Accessibility Pass logo

Accessibility Pass

Organization
microsoft
accessibility-pass

Use this skill whenever the user asks to check, review, or fix the accessibility of a PowerPoint deck, Word document, HTML page, or Markdown file, and before handing over any deck or document this agent just generated, so it does not ship with missing alt text, untitled slides, or unreadable colour contrast.

Overview

Publishermicrosoft
Repositorycat-agent-skills
Skill nameaccessibility-pass
Stars
95
Forks
115
Bundled files
2
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by microsoft on GitHub. Read the source before you install it.

Installation

Install the Accessibility Pass AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/microsoft/cat-agent-skills/tree/main/submissions/accessibility-pass
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/microsoft/cat-agent-skills.git /tmp/cat-agent-skills
mkdir -p .claude/skills
cp -r /tmp/cat-agent-skills/submissions/accessibility-pass .claude/skills/accessibility-pass
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Accessibility Pass in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Accessibility Pass on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Accessibility Pass is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Review the artefact against Microsoft's Accessibility Checker rules, report what fails, and fix it on request. Use the rule names below verbatim. They are the same ones the author sees when they run the checker themselves in Office, so the findings are recognisable and searchable.

Instructions

  1. Get the file. Supported: .pptx, .docx, .html, .md. If the user pasted raw HTML or Markdown instead, save it to a file first.

  2. Run the bundled checker when a Python environment is available, from the skill's own root directory (the folder containing scripts/), so the relative path resolves:

    bash
    python scripts/a11y_check.py <file> --json

    It needs python-pptx for .pptx and python-docx for .docx; HTML and Markdown need nothing beyond the standard library. If the environment has no Python, or the import fails: for HTML and Markdown, read the file directly and check it against the rules below by hand. For .pptx and .docx, that fallback doesn't really work, these are zipped Office XML, not something to eyeball as text, so instead say plainly that the file couldn't be checked, and ask the user to run Office's own Accessibility Checker (File > Info > Check for Issues > Check Accessibility) and share what it reports.

  3. Add the judgement calls the script deliberately does not make:

    • Alt text that exists but is useless. "image1.png", "chart", or a filename is a failure even though the attribute is populated. Read every alt string and judge whether someone who cannot see the image learns the same thing from it.
    • Wrongly decorative. Anything marked decorative that actually carries information is a failure the script cannot see.
    • Reading order that is flagged but fine. The script compares screen-reader order with visual top-left order; deliberate multi-column layouts trip it. Confirm before reporting.
    • Meaning carried by colour alone, such as red/green status dots or "the items in orange," which no file-level check can detect.
    • Captions and transcripts for embedded audio or video.
    • Contrast the script skipped. It only compares explicit RGB values. Theme colours, gradients, and picture fills come back unchecked. Report them as unchecked, never as passing.
  4. Report findings grouped by severity, worst first, one row each:

    SeverityRuleWhereFix
    ErrorAll slides have titlesSlide 4Add a title in the title placeholder

    Open with a one-line count (3 errors, 5 warnings, 1 tip). If nothing fails, say so and list what could not be checked mechanically.

  5. Offer to apply the fixes. Apply directly: adding a missing slide title, adding a table header row, rewriting vague link text, adding lang to <html>, fixing heading-level jumps. Ask first before writing alt text for an image whose content is uncertain, marking anything decorative, changing colours, or restructuring merged tables, since each of those changes meaning or design.

  6. Re-run the checker after fixing and report what is left.

The rules

Errors: content that is unusable for someone relying on assistive tech.

  • All non-text content has alternative text (alt text)
  • Tables specify column header information
  • All slides have titles
  • All sections have meaningful names (the bundled script doesn't check this; ask the user or inspect the file's sections manually)
  • Document access is not restricted (the bundled script doesn't check this; verify manually whether the file has IRM/password protection applied)

Warnings: content that is hard to use.

  • Sufficient contrast between text and background (WCAG AA: 4.5:1 normal text, 3:1 for text 18pt+ or bold 14pt+)
  • Table has a simple structure
  • The reading order of the objects on a slide presentation is logical
  • Closed captions are included for inserted audio and video
  • Hyperlink text is meaningful

Tips: content that could be easier to use.

  • Slide titles in a deck are unique
  • Documents use heading styles
  • Document language is set

Guardrails

  • Never invent alt text for an image whose content is unknown. Ask the user what it shows, or describe only what surrounding text supports and say the description needs confirming.
  • Never mark an object decorative to clear a finding. Decorative means it carries no information; use it only when that is true.
  • Never edit visible wording, layout, or branding beyond what a fix requires.
  • Never report an unchecked item as a pass.
  • Do not claim WCAG or EN 301 549 conformance. This is a pass over known rules, not a certification. A clean result means no detected issues.

Tone

Direct and specific. Name the rule, the location, and the concrete fix. No lecturing about why accessibility matters; the user already asked.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Accessibility Pass AI skill do?

Use this skill whenever the user asks to check, review, or fix the accessibility of a PowerPoint deck, Word document, HTML page, or Markdown file, and before handing over any deck or document this agent just generated, so it does not ship with missing alt text, untitled slides, or unreadable colour contrast.

Why use Accessibility Pass on TypingMind?

Because you install it once and use it with any model. Accessibility Pass is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Accessibility Pass in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/microsoft/cat-agent-skills/tree/main/submissions/accessibility-pass. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Accessibility Pass?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Accessibility Pass?

As many as you like. As long as a model supports skills, you can use Accessibility Pass with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Accessibility Pass AI skill free?

Yes. It is published on GitHub by microsoft under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇