a11y-ui-kit
doggonug-cpu.github.io
Small, static, no-signup tools built in Wando, Republic of Korea — published as a GitHub Pages user site at https://doggonug-cpu.github.io. No framework, no bundler, no tracking, no server-side storage.
한국어 — 완도에서 만드는 작은 정적 웹도구 모음입니다. 회원가입·추적·서버 저장이 없고,
모든 계산과 문서 작성은 방문자의 브라우저에서만 이뤄집니다. 아래 네 서비스는 지금 바로
쓸 수 있고, AI 에이전트용 인덱스(llms.txt)와 MCP 서버도 함께 배포합니다.
Live services
Project | What it is | For |
18 senior-first accessible UI components (48px+ targets, 18.4:1 contrast, synthesized sound) + a local MCP server + AI rule packs. 10 components MIT. | AI builders, agencies, kiosk and care teams | |
견적서 · 거래명세서 · 간이영수증 in ~10 seconds, printed or saved as PDF. Offline PWA, 20 currencies, KO/EN. | Small businesses, freelancers, on-site sales | |
Free 40-minute AI-literacy lesson kit (plan, command cards, worksheet) for age 5–grade 2, zero devices needed. | Teachers, digital tutors, kindergarten staff | |
Net-profit, margin and break-even price across 21 Korean sales channels, incl. VAT, ads and returns. | Sellers, freelancers, own-shop owners |
Related MCP server: ui-unslop-mcp
Also running
한미 공시 3줄 요약 — a free Telegram channel posting five Korean and US filings as three factual lines twice a day. Custom keyword/ticker alert bots are quoted at https://doggonug-cpu.github.io/#disclosure (software build; not investment advice).
For AI agents
Hub index: https://doggonug-cpu.github.io/llms.txt
A11yKit index: https://doggonug-cpu.github.io/a11y-ui-kit/llms.txt · full text: https://doggonug-cpu.github.io/a11y-ui-kit/llms-full.txt
MCP descriptor: https://doggonug-cpu.github.io/a11y-ui-kit/mcp/server.json · discovery: https://doggonug-cpu.github.io/.well-known/mcp.json
Rule files: https://doggonug-cpu.github.io/a11y-ui-kit/ai-rules.json
robots.txt explicitly allows GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot,
Google-Extended and the other main AI crawlers.
Repository notes
이 저장소는 완도애협동조합 공식 사이트가 아닙니다. 조합 사이트는 별도 저장소
doggonug-cpu/wandoae이며 https://wandoae.me 에서 운영됩니다.이 저장소에는
CNAME파일을 두지 않습니다. (같은 도메인을 두 저장소가 주장하면 GitHub Pages HTTPS 인증서 발급이 중단됩니다.)작업 규칙은 AGENTS.md를 따릅니다.
main푸시가 곧 배포입니다.
Available Tools
4 toolscheck_wcag_checklistGet a WCAG manual checklistARead-only
Return the WCAG criteria that apply to a component, each with the manual check a human must run and the component-specific accessibility notes.
Inputs: id (required), level.
get_ai_promptGet an implementation promptARead-only
Return a paste-ready implementation prompt for Claude Code, Cursor, Lovable or a generic agent, in English or Korean. Pro components return a lock notice.
Inputs: id (required), target, language.
get_componentGet component sourceARead-only
Return one component: plain HTML/CSS/JS source, a React version, WCAG criteria, accessibility notes and labels. Pro components return the spec and a lock notice instead of source.
Inputs: id (required), include_code.
list_componentsList A11yKit componentsARead-only
List the 18 senior-first components with tier (free/pro), WCAG criteria and tags. Filter by tier, tag or keyword. Free components include MIT source; Pro components list the spec only.
Inputs: tag, tier, limit, query.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool has a largely distinct role: catalog listing, single-component retrieval, prompt generation, and WCAG checklist lookup. The only overlap is that get_component also returns WCAG criteria, which check_wcag_checklist expands with manual checks and notes, so an agent could occasionally hesitate between the two.
All four names follow a clear verb_noun pattern (list_components, get_component, get_ai_prompt, check_wcag_checklist). No mixed casing or inconsistent verb styles.
Four tools is a tight, well-scoped surface for a component catalog with three read paths plus a checklist helper. It is slightly thin (no search or tier/version discovery tool separate from list), but nothing feels redundant.
The read-only lifecycle (list, get, prompt, accessibility check) is covered, and filtering by tier/tag/keyword substitutes for a dedicated search tool. Minor gaps remain: no way to enumerate WCAG criteria globally, no versioning/changelog access, and Pro content is intentionally spec-only, which agents must work around.
Maintenance
Related MCP Connectors
Live React design-system APIs, patterns, and code validation so AI agents build real UI, not slop.
500+ typed UI atoms an agent composes into real rendered interfaces. Free, no API key.
React, Tailwind CSS and shadcn/ui components, sections and landing pages for coding agents.
React components and design rules your coding agent installs and follows.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables AI coding assistants to generate bespoke design systems by synthesizing OKLCH color palettes, chamfer highlights and spring physics, teaching layout composition grammar, and returning micro-interaction primitives and component blueprints. Also audits JSX/TSX/HTML through a senior design director's lens to score and critique UI quality.1MIT
- AlicenseNot gradedqualityCmaintenanceEnables coding agents to deterministically audit markup and stylesheets for WCAG contrast failures, accessibility errors and design-token sprawl, and to compute contrast ratios and inventory hardcoded colors, sizes, radii and spacing before shipping UI code. Works across any web stack with Claude Code, Claude Desktop, Codex and Antigravity, and can also run as a CI check that exits non-zero on errors.195 npmMIT
- AlicenseBqualityCmaintenanceEnables AI coding agents to audit Apple HIG compliance across macOS, iPadOS, and iOS, extract multi-format design tokens, compare UI patterns, and scaffold production-ready React/TypeScript components.9MIT
- AlicenseAqualityBmaintenanceEnables AI agents to audit HTML snippets against 18 WCAG 2.2 static checks, look up success criteria and their applicability, retrieve browser-verified failing/corrected markup examples with manual test steps, and resolve ADA Title II, Section 504, and other accessibility deadlines by entity type and date. Runs fully offline with bundled datasets and no API keys, while explicitly reporting that a clean audit is not a conformance claim.442 npmMIT