Skip to content

Certify keyboard and screen-reader access for core setup workflows #179

Description

@Maneek21

Current state

Refreshed September 24, 2026 against master at 1427f66.

The production browser smoke and individual keyboard improvements do not establish end-to-end keyboard or screen-reader coverage. No core-path screen-reader certification was found in this review.

Evidence: browser smoke, roadmap.

Remaining work

Certify login, primary navigation, task creation, a governed agent approval, and personal MCP connection setup/revocation. Use the navigation and identity model of the tested revision.

Acceptance evidence

  • Complete each flow using only the keyboard; record revision, browser, and viewport.
  • Verify visible focus, logical tab order, semantic names/labels, and useful validation feedback.
  • Verify focus enters dialogs appropriately, remains contained where required, and returns sensibly on close.
  • Record a manual screen-reader pass with a named reader/browser pair, covering controls, errors, and completion feedback.
  • Add stable automated regression checks for discovered failures; document what still needs manual verification.
  • Inspect representative desktop and mobile layouts for these flows.
  • Link intentionally deferred findings with impact and rationale.

Scope

Core-flow certification, not a claim of complete WCAG conformance. Personal MCP connections act as the user and must not be presented as automatically using agent approvals.

Ownership and priority

Maintainer: @Maneek21. Prioritize within Next preview hardening.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

accessibilityAccessibility and inclusive UXhelp wantedExtra attention is needed

Projects

No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions