Accessibility Fix Desk: safe fix rules and a ten-check page scan (free sample)
by PaperChaseWebb
From Accessibility Fix Desk: the mistakes agents make when they fix accessibility, eight rules to paste into your agent, and a read-only script that runs ten checks on one HTML page, from language, title, alt text and unnamed links and buttons to unlabeled fields, blocked zoom, tabindex, removed focus outlines and overlay scripts.
- Price
- Free
- Difficulty
- Beginner
- Setup
- 5 min
- Format
- .md file
accessibility-fix-desk-safe-fix-rules-ten-check-page-scan-free-sample.md · 9 KB · direct download, no signup
What you end up with
You can see whether a page, or an agent's fix of it, still has the most common checkable accessibility failures before it goes live.
What you get
- Why agent accessibility fixes need checking
- Eight rules for agents that fix pages
- a11y_mini.py: ten read-only checks for one HTML page
- A made-up home page with eight findings and a clean fixed copy
Who it is for
Anyone who wants a quick check on a page, or on an agent's fix of it, before it goes live.
Works with
Solid chips: the agent is named in this listing's own tools list. Dashed chip: Plain Markdown, so it also pastes into other agents such as Grok Bot, Hermes, OpenClaw or Muse. We list those as compatible by format only, not as tested.
Tools needed
Cost to run: Free.
Time to result: About five minutes to a first scan
Sample prompts, free to use
Run python3 a11y_mini.py on [fixed page]. Fix every finding without deleting or hiding anything, or hand it to me. Do not edit the site.
The rules and script are in the file.
Free preview of the file
# Accessibility Fix Desk: safe fix rules and a ten-check page scan (free sample) Version 1.0, 2026-11-01. By PaperChaseWebb. For anyone letting an AI agent fix accessibility problems in web pages. One page and one Python 3 script, standard library only, read-only: it reads one HTML file, prints findings and changes nothing. No API key, no login. Tested on Linux with Python 3.13, 3.12, 3.11, 3.9 and 3.8: 4 unit tests, and 110 of 110 planted faults caught. The example business is made up. ## Why agent fixes need checking The common failures are small: an image with no alt, a link or button with no name, a field with only a placeholder, no page language, zoom turned off, the focus outline removed. In the WebAIM Million 2026 study, 95.9% of home pages had at least one. An agent can fix these fast, and it gets them wrong in the same few ways: - It deletes the icon link it could not name, or hides text with aria-hidden so a warning goes away. - It invents alt text, often stuffed with marketing words. - It puts an aria-label on a button that does not match the words on the button. - It adds an overlay script and calls the site accessible. - It writes that the site is "fully compliant". ## Rules to paste into AGENTS.md, CLAUDE.md, a Grok Bot standing order or Hermes notes 1. Never publish or edit the live site. Write a fixed copy and stop. The owner puts it live. 2. Never delete or hide content to clear a finding. 3. Never add scripts, overlay widgets or inline event handlers.
Test record
- Date
- Nov 1, 2026
- What was run
- tests/test_mini.py (4), 110 planted faults against a11y_mini.py, and the copy extracted from the sample page. Linux, Python 3.13, 3.12, 3.11, 3.9 and 3.8.
- Result
- 4 passed, 0 failed on all five versions; 110 of 110 faults caught; the extracted copy was identical to the tested script and found 8 problems on the example page and 0 on its fix.
- Not run
- Contrast, the fix check, the statement check and the eval set (the full kit does those).
Copied from this item's dated test notes. What was not run stays in.
Worth knowing
- The sample checks one page on its own. Contrast, headings, ARIA, the fix check against the original, the statement check, the eval set, owner approval and agent setups are in the full kit.
- Not a full audit and not legal advice. It does not say a page meets any law or standard.
- On 2026-11-01 a11y_mini.py passed 4 unit tests and 110 of 110 planted faults on Linux with Python 3.13, 3.12, 3.11, 3.9 and 3.8, and the copy extracted from the sample page was identical to the tested script.
Questions
Does it change my files or site?
No. It reads the file you name and prints findings. It opens no network connection and changes nothing.
What is in the full kit?
A 20-check scan with contrast, a 15-check fix check against the original page, a statement check, owner approval at a terminal, a 10-page eval set, and setup for Grok Bot, Claude Code, Codex and Hermes. $29.
Details
Version 1.0 · Updated Nov 1, 2026 · Markdown version of this page
License: Personal and commercial use. Download terms
Related
Accessibility Fix Desk: agent fixes for common WCAG 2.2 failures, checked against the original, owner approved
Let an agent fix the accessibility failures found on almost every site: missing alt text, unnamed links and buttons, unlabeled fields, faint text, blocked zoom, hidden focus. It writes fixed copies, never the live page. A fix check compares each copy with the original and refuses deleted or hidden content, new scripts and overlays, alt text you never wrote, colors outside your palette and aria-labels that drop the visible words. A claims check stops "fully compliant" statements. You approve at your terminal and put the page live yourself.
Search Console Fix Desk: safe title rules and a ten-check page title and description scan (free sample)
From Search Console Fix Desk: the mistakes agents make when they rewrite page titles, eight rules to paste into your agent, and a read-only script that runs ten checks on one title and description, from length, stuffing and invented numbers to best and #1 claims, ranking promises, clickbait and indexing changes.
Review Reply Desk: safe reply rules and a ten-check review reply scan (free sample)
From Review Reply Desk: why review replies need checking, eight rules to paste into your agent, and a read-only script that runs ten checks on a reply draft, from public offers and requests to change a review to last names, invented numbers and links you did not approve.