Index of every page on this site: llms.txt
All skills
rshankras avatar

/accessibility-audit

@f958982
Run it yourself

Run a structured accessibility audit on an iOS/macOS app — automated XCUITest audits, Accessibility Inspector, manual VoiceOver/Dynamic Type passes, and App Store Accessibility Nutrition Label evaluation. Use before release, when preparing Nutrition Label declarations, or for EU Accessibility Act compliance.

Use this Skill: https://skilld.dev/gh/rshankras/claude-code-apple-skills/accessibility-audit

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈83 tokens for metadata: the name and description. ≈1.5k when used: this file. ≈3.3k more on demand in 2 files.

Description uses 4.1% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈331 characters, or 4.1%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Accessibility Audit

A repeatable audit workflow that takes an app from "we think it's accessible" to evidence: automated audits in CI, an Inspector pass, manual assistive-tech passes, and an Accessibility Nutrition Label evaluation you can defend. Distilled from Apple's WWDC23 "Perform accessibility audits" (10035), WWDC19 "Accessibility Inspector" (257), WWDC25 "Evaluate your app for Accessibility Nutrition Labels" (224), and the 2026 Tech Talk "Prepare your app for Accessibility Nutrition Labels" (111433).

Why it matters: Accessibility Nutrition Labels on the App Store make your support (or its absence) visible before download.

When This Skill Activates

  • User asks for an accessibility audit, review, or compliance check
  • User is preparing Accessibility Nutrition Label declarations for App Store Connect
  • User asks about performAccessibilityAudit, Accessibility Inspector, or automated a11y testing
  • User mentions EU Accessibility Act / accessibility compliance
  • Pre-release checks (pairs with release-review)

The Audit Workflow

1. Define common tasks first (WWDC25 224)

List the primary tasks people download the app for, plus the fundamentals: first-launch experience, login, purchase, settings. Every subsequent pass evaluates these tasks, on every device family the app supports.

2. Automated pass — XCUITest audits in CI

try app.performAccessibilityAudit() audits the current view exactly as the Inspector does; the test fails automatically on findings. Full API, audit-type table, issue filtering, and CI patterns: automated-audits.md. Baseline rules:

  • Audits cover only what's on screen — write one audit per distinct screen/state.
  • continueAfterFailure = true before the audit to surface all issues in one run.
  • Filter accepted issues via the issue handler — never by skipping whole audit types.

3. Inspector pass (WWDC19 257)

Xcode → Open Developer Tool → Accessibility Inspector, target the app:

  1. Run Audit → each finding highlights the view and offers a Help suggestion.
  2. Fix, re-run to zero.
  3. Auto Navigate through the screen with the speaker button — hear exactly what VoiceOver would say, in order; wrong reading order shows up here.
  4. Point Inspection for spot checks; Color Contrast Calculator (Window menu) for failing pairs.

Classic findings and fixes: filename-as-label (give a real accessibilityLabel; move technical IDs to accessibilityIdentifier), text drawn via CATextLayer invisible to VoiceOver (isAccessibilityElement = true + label), contrast below threshold (Inspector flags pairs; fix with darker/lighter variants until it passes).

4. Manual assistive-tech passes (per common task)

Pass How Pass criterion
VoiceOver Swipe right through every element; double-tap to activate; complete each common task eyes-free Every element speaks label + trait + value; task completable with gestures only
Voice Control Complete tasks by voice only Every interactive element has a speakable label (accessibilityInputLabels for synonyms)
Dynamic Type Test at 200% and at the largest accessibility size (310%) Text wraps (never truncates), fields grow, layout adapts
Sufficient Contrast Light + dark appearance, with Increase Contrast on Legible everywhere
Dark Interface Dark mode + Smart Invert Photos/video NOT inverted (accessibilityIgnoresInvertColors)
Reduced Motion Reduce Motion on Zoom/slide transitions, autoplay, parallax replaced (modified, not just removed)
Keyboard (Mac / iPad FKA) Complete tasks keyboard-only Focus reaches everything; no hover-only affordances

5. Nutrition Label evaluation and declaration

Map the evidence from passes 2–4 onto the nine App Store features and declare only what holds for all common tasks on all supported device families. Per-feature criteria, disqualifier examples, and the declaration flow: nutrition-labels.md. The model behavior (Apple's own demo): find bugs at 235%/310% text size → fix first, claim after.

Output Format

Report findings as:

  • 🔴 Critical — blocks an assistive-tech user from completing a common task (unlabeled control on the purchase path, VoiceOver trap, text that doesn't scale)
  • 🟠 High — feature claim at risk (truncation at accessibility sizes, color-only state, missing captions)
  • 🟡 Medium — friction (bad reading order, unclear labels, missing rotors/actions)
  • 🟢 Low — polish (verbosity, missing synonyms, hint quality)
  • ✅ Strengths — passes worth keeping (cite the pass that proved them)

For each finding: the screen/task, the failing feature category, the fix (API-level), and which audit pass detects the regression.

Cross-References

  • generators/accessibility-generator — implementation patterns (labels, Dynamic Type, custom actions, rotors) to fix what the audit finds
  • macos/ui-review-tahoe/accessibility.md — Mac-specific VoiceOver/keyboard depth
  • ios/assistive-access — the separate Assistive Access experience (cognitive disabilities)
  • design/typography — Dynamic Type design rules
  • release-review — this audit slots into the pre-release gate

References

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts3mo3 checks · Risk SAFE
  • Gen Agent Trust Hub3mo

    The accessibility-audit skill is a comprehensive documentation set that guides developers through auditing iOS and macOS apps for accessibility. It covers automated XCUITest APIs, Accessibility Inspector workflows, and App Store Accessibility Nutrition Label requirements using official Apple guidance. No malicious patterns or security risks were detected.

  • Socket3mo

    No alerts

  • Snyk3mo

    Risk: LOW · No issues

Signed by skilld at f958982. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last week.

Activeupdated 4 months ago
What it can do
Reads files Runs commands
last_verified
2026-07-16
review_by
2027-06-22
All 4 allowed tools
ReadGlobGrepBash

README badge

README badge for rshankras/claude-code-apple-skills/accessibility-audit