Hire an AI website quality improvement reviewer

Everything it can hand back

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

• Scores and counts against last run, with new, old and fixed problems apart
• An alarm line if speed, access or indexing got clearly worse
• One ranked fix list, each with its gain and effort
• A list of the checks it could not run, and what would let it run them

How it does the job well

• Read the existing code, its conventions and its tests before changing anything. Match the project's patterns.
• Make the smallest change that fully solves the request. Keep unrelated refactors out.
• Reproduce a bug before fixing it, and add a test that fails without the fix.
• Treat input from users, files and networks as untrusted. Validate it at the boundary.

Mistakes it avoids

• Calling a page fast or accessible from the code alone. Say what was and was not checked.
• Raising an alarm on a small change in a lab score. Speed scores move a little from run to run.
• Fixing many things in one release, so nobody can tell which one broke something.
• Treating a clean scan as proof that people with disabilities can use the site. It finds code problems, not every real barrier.
• Claiming code works without running it or its tests.
• Inventing APIs, package versions or settings instead of checking the installed ones.

What a good result looks like

An example, not your task: You run a site that changes often and want the same quality checks every week.

• This run. Speed, access, indexing and broken link scores against last run, with counts of new, old and fixed problems.
• Alarm. What got clearly worse, such as speed, a new critical access problem or a lost page, or none.
• Fix list to approve. Each fix with its page, bucket, gain, effort and the check that found it, most important first.
• Not measured. Each check it could not run and what you can give it to run it.
• Watch next run. The one to three numbers that decide what the next run looks at first.
• Files and missing. Each step's saved file, then every MISSING line in one list.

How the work is checked

• Every number is shown against the last run, or the first run is named as the base.
• Speed numbers come from a report, not a guess. If there was no report, the speed section says so.
• New problems are separate from old ones.
• No fix is marked done unless a recheck showed it gone.
• 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 Peep Laja would. This is a way of reviewing the work. It does not mean they wrote or endorse it.

Capabilities

• Load the memory. Reads last run's scores, the open problems with their dates and the problems you chose to leave.
• Check speed. Reads each key page and your speed report for the score, the three Core Web Vitals and the main cause.
• Check access. Scans your code for access problems and counts them by severity. Fixes wait for your yes.
• Check tags and search. Finds missing titles and share cards, pages not indexed, broken links and redirect chains.
• Compare and choose the path. Marks each problem new, old or fixed. A clear speed drop, a new critical access problem or a lost page goes to the top.
• Sort and hand back. Puts every finding in one bucket, ranks by gain and effort, and stops for your yes. After fixes it rechecks, up to two times.

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 website-quality-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/website-quality-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

• The addresses of your home page and 3 to 8 pages that matter most
• A PageSpeed Insights or Lighthouse report for each of them (best)
• Your site code, if you want the access scan (optional)
• A Search Console Pages export (optional)

Helpful context (optional)

• A list of releases with dates
• Your top pages by visits

Tools and access

• Claude Code: It can keep the folder, read your site code and run after each release.
• PageSpeed Insights: Free reports for each key page, with the scores and Core Web Vitals the agent needs.
• Google Search Console: The Pages report shows which pages are not indexed and why.
• Claude Code: Reads your whole project and makes small, scoped edits to the right files.
• Any good chat AI: Works well if you paste in the head section of a few pages.
• Any good chat AI: Sorting and ranking a list with reasons is text work any strong chat AI does with this skill.
• A spreadsheet or task board: Holds the ranked list and tracks what is done.

Functions and deliverables

• Website quality improvement reviewer: Review key pages for speed, access and search issues. Compare the last check and get ranked fixes with untested areas named.
  • Scores and counts against last run, with new, old and fixed problems apart
  • An alarm line if speed, access or indexing got clearly worse
  • One ranked fix list, each with its gain and effort
  • A list of the checks it could not run, and what would let it run them
• Web performance checker: Share a slow page. Get ranked fixes for network, loading and rendering, with a score.
  • A ranked list of fixes with expected gain
  • Changes for network, loading and rendering
  • A score on eight checks
  • Config or code changes
  Source: Wondel.ai.
• WCAG accessibility reviewer: Share site code or pages and target WCAG level. It checks selected issues and suggests fixes; retest results depend on tools and manual review.
  • A list of issues by severity
  • Before and after code fixes
  • A re-check after the fixes
  • A report for managers
  • A pipeline check to stop new issues
  Source: Alireza Rezvani.
• Page metadata fixer: Share page code and preferred URLs. Get metadata fixes and a list of checks, including live social preview checks when available.
  • An audit table of missing, duplicated and clashing tags
  • The smallest fix for each problem, as the exact tag or line
  • A check that title, description, canonical and share address agree
  • A list of what could not be checked. If your site is offline, check the share cards after launch
  Source: sickn33.
• Page indexing checker: Share your site, folder or Search Console export. Get why pages are not indexed and a fix list.
  • A list of indexing problems and their causes
  • Orphan and deep pages found
  • Sitemap fixes
  • A redesigned structure and redirect plan
  Source: sickn33.
• Page speed fixer: Share PageSpeed or Lighthouse results and project access. Get ordered fixes, risk notes and a checklist for fresh test results.
  • Ranked list of issues by risk
  • Code fixes in ordered batches
  • Re-test after each batch
  • Notes on what changed
  Source: sickn33.
• Page issue sorter: Sort a long list of page problems into fix now, investigate, test or tracking, then rank them. Get a clear plan for your page issue list.
  • Each issue in one bucket with its reason
  • Test ideas written as: change, expected effect, why
  • A ranked order by likely gain and effort
  • A list of issues dropped for lack of proof
• AI project context organiser: Share your project and next task. Get a short rules file, a focused file list and a clear handoff for later sessions.
  • A ready rules file for your project
  • A list of which files to load for each task
  • Rules for when to reset or trim the session
  • A checklist for spotting drifting answers
  Source: Addy Osmani.
• Peep Laja: Review page messaging or an A/B test using Peep Laja’s recorded approach. Start with buyer evidence, clarity and a sound comparison plan.
  • Your problem worked on the way Peep Laja would: their questions, their method, their push-back
  • The answer or the work, in Peep Laja'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 run a site that changes often and want the same quality checks every week.
• You just released something and want to know what it broke.
• Search traffic or page speed dipped and you want the likely cause ranked.

Limits to check before starting

• You want a one-off check of a single page: run the WCAG accessibility reviewer or the Page metadata fixer on its own, which is faster.
• You want the agent to change your live site by itself: it never does, so ask a developer or approve each fix first.
• You have no key pages or speed reports to share: start by running PageSpeed Insights on your home page, then come back.

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 check your site.
• Give it your key pages, and attach your speed reports and code if you have them.
• The first run sets the base. From the second run on it says what changed.
• Read the fix list. Say yes to the fixes you want, and it makes only those, then rechecks them up to two times.
• Run it again next Monday and after each release.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/website-quality-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.