Standards coverage

Every success criterion of WCAG 2.2 - all 86 of them, up to Level AAA - plus ARIA conformance and Silktide's own best-practice checks, and how Silktide covers each one.

This page is about testing websites. For PDF documents, see PDF standards coverage.

How much does Silktide cover?

No automated tool - Silktide's or anyone else's - can test all of WCAG. Most success criteria hinge on human judgment: whether alternative text is appropriate, whether a page works with a real keyboard, whether an error message actually helps. What automation can do is find the definitive failures, flag the likely ones, and make sure the judgment calls are never silently skipped. That is how this page is organized - three separate columns, never rolled into one inflated percentage.

Coverage levels

• ✅ Machine-checked - Silktide finds these failures automatically and reports them as errors. High confidence; fix them without second-guessing the tool.
• ☑️ Machine-assisted - Silktide flags likely failures as warnings. The evidence is real but the verdict needs a person.
• 🔍 Guided review - automation cannot decide this criterion. Silktide detects the relevant content (a video, a form, a time limit) and prompts a structured manual review so it is never silently skipped.
• ❌ Not covered - Silktide has no check for this criterion today.

Where a criterion has checks at several levels, the table shows the strongest. Follow the check links for exactly what each one tests, and its limits.

WCAG 2.2 coverage

Each criterion links to the official W3C explanation. Since is the WCAG version that introduced it - testing against WCAG 2.2 includes everything from 2.0 and 2.1.

WCAG 2.2 removed 4.1.1 Parsing as obsolete. Silktide still runs its former checks - Duplicate IDs and Semantic lists - because the underlying problems still break assistive technology, and because they still count when you test against WCAG 2.0 or 2.1.

Reading this honestly

The caveats behind the table:

• A ✅ or ☑️ row covers that criterion's machine-detectable failures, not the whole criterion. 2.1.1 Keyboard is the clearest example: Silktide flags elements with mouse-only event handlers, but proving a whole page works with a keyboard requires using one. The same applies to 1.4.1 Use of Color, where Silktide tests link contrast but cannot judge a chart that relies on color alone.
• Quality judgments are prompted, not decided. Silktide reliably finds an image with missing alternative text (error). Whether existing alt text is appropriate is a review prompt, not a machine verdict.
• Guided review is a workflow, not a test result. A 🔍 criterion passes no audit by itself. Its value is completeness: every applicable criterion appears in your Guidelines view with something recorded against it - an automated result, a review decision, or an open prompt - so nothing falls through the cracks. Testing that still needs a person is what a manual audit is for.
• Silktide's web accessibility checks do not yet use AI judgment. The checks here are deterministic. When that changes, this page will change.

What we don't cover yet

Every WCAG 2.2 success criterion - all 86, Levels A through AAA - has at least a guided review in Silktide. There is no uncovered criterion today.

That does not mean every criterion is fully machine-testable. The six Level AAA criteria that previously had no check (1.3.6, 2.1.3, 2.3.3, 2.4.8, 2.4.9, 2.4.10) are tracked as page-level reviews because no mainstream tool - including Silktide - can decide them automatically. Follow each check's help page for exactly what a human must verify.

If a criterion's automated coverage is thinner than you need, tell us - the table above is exactly what we use to prioritize deeper checks.

ARIA coverage

Silktide tests ARIA 1.2 usage separately from WCAG, because broken ARIA can make a page less accessible than no ARIA at all. Most ARIA rules also count toward a WCAG criterion (usually 4.1.2 or 1.3.1), so these checks appear in both views.

Best practices beyond the standards

Some things hurt real users without failing any WCAG criterion. Silktide tests these at its own discretion; they contribute to the Best practice view of your Accessibility score, not to WCAG conformance:

Silktide check	Coverage	What it catches
Accessibility overlay	☑️ Machine-assisted	Third-party overlay widgets that claim to fix accessibility and routinely make it worse.
Title attribute labeling	☑️ Machine-assisted	Controls labeled only by title attributes, which most assistive technology handles poorly.
Small desktop text	☑️ Machine-assisted	Body text rendered too small to read comfortably, even where contrast passes.
Icon-only control	☑️ Machine-assisted	Buttons and links that are only an icon, with no accessible name.
Modal focus trap	☑️ Machine-assisted	Modal dialogs that fail to trap focus while open or restore it on close.

PDFs on your website

PDFs found while testing your website get their own accessibility checks (tagging, reading order, contrast, titles - the PDF-prefixed rows in the WCAG table above). See PDFs for how they surface in the Accessibility module, and PDF standards coverage for the full PDF/UA-1 picture, including AI-assisted remediation.

Related

• Accessibility guidelines - this same coverage, applied to your website with evidence per requirement.
• Manual audits - ask Silktide's specialists to test what automation cannot.
• Checks - every Silktide check, with its standard.
• PDF standards coverage - the equivalent page for PDF documents.