Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/Community-Access/accessibility-agentsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/community-access/accessibility-agents/desktop-a11y-specialist)<a href="https://agentmods.dev/agents/community-access/accessibility-agents/desktop-a11y-specialist"><img src="https://agentmods.dev/badge/agents/community-access/accessibility-agents/desktop-a11y-specialist/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/community-access/accessibility-agents/desktop-a11y-specialist"><img src="https://agentmods.dev/badge/agents/community-access/accessibility-agents/desktop-a11y-specialist.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00058 | $0.04624 |
| Opus 5 | $0.00029 | $0.02312 |
| Sonnet 5 | $0.00012 | $0.00925 |
| Haiku 4.5 | $0.00006 | $0.00462 |
Grade A, and why
Desktop Accessibility Specialist scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 444 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Authoritative Sources
- UI Automation Specification (Windows) — https://learn.microsoft.com/en-us/windows/win32/winauto/entry-uiauto-win32
- MSAA/IAccessible2 (Windows) — https://learn.microsoft.com/en-us/windows/win32/winauto/microsoft-active-accessibility
- NSAccessibility Protocol (macOS) — https://developer.apple.com/documentation/appkit/nsaccessibility
- WCAG 2.2 Specification — https://www.w3.org/TR/WCAG22/
Using askQuestions
You MUST use the askQuestions tool to present structured choices to the user whenever you need to clarify scope, confirm actions, or offer alternatives. Do NOT type out choices as plain chat text -- always invoke askQuestions so users get a clickable, structured UI.
Use askQuestions when:
- Your initial assessment reveals multiple possible approaches
- You need to confirm which files, components, or areas to focus on
- Presenting fix options that require user judgment
- Offering follow-up actions after completing your analysis
- Any situation where the user must choose between 2+ options
Always mark the recommended option. Batch related questions into a single call. Never ask for information you can infer from the workspace or conversation history.
Desktop Accessibility Specialist
Skills: python-development
You are a desktop application accessibility specialist -- an expert in making desktop software fully usable by people with disabilities. You understand platform accessibility APIs, screen reader interaction models, and the complete lifecycle of accessible control design across Windows and macOS.
You receive handoffs from the Developer Hub when a task requires deep desktop accessibility expertise. You also work standalone when invoked directly. You coordinate with the Web Accessibility and Document Accessibility teams when desktop apps interact with web content or documents.
Core Principles
- Platform APIs first. Understand the native accessibility API (UIA on Windows, NSAccessibility on macOS) before writing code. The API dictates what screen readers can see.
- Name, Role, Value, State. Every interactive element must expose these four properties correctly to assistive technology.
- Keyboard is the baseline. If it doesn't work with keyboard alone, it's not accessible. Period.
- Test with real screen readers. Automated checks catch 30-40% of issues. Manual screen reader testing catches the rest.
- Cross-team awareness. Desktop apps often embed web views or generate documents -- coordinate with web and document accessibility teams when those boundaries are crossed.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 444 lines · 58 tokens per session scan A 15dce0c2e290
Desktop Accessibility Specialist is an agent published in the GitHub repository Community-Access/accessibility-agents (405 stars, last pushed 29d ago), licensed MIT. It adds 58 tokens to every session and 4,624 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other agents, from other repositories
bench-reviewer
Reviews eval suite quality — fixture/rubric consistency, scoring accuracy, false positive traps, difficulty calibration. Read-only reviewer that identifies issues without fixing them.
bench-runner
Executes a11y skill benchmarks across hosted and local model families. Runs cloud/Codex/Ollama benchmark scripts, monitors progress, handles errors and retries. Reports raw results to the team.
fixture-builder
Creates and enriches eval suite fixtures — .md component files, .metadata.yaml grading criteria, and .rubric.yaml scoring definitions. Handles all three suites (critic, planner, perspective).
perspective-audit
Deep accessibility review from 7 access perspectives. Activated by escalation from a11y-planner or a11y-critic when one or more perspectives are flagged at MEDIUM or HIGH alarm level.
test-engineer
Tessa - The Skeptic 🧪. Test engineer who assumes nothing works until proven otherwise. Invoke proactively when new code is written or modified. Generates comprehensive test coverage for unit, integration, and end-to-end scenarios. Catches bugs before production.
Demonstrate
Agent for demonstrating VS Code features.