agentleFS
DnsQueenDnsQueen/orcinus-skills/skills/accessibility/SKILL.md

accessibility

Skill1 starsChanged 5 days agoMITView on GitHub

Audit a web UI for accessibility with the ultra11y engine (RGAA 4.1.2 on a WCAG 2.2 AA core, 102 static checks), fix what it finds, and re-audit. Use before shipping any web UI change and in CI on every UI change. A page that was never audited is not accessible. Only invoke via the orcinus router.

123 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use author-contributions, tool-rename-deprecation and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • MicrosoftSkill

    author-contributions

    Identify all files a specific author contributed to on a branch vs its upstream, tracing code through renames. Use when asked who edited what, what code an author contributed, or to audit authorship before a merge. This skill should be run as a subagent — it performs many git operations and returns a concise table.

    Also in 7 projects that use this

  • gitpod-ioSkill

    tool-rename-deprecation

    Ensure renamed built-in tool references preserve backward compatibility. Use when renaming a toolReferenceName, tool set referenceName, or any tool identifier. Run on ANY change to tool registration code. Covers legacyToolReferenceFullNames for tools and legacyFullNames for tool sets.

    Also in 5 projects that use this

  • DepressionCenterSkill

    project-preferences

    Apply repository-specific preferences when planning, implementing, or reviewing changes in this project.

    Also in 6 projects that use this

  • MicrosoftSkill

    chat-perf

    Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.

    Also in 5 projects that use this

  • DepressionCenterSkill

    create-media-pack

    Create a coordinated project branding and media pack with three accessible design options, approved vector artwork, image exports, a branding guide, and README integration. Use when asked for a repository media pack, brand kit, or complete project visual identity.

    Also in 5 projects that use this

  • +7 more from microsoft

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 112 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every finding names its criterion, its location, and its remediation.

    line 15
  2. Very specific:

    Never declare conforming what was not verified.

    line 16
  3. Distinctive:

    Blocking / major: must be fixed before shipping.

    line 32

“Common” is measured: kept by at least 6 in 10 of the 112 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

ACCESSIBILITY — can everyone use this page?

An automated check sees only part of accessibility. This skill owns that part with a deterministic engine, and says plainly what the engine cannot see. A page that was never audited is not accessible, and a page whose audit was never read is not fixed.

Rules

  • Audit before shipping any web UI change. A blocking or major finding is a FAIL.
  • Every finding names its criterion, its location, and its remediation. A finding without a criterion is an opinion, not a result.
  • Never declare conforming what was not verified. The engine's check and verify gates exist to reject invented non-conformities — and invented conformities.
  • Fix, then re-audit. A fix that was not re-audited is a claim, not a fix.
  • A check that did not run is never reported as passed.

Steps, in order

1. Audit

python3 skills/accessibility/scripts/check_accessibility.py <page-or-glob> --out <dir>

The audit runs through the ORCINUS gate wrapper (scripts/check_accessibility.py), which calls the engine (node skills/accessibility/scripts/ultra11y.mjs audit, --fail-on major by default) after pre-checking every input. A missing input, a glob that matches nothing, or an audit of zero files is INCOMPLETE (exit 2) — the engine on its own warns and exits 0 with zero files audited and 100% conformance, which is never a pass. The engine is self-contained and install-free (Node only, no dependencies). It audits the page against RGAA 4.1.2, the French State's accessibility standard, on a WCAG 2.2 AA core; other country standards are pluggable packs. Read the JSON report it writes: each criterion carries a status (conforming, non-conforming, not applicable, manual), and each finding carries its rule, criterion, file, line, severity, message, and remediation.

2. Classify

  • Blocking / major: must be fixed before shipping. A missing main landmark, an image without an alternative, a form field without a label.
  • Minor: fix in the same change if it is cheap; otherwise record it in the report with a reason, not silently.
  • Manual: the engine cannot decide these (for example, whether an image alternative actually describes the image). A human must judge them — the engine saying "manual" is not a pass.

3. Fix

Fix one finding at a time, starting with the blocking and major ones. Prefer native HTML first: a real <button>, a real <label>, a real <main>. Do not fix an accessibility finding by adding ARIA where a native element would do — ARIA is a patch, not a design.

4. Re-audit and gate

Re-run the audit on the fixed page. In CI, gate on it:

python3 skills/accessibility/scripts/check_accessibility.py <page-or-glob> --fail-on major

The wrapper defaults to --fail-on major: it exits 1 when a blocking or major finding exists, and exits 2 (INCOMPLETE) when an input is missing or no file was audited — a missing page is never reported as passed. (The bare engine without --fail-on reports findings and still exits 0 — a gate that cannot fail is not a gate — which is why the audit always goes through the wrapper.)

5. Verdict

  • PASS: no blocking or major findings, and every manual criterion has a recorded human judgment.
  • FAIL: any blocking or major finding, or any manual criterion left undecided.

Report

Return the report in this form:

  • The exact engine command, its version, and its exit code.
  • Each finding, with criterion, file, line, severity, and remediation.
  • Each manual criterion and the human judgment recorded for it.
  • The fix applied per finding, and the re-audit result.
  • Verdict: PASS or FAIL, and what remains.

Scope

Static checks only. The engine checks the markup it can see; it does not see rendered behaviour — keyboard traps, focus order in a dynamic flow, or a live region that never announces. Those need a real browser (use webapp-testing) and a human with a keyboard and a screen reader. This skill does not claim a page is accessible; it claims the engine's checks passed and the manual criteria were judged.

Sources

Method, in our own words, using the ultra11y engine (maxgfr/ultra11y, MIT, licence verified from the primary LICENSE file) as-is. No code was copied from ultra11y except the engine script itself, which is bundled verbatim (see ATTRIBUTION.md for why the upstream SKILL.md was not copied verbatim).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.