accessibility-skills
Static, dependency-free WCAG 2.2 code review for web markup (HTML, JSX, Vue, Svelte, Astro) and CSS. Checks source against the a11yproject.com checklist (mapped to WCAG 2.2 success criteria) and grades each finding by evidence basis (verified / flagged / human-required) and severity (critical/serious/moderate/ minor). No browser, no npm CLI, no network access — pure Read/Grep over source, so it works in any repo immediately. Use for "review this component for accessibility", "check this page/PR for a11y issues", "is this markup WCAG compliant", or before merging a UI change. It cannot verify computed contrast, actual keyboard/screen-reader behavior, or anything needing a rendered page — those get flagged for a real scanner (e.g. axe-core) or human/AT testing, not guessed at.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-15
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- accessibility-skills
- license
- Apache-2.0
- metadata
- { "version": "0.1.0" }
- description
- Static, dependency-free WCAG 2.2 code review for web markup (HTML, JSX, Vue, Svelte, Astro) and CSS. Checks source against the a11yproject.com checklist (mapped to WCAG 2.2 success criteria) and grades each finding by evidence basis (verified / flagged / human-required) and severity (critical/serious/moderate/ minor). No browser, no npm CLI, no network access — pure Read/Grep over source, so it works in any repo immediately. Use for "review this component for accessibility", "check this page/PR for a11y issues", "is this markup WCAG compliant", or before merging a UI change. It cannot verify computed contrast, actual keyboard/screen-reader behavior, or anything needing a rendered page — those get flagged for a real scanner (e.g. axe-core) or human/AT testing, not guessed at.
- allowed-tools
- Read, Grep, Glob, Bash
- argument-hint
- [file|directory|diff] [--level AA|AAA]
02 资源文件 10
- 参考文档.claude-plugin/plugin.json705 B
- 参考文档evals/accessibility-skills-does-not-certify-compliance/case.yaml3.2 KB
- 参考文档evals/accessibility-skills-reviews-component/case.yaml4.8 KB
- 参考文档evals/accessibility-skills-reviews-stylesheet/case.yaml3.1 KB
- 参考文档evals/accessibility-skills-stays-quiet-on-aria-question/case.yaml2.3 KB
- 参考文档evals/accessibility-skills-stays-quiet-on-render-perf/case.yaml2.7 KB
- 许可证LICENSE11.1 KB
- 参考文档NOTICE.md3.1 KB
- 参考文档references/checklist.md12.1 KB
- 参考文档references/grading.md5.2 KB
03 SKILL.md 正文
Adapted from The A11Y Project (checklist
content, Apache-2.0) and AccessLint
(evidence-basis/severity grading approach, MIT). See NOTICE.md for
provenance.
This skill locates and grades; it doesn't fix (hand findings to the user or a follow-up edit) and it doesn't scan a live page (hand that to axe-core or a browser-based tool). Say which you did and didn't do.
Reference material
references/checklist.md— the WCAG 2.2 checklist items, organized by content area (images, headings, forms, color contrast, etc.), each mapped to its success criterion. Grep it for a keyword rather than reading it end to end.references/grading.md— the evidence-basis (●/◐/○) and severity grading rules, adapted for what static source alone can support. Read it before writing findings; the rules that always apply are summarized below.
Grading, in brief
Every finding gets two independent grades:
- Evidence basis: ● verified (a deterministic fact from source that's also a
real conformance failure: a missing attribute, wrong element, positive
tabindex) · ◐ flagged (real evidence, but confirming the failure needs computed styles, a rendered page, or a judgment call — a heading-rank skip orautofocusis a verifiable fact but not an automatic failure, seereferences/checklist.md's Advisory section) · ○ human-required (needs a screen reader or lived experience). - Severity: critical (blocks a core task) · serious (major barrier) · moderate (friction, still completable) · minor (polish).
When unsure between two evidence grades, use the lower one. Full definitions and
worked examples are in references/grading.md.
Procedure
- Scope.
$ARGUMENTSis a file, directory, or diff (git diff, a PR's changed files) to review, plus an optional--level AA|AAA(default AA — AAA items in the checklist are still worth flagging as informational at AA). Given nothing, ask what to review rather than sweeping the whole repo unprompted. - Read the target. For a diff, review the changed lines plus enough
surrounding markup to judge structure (a changed
<div>needs its parent to judge heading/landmark context). For a directory, find markup/template files via Glob rather than assuming a framework. - Check against the checklist. Work through
references/checklist.mdby category, but only the categories relevant to what's actually present in the target (no<table>in the diff → skip the Tables section; note it as N/A, not silently dropped). Grep the source for the syntactic signals each item implies (missingalt=,<div onClick, un-labeled<input, skipped heading levels,autofocus, positivetabindex, literal low-contrast color pairs) rather than reading every line by eye. - Grade each finding per
references/grading.md. Cite file:line and the exact snippet for every finding; never fabricate a location. Computed contrast (even a literal color pair like#999on white) andautofocusalways go under Flagged, never Verified; only a missing attribute, wrong element or positivetabindexis Verified. - Report.
# Accessibility review — <target>
WCAG 2.2 Level <AA|AAA> · static source review (no browser, no live scan)
Severity: <c> critical · <s> serious · <m> moderate · <n> minor
Basis: ● <v> verified · ◐ <f> flagged · ○ <h> human-required
Checklist categories reviewed: <list> · N/A (not present in target): <list>
## ● Verified
- [severity] <what> — SC <x.x.x>
where: <file>:<line> — `<snippet>`
fix: <mechanical change> | TODO(<SC>): needs judgment (<what a human decides>)
## ◐ Flagged
- [severity] <what> — SC <x.x.x>
where: <file>:<line> — `<snippet>`
confirm with: <specific check — e.g. "axe-core color-contrast rule", "render at 200% zoom">
## ○ Human-required
- <what only AT or lived experience reveals> — SC <x.x.x>
needs: <functional ability + assistive technology>
where: <file>:<line> (context, not proof)
## Not covered by this review
- Computed contrast, actual keyboard/screen-reader behavior, anything needing a
rendered page — recommend axe-core/Lighthouse and human/AT testing for these.
Copy the Checklist categories reviewed: line verbatim, label included; do not
shorten it to Categories reviewed:.
Spend the report's words on findings and handoffs; a clean category is its name in "reviewed," not a paragraph. Don't let "not present in target" (N/A) or "not covered by this review" read as "passed" — they're honestly different things.
Notes
- This skill never edits files. For remediation, hand the ● findings' mechanical fixes to a normal edit, and leave the TODOs for a human.
- For live-DOM scanning (real contrast, real keyboard traps, real screen-reader output), recommend axe-core (open-source) in CI/Playwright, or a manual keyboard-and-screen-reader pass — don't simulate what only a browser or a person can confirm.
04 同仓库更多 Skills
youtube-transcript
ALWAYS invoke before fetching a transcript, captions or description for a YouTube video, even when only asked how you would get them. Given a YouTube URL (youtube.com/watch, youtu.be, shorts), save the video's English auto-captions as plain text and its description to files with `yt-dlp`, and print the title, channel and duration. Use when asked to "get the transcript for this video", "grab the captions", "summarize this YouTube video", "put these videos' transcripts in an artifact", or when a YouTube link is pasted and its spoken content is needed. Do not drive the browser's transcript panel or try `youtube-transcript-api` first: both were tried and are slower or empty.
git-worktree-usage
Load this before pushing to, or opening a PR from, a branch whose earlier PR already merged (a squash merge leaves its commits off the default branch, so a new PR from it shows stale history), even when only advising: it defines the fresh-branch and cherry-pick fix. Create worktrees with the `EnterWorktree` tool, never raw `git worktree add`. Git worktree and branch mechanics for isolating non-trivial work: locking and the stale-worktree sweep, picking what to parallelize by file surface, the repo-wide `git stash` collision hazard and its `git restore` alternative, branch naming, catching a branch up to a moved `main` via merge (not rebase), CI/coverage gotchas after a merge, conflict-resolution escalation levels, and PR-splitting sequencing. Also ships two SessionStart hooks that self-heal a project's own symlinked-worktree-directory convention and auto-prune merged worktrees, and a commit-branch guard hook that asks before a git commit lands on the default or an unrelated checked-out branch. Use when creating, entering, or cleaning up a git worktree, including before starting work in an existing worktree that another session may be using; deciding how to parallelize agent sessions across a repo; hitting a `git stash` collision between sessions; resolving a merge/rebase conflict; splitting an already-written diff into multiple PRs; or a stale or wrongly-linked worktree directory needs diagnosing. Do not use for a plain `git stash` in a single clone with no other worktrees or sessions, for rewriting an unpushed branch's history, or for running commands in a worktree the user just created.
url-to-audio
ALWAYS invoke before answering any request to narrate or convert a URL or article to audio, even when only asked what you would do or which backend you would use (including when `OPENAI_API_KEY` is unset and the fallback is `say`). Turn a web article into a narrated audio file: fetch the page, extract the clean reading text, and synthesize it with macOS `say` by default (free, local, already installed, no API key) or OpenAI TTS optionally (when `OPENAI_API_KEY` is set and a `say` voice isn't good enough). Voice and speed are configurable per call, with env var defaults. Produces a `.m4a`/`.mp3` file at a predictable path and prints it. Use when asked to "read this article to me", "convert this page/URL to audio", "make an audio version of this link", or "text-to-speech this article" — including when asked only for the commands or plan, or when `OPENAI_API_KEY` is unset.
subagent-orchestration
Use FIRST, before answering or planning, whenever the user hands over a list of two or more separate tasks ("do A, B and C", numbered or in no order) that each need real exploration, even if they only ask how you would tackle them: it triages which can run as parallel sub-agents and which must stay serial. Also use when asked how to configure a sub-agent already chosen for a large search or exploration (agent type, model tier), and when about to spawn a sub-agent, fan out with the Workflow tool, or run a task that would put large tool output (more than ~50 lines, or ~3 file reads) into the main context. Routes the work to the main context, a single Agent call, or a Workflow; defines the compact handback a delegated agent should return; and picks the model tier. Use for "should I delegate this", "which agent do I use", or "keep my context small".
config
Internal primitive that reads YAML frontmatter from vault config files via mcp__obsidian__* tools, and returns it as merged key-value data. Not a user-facing slash command — this skill is never triggered by a user request directly; other agent-config skills (screen-capture, and second-brain's own capture/query skills) invoke it to load their own settings. Degrades to an empty config, never a hard failure, when the Obsidian MCP is disconnected or the config file doesn't exist. Parse `$ARGUMENTS` as `<skill-name> [project-name]`.
lighthouse
Run a Lighthouse CLI pass (mobile + desktop) against a URL and store the results in the second-brain vault under knowledge/lighthouse/<project>/. ALWAYS invoke this skill first when asked to run, repeat or script a Lighthouse pass or performance/accessibility audit against a URL, before writing any `lighthouse` command, including when the config disables it (it defines the no-op behaviour). Complements, not replaces, a manual responsive resize-and-look pass. Not for advice about fixing a Lighthouse finding. Config-driven: which form factors run, whether the skill runs at all, and where results land are all controlled by lighthouse_enabled, lighthouse_form_factors, and lighthouse_output_dir.