Praxity Check v0.2.1 (alpha)

Praxity Check v0.2.1 fixes baseline comparisons for assessment controls whose generated IDs change on each page load.

Fixed

• Generated fragments in rs-choice, rs-match, rs-cat, rs-hs, and rs-matrix IDs no longer change occurrence identity.
• Existing v0.2.0 schema-v4 reports remain valid baselines. Praxity Check accepts their original occurrence IDs and re-keys them internally.
• A replay of the affected Praxity Studio report now produces 0 new, 224 existing, and 0 resolved occurrences instead of 10 false new and 10 false resolved occurrences.

You can continue using an existing v0.2.0 baseline. The JSON report schema remains version 4.

Praxity Check v0.2.0 (alpha)

Praxity Check v0.2.0 can inspect interface states reached through user actions and compare the results with an earlier report. Reports now identify checks and page areas that could not be tested.

New

• The project and CLI are now named Praxity Check. Use praxity-check instead of prax-audit.
• Use --baseline with a previous schema-v4 report to mark current results as new or existing and list results that have been resolved.
• Reviewed baseline results can be marked accepted or falsePositive with a reason, owner, and review date. These decisions remain visible and do not suppress findings.
• Use --scenarios to inspect content revealed by clicks, keyboard actions, selections, and waits. Scenario files cannot execute arbitrary JavaScript.

Improved

• JSON reports now include stable IDs for individual findings and review questions, rule metadata, ACT rule IDs where available, browser details, and explicit records of checks that could not run.
• Package checks now report missing local images, audio, video, documents, stylesheets, fonts, and resources referenced from CSS.
• Cross-origin frames that cannot be inspected now appear as untested. Reports preserve their provider, title, URL, and the reason inspection was unavailable.
• Visual checks now cover declared dark colour schemes and 200% operating-system text scaling on pages that opt in.
• Interaction review now covers keyboard reordering, custom checkboxes, sliders, and additional tooltip and Escape-key behaviour.

Fixed

• Removed random local-server ports from report evidence so unchanged scans produce stable results.
• Kept selectors stable when generated component IDs change during hydration.
• Prevented arrow-key scrolling checks from triggering page-level navigation shortcuts.
• Stopped the Chromium default focus ring from being reported as author-defined focus styling.

For contributors

• Added a repeatable benchmark covering 40 defect-and-clean fixture pairs and eight selected ACT examples.

Compatibility

JSON reports now use schema v4. Existing findings and needsReview fields are preserved, but integrations restricted to schema v3 need updating.

Requirements: Node 22.18+, pnpm 11.5.3, and Chromium installed through Playwright.

Praxity Audit v0.1.0 (alpha)

Praxity Audit is ready for public alpha testing.

Included:

• automated local checks for keyboard use, focus, contrast, layout, motion, audio, image alternatives, and link names
• experimental interaction evidence for a separate human-approved LLM review
• experimental, opt-in VoiceOver evidence on macOS
• local folder and zip input with bounded extraction and blocked ordinary web requests by default
• community source licensing and private security reporting

Requirements: Node 22.18+, pnpm 11.5.3, and Chromium installed through Playwright.

Interaction and VoiceOver review remain experimental. Audit only packages you trust, and review reports before sharing them.