priority page:

• Skip link present and functional (first Tab focus).
• Tab order logical (matches visual reading order).
• Focus visible at every interactive element (WCAG 2.4.7).
• No keyboard traps (WCAG 2.1.2).
• Custom widgets (combobox, menu, tabs, dialog) operable via expected keys per WAI-ARIA Authoring Practices.
• Modal dialogs trap focus correctly AND restore focus on close.
• WCAG 2.2 SC 2.4.11 Focus Not Obscured: focused element not hidden by sticky header / cookie banner.
• WCAG 2.2 SC 2.5.7 Dragging Movements: every drag operation has a non-drag alternative (e.g., click buttons).
• WCAG 2.2 SC 2.5.8 Target Size: interactive targets ≥ 24×24 CSS pixels (exceptions noted).

Generate keyboard_findings.md with per-page checklist and violations.

VALIDATION: Every interactive element tested. Findings reference specific WCAG SC.

============================================================ === PHASE 3: SCREEN READER AUDIT ===

Test priority pages with NVDA (Windows + Chrome/Firefox) and VoiceOver (Mac/iOS). Report per page:

• Page title announced and meaningful (not "untitled").
• Heading hierarchy (h1 → h6) maps to visual structure.
• Landmarks (<header>, <nav>, <main>, <aside>, <footer>) present.
• Form labels associated via <label for> or aria-labelledby.
• Error messages announced (aria-live or focus-shift).
• Images: meaningful images have alt; decorative use alt="" or role="presentation".
• SVG: <title> + aria-labelledby or hidden via aria-hidden.
• Tables: <th> + scope for data tables; layout tables avoided.
• Live regions for dynamic updates (aria-live="polite" or "assertive").
• Custom widgets: announce role, state (expanded/collapsed), value.

Generate sr_findings.md per page with screen-reader-specific issues.

VALIDATION: Findings cite specific SC + reproducer steps.

============================================================ === PHASE 4: COGNITIVE / READABILITY AUDIT ===

WCAG 2.2 added several cognitive-load SC:

• 3.2.6 Consistent Help — help mechanism (contact, FAQ, chat) at consistent position across pages.
• 3.3.7 Redundant Entry — info previously entered is auto-populated or selectable.
• 3.3.8 Accessible Authentication (Minimum) — no cognitive function test (typing CAPTCHA, recalling password from memory) unless alternative provided.
• Plain language — Flesch Reading Ease ≥ 60 for civic content (state/federal guideline).
• Forms — autocomplete attributes set (autocomplete="email", autocomplete="given-name").
• Session timeout warning + extension UI.
• Errors explain WHAT and HOW TO FIX, not just "invalid input."

VALIDATION: Each page has cognitive checklist run with findings.

============================================================ === PHASE 5: VPAT 2.5 GENERATION ===

Generate vpat.docx (or vpat.md if user prefers) following the VPAT® 2.5 Rev 508 template from ITI / Section508.gov:

Sections required:

• Cover sheet (product, version, contact, evaluation method, date)
• Chapter 1: Application of Revised 508 Standards (which chapters apply)
• Chapter 2: 36 CFR Part 1194 Appendix A (Functional Performance Criteria)
• Chapter 3: WCAG 2.0 Level A & AA criteria (incorporated into 508)
• Chapter 4: Hardware (if applicable, often N/A for web)
• Chapter 5: Software (web applications fall here mostly)
• Chapter 6: Support Documentation and Services
• Conformance terms per criterion: Supports / Partially Supports / Does Not Support / Not Applicable + Remarks and Explanations

Per criterion remarks must be specific. "Supports" without explanation is suspect to procurement officers.

VALIDATION: VPAT covers ALL applicable WCAG 2.0 A/AA criteria + all 508 chapters. Every "Partially Supports" / "Does Not Support" has a remarks explanation.

============================================================ === PHASE 6: ACCESSIBILITY CONFORMANCE REPORT (ACR) ===

The ACR is the VPAT-formatted document signed and dated, typically delivered with the procurement response. Generate as acr.pdf (via pandoc + wkhtmltopdf or weasyprint) with:

• Signed conformance attestation
• Evaluation methodology summary (tools used, manual testing scope, sample size)
• Known issues with severity + planned remediation
• Roadmap dates for in-progress fixes
• Contact for accessibility questions

VALIDATION: ACR is a single PDF, signed, dated.

============================================================ === PHASE 7: REMEDIATION PLAN ===

Generate remediation_plan.md with findings ranked:

Priority	Finding	WCAG SC	Sample URL	Estimated effort	Suggested fix
P0	Keyboard trap in date picker	2.1.2	/apply	4 hr	Use <input type="date"> or trap-aware Floating UI
P0	No skip link	2.4.1	All pages	1 hr	Add <a href="#main" class="skip-link">Skip to main</a> in layout
P1	Insufficient color contrast on link hover	1.4.3	All pages	2 hr	Update --color-link-hover to ≥ 4.5:1
P2	Missing lang attribute	3.1.1	/es/*	1 hr	Set <html lang="es"> for Spanish pages

P0 = WCAG fail that prevents access entirely. P1 = fail that hinders. P2 = polish.

Output also as JIRA / GitHub Issues import CSV.

VALIDATION: Every finding from phases 1-4 appears in the plan. P0 items have effort estimates.

============================================================ === SELF-REVIEW ===

Score 1–5:

• Complete: All 7 phases delivered? VPAT + ACR + remediation plan present?
• Robust: Automated + manual + screen reader + cognitive covered?
• Clean: VPAT criterion remarks are specific (not "supports — no issues")?
• Federal-credible: Would a federal accessibility officer (or Section508.gov reviewer) accept this trio as procurement-ready?

Common gap: VPAT with vague "Supports" remarks. Each must reference specific testing evidence.

============================================================ === LEARNINGS CAPTURE ===

Append to ~/.claude/skills/508-audit/LEARNINGS.md:

— <site, framework, scope>

• What worked:
• What was awkward:
• Suggested patch:
• Verdict: [Smooth / Minor friction / Major friction]

============================================================ === STRICT RULES ===

• Never deliver a VPAT with only automated-scan findings. 60-70% of real issues require manual testing.
• Never claim "Supports" on a criterion without explaining the testing evidence.
• Never silently overlook WCAG 2.2 SCs. Even though 508 floor is 2.0 AA, agencies increasingly test against 2.2.
• Never sign an ACR without management review. The conformance attestation has legal weight.
• Always flag CSR-only sites as requiring rendering before audit. Bare-bones HTML scans miss the entire SPA.