Hire an AI website accessibility reviewer

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• Gaps by WCAG level, against last quarter
• At most ten fixes, task-blocking first, each with page, criterion and where it goes
• A verdict on each fix from last quarter, held or reopened
• Failing colour pairs and the closest passing colours
• Dated notes for your accessibility statement, with no over-claim
• A five-task human test to run with people who use assistive technology

How it does the job well

• Start from the user's task and the content, then the layout. Decoration comes last.
• Design every state: empty, loading, error, long text and mobile.
• Keep one visual system. Reuse spacing, type and colour tokens instead of one-off values.
• Check text contrast (at least 4.5:1 for body text) and keyboard focus as you go.

Mistakes it avoids

• Claiming a level because the automated checks passed.
• Fixing colour contrast across the site in one go without a designer.
• Ranking a criterion failure above a blocked task.
• Letting a redesign ship and checking it the next quarter.
• Designing only the happy path at desktop width.
• Copying a reference's logo, text or claims instead of learning from its structure.

What a good result looks like

An example, not your task: You fixed accessibility once and are not sure it held.

• This quarter. One line with gaps by level against last quarter, fixes that held and fixes reopened.
• Fix first. At most ten fixes, task-blocking first, each with page, criterion, level and where in the code.
• Re-checks. Each fix from last quarter, marked held or reopened with the reason.
• Colour pairs. Failing colour pairs with their ratio and the closest passing colour, if the palette changed.
• Government notes. Section 508 gaps and report notes, only if a government buyer applies.
• Statement notes. What is true now, dated, with known gaps and no conformance claim.

How the work is checked

• Every gap names a page and a WCAG criterion.
• Last quarter's fixes were re-checked before new ones were ranked.
• No more than ten fixes were proposed, task-blocking first.
• The statement notes make no conformance claim.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

A second opinion

Before it hands back, your AI reviews the draft the way Jakob Nielsen and Don Norman would. This is a way of reviewing the work. It does not mean they wrote or endorse it.

Capabilities

• Read last quarter. Lists every fix marked done, to re-check before anything new.
• Check the whole site. Finds WCAG gaps by level, what to fix first and notes for the statement, against last quarter.
• Check the pages that matter. Ranks problems on key and changed pages against WCAG 2.2, with fixes held as proposals, task-blocking ones first.
• Check the colours when they changed. Tests every light and dark pair and names the closest passing colour for each failure.
• Check for government buyers. Runs the Section 508 check and writes report notes only when a government buyer applies.
• Rank and plan. Merges the gaps, sorts by blocked tasks then level, and proposes at most ten fixes with their place in the code.
• Update the statement notes. Writes what is true now, dated, and leaves the conformance claim to you after a human test.
• Log and hand back. Saves the gaps and notes, then stops.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id quarterly-accessibility-keeper only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/quarterly-accessibility-keeper/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• Your site address and the pages where people buy, sign up, pay or get help
• Your code or templates, so fixes can be placed (advised)
• Your colour tokens or design direction (advised)
• What changed since last quarter (advised)
• Whether any buyer is a government body (optional)

Helpful context (optional)

• Support tickets that mention a screen reader, keyboard or colour
• Your statement as it stands today, so the notes update it rather than start over

Tools and access

• Claude Code: It can read the theme or app code, keep the gaps file between quarters and run on a schedule.
• A screen reader: For the human test. Use the one your users use most on the device they use most.
• Claude Code: It can read your project files, run the script and apply the fix once you agree.
• Any AI that can run Node: The script needs Node 18 or newer. A plain chat AI can do the hand-worked version.

Functions and deliverables

• Website accessibility reviewer: Review key pages for access barriers. Recheck earlier fixes and get ranked issues, statement notes and tasks for human testing.
  • Gaps by WCAG level, against last quarter
  • At most ten fixes, task-blocking first, each with page, criterion and where it goes
  • A verdict on each fix from last quarter, held or reopened
  • Failing colour pairs and the closest passing colours
  • Dated notes for your accessibility statement, with no over-claim
  • A five-task human test to run with people who use assistive technology
• Accessibility standards checker: Check your site against WCAG accessibility rules. Give it your page code or screenshots. You get gaps by level, ranked fixes and draft statement notes. Not legal advice.
  • A gap table with status for each criterion
  • Fixes ranked by priority
  • Annotated code with corrections
  • Draft notes for an accessibility statement
  • A map from rules to laws
  Source: Hemant Naik.
• Web accessibility checker: Check a web page or code for WCAG 2.2 accessibility problems. Get a ranked list with where each one is, and fixes if you ask.
  • A ranked list of WCAG 2.2 problems
  • The place in the page or code for each
  • Fixes applied and re-checked, if you asked
  • A note on what automated checks cannot cover
  Source: sickn33.
• Colour contrast checker: Share your colour tokens. Get light and dark theme pairs checked for readability, with the closest passing colours.
  • A pass or fail line for each text and background pair, with its ratio
  • Separate results for light and dark mode
  • The closest passing colour for each failure, with the same hue
  • One value per token that passes on every background
  • A list of where each token is used, and what to check by eye
  Source: sickn33.
• Federal accessibility checker: Share federal site, app, or document details for Section 508 gap guidance and ranked fixes.
  • Gaps tied to WCAG criteria
  • A priority fix plan
  • Draft conformance report remarks
  • Procurement wording if needed
  Source: Hemant Naik.
• Brand profile builder: Share your website. Get a reusable profile covering your products, customers, voice and approved claims, with sources.
  • A folder of eight plain files, from a start file for your AI to a claims list and a change log
  • A tag on every fact, and every claim marked approved, pending or banned
  • A two-sentence diagnosis of your biggest gap, and a list of what's missing and who can answer it
  • A short slice for your next step, if you name one
• Jakob Nielsen: Review a website, app or form through Jakob Nielsen’s recorded approach. Find likely usability problems and plan small rounds of realistic user tasks.
  • Your problem worked on the way Jakob Nielsen would: their questions, their method, their push-back
  • The answer or the work, in Jakob Nielsen's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step
• Don Norman: Find why a product, form or service is hard to use through Don Norman’s public approach. Check cues, controls, feedback and recovery from mistakes.
  • Your problem worked on the way Don Norman would: their questions, their method, their push-back
  • The answer or the work, in Don Norman's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step

Tasks you can use it for

• You fixed accessibility once and are not sure it held.
• A redesign or brand refresh just shipped.
• A buyer or a regulator asked for your accessibility statement and you do not have a current one.

Limits to check before starting

• You need a one-off check of a single page: run the standards checks by hand in your AI, or ask a developer to review that page.
• You need a formal conformance claim or a legal opinion: hire an outside accessibility auditor. This agent never claims a level.
• You want the fixes made for you: ask your developer to apply the list. This agent only proposes changes.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Start a chat with the agent. In Claude Code, add it as a skill or save it in .claude/agents, then ask it to run this quarter.
• Give it the site address and your key pages. Point it at the code and the colour tokens if you can.
• Say what changed since last quarter.
• Read the fix list. Make the changes you agree with, then tell the agent which were done.
• Run the human test with five users. Run the agent again next quarter; it re-checks what you fixed.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/quarterly-accessibility-keeper for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.