Skip to main content
IMPACT HACKERSACCESSIBILITY RECEIPT
ACCESSIBILITY RECEIPT / A11Y-2027-01

Access is part of the product.

This receipt shows what we tested, what automation can and cannot prove, and which checks still need a person using real assistive technology.

IMPACT HACKERS / ACCESSIBILITYACCESSIBILITY RECEIPT
v1.0
VERIFIED CODE STATE2026-08-03
SCOPELocal release candidate
TARGETWCAG 2.2 A + AA
01 / AUTOMATED EVIDENCE

What the release gate checks.

  • Every public page source is manifested or explicitly excluded.
  • Critical routes are checked for structure, accessible names, reflow, zoom, color modes, and WCAG A and AA rules.
  • Keyboard-only journeys cover diagnosis, application review, workshop selection, and artifact creation.
  • Accessible text downloads are checked for filename, media type, content, contract version, and evidence boundary.
  • Visual baselines cover mobile, tablet, desktop, dark mode, reduced motion, and forced-colors emulation.
02 / REAL ASSISTIVE TECHNOLOGY

A proxy is not a person.

We show incomplete evidence instead of turning an automated result into a broader claim.

Manual verification required

Safari 27 on macOS 27

Safari and safaridriver are available, but Safari Remote Automation is disabled on the verification machine.

No Safari result is claimed. Enabling persistent remote automation requires explicit owner approval, and a human journey is still required.
Manual verification required

VoiceOver with Safari

VoiceOver is available on the verification machine.

Automated ARIA snapshots are a semantic proxy, not VoiceOver output.
Automated proxy passed

Windows High Contrast

Chromium forced-colors emulation passes structure and visual regression checks.

A real Windows High Contrast run is still required.
Manual verification required

NVDA with Chrome or Firefox on Windows

No compatible Windows and NVDA environment is available in this workspace.

No NVDA result is claimed.
03 / COGNITIVE ACCESSIBILITY

Make the next move easier to understand.

01

Orient before asking

Name where the person is, what this step is for, and what happens next before requesting effort.

02

Ask for one decision at a time

Keep each control focused on one choice and separate explanation from the action label.

03

Make labels predict the result

Use specific verbs and objects so links and buttons still make sense when read out of context.

04

Make recovery obvious

Explain what needs attention, preserve prior work, and return focus to the place where recovery starts.

05

Separate evidence from promise

State what an estimate, recommendation, or artifact can show and what it cannot prove.

04 / EVIDENCE BOUNDARY

Passing checks is not the same as universal access.

Automated checks and browser emulation can find regressions, but they do not prove that every person, assistive technology, browser, or operating system can complete every task.

  • No automated tool can establish complete accessibility or usability.
  • The current screen-reader journey validates semantic orientation, not speech quality or user comprehension.
  • Production accessibility must be rechecked after deployment because the verified scope is local.
  • New customer content and third-party integrations can introduce issues outside this receipt.
STANDARDS + METHOD / 05

A stable standard and a living practice.

The release gate targets WCAG 2.2 A and AA. Cognitive guidance and real assistive-technology checks extend the practice beyond what one automated scan can establish.

WCAG overview Cognitive accessibility at W3C Explore the design system