官方说明开源项目7 天前
gabrieldcpaiva/not-optional
Accessibility is not a request. A practical, lived-experience toolkit for making digital products and printables usable by default.
在 GitHub 查看 ↗直链 · 无中转页
- ★ 数取值
- 1 ★ · gabrieldcpaiva/not-optional · 取自该仓库的 GitHub 页面 · （2026-10-16 20:00:20 核对）
- 来源明细
- 从该仓库的 GitHub 页面直接读取 · 条目类型：开源项目
- 收录与活跃
- 收录于 2026-10-10 · 最近见到 2026-10-10 · 共见到 1 次 · 被 1 个来源列出
- 评分构成
- 综合分 103 · 1 ★ · 1 个来源 · repo
- 安装
npx skills
常见问题（答案都来自我们的记录）
怎么安装？
该仓库给出的安装命令是
npx skills。我们的记录里它是仓库自己提供的；不保证在你的客户端上直接可用。它适合谁 / 用来做什么？
我们把它归在这些标签下: 开源项目。标签来自条目自身，不是我们强加的分类。
谁维护它、还在更新吗？
仓库
gabrieldcpaiva/not-optional；我们收录于 2026-10-10；最近见到 2026-10-10。这里没有"官方认证"或"质量评级"——我们只记录来源与时间。第三方质量标记
not-optional
以下摘录自该仓库的 README（片段） · 在 GitHub 读完整 README →
Accessibility is not a request.
It is not something you add because of a fine or an invoice.
It is not a premium feature.
It is baseline competence inside the workflow.
This repository is a technical toolkit for web interfaces and physical printables. It combines WAI / WCAG 2.1 AA practice with four additional constraints that checklists often skip:
1. Physiological contrast (avoid pure black-on-white visual vibration)
2. Executive-function chunking (short measures, hard paragraph caps, empty space)
3. Dual medium (web code ≠ print geometry)
4. Multi-channel state (never hue alone)
It is for any surface a human has to read, tap, or print. Not for one brand, one shop, or one product line.
Quick Start for AI (Agents & LLMs)
If you are instructing an AI agent (like a GitHub Agent, Copilot, or Cursor) to build an accessible UI, provide them with the master context file:
1. Point your agent to read agent-instructions.md.
2. Or, copy the contents of agent-instructions.md into your agent's system prompt / .cursorrules file.
- **Direct Link for Remote Agents**: Provide the raw GitHub URL so web-browsing agents (ChatGPT, Claude, Gemini) can ingest the master rules in one turn:
https://raw.githubusercontent.com/gabrieldcpaiva/not-optional/main/agent-instructions.md
- **1-Line CLI Drop-In**: Add curl snippets for terminal and IDE agents (Cursor, Windsurf, Claude Code, Aider):
Agent Skills Registry Command
For systems adhering to the agentskills.io specification:
*To rebuild this file after modifying rules, run ./build-agent-context.sh.*
Quick Start for Humans
To understand what these rules look like in practice, see the concrete baseline examples:
- **Web:** examples/web-baseline.html (Demonstrates tokens, focus rings, micro-chunking, and semantic HTML).
- **Print:** examples/print-baseline.html (Demonstrates physical geometry, safe margins, and vector text).
Dual-medium boundary
- **Web:** semantic HTML, :focus-visible, WCAG 2.1 AA, ARIA only when a native element cannot express the state.
- **Print:** vector text, ≥12pt body, ≥12.7mm safe margins, high-contrast ink-saving layers. No ARIA, no focus logic, no live regions in PDFs.
What this is
- Nine agent skills (skills/*/SKILL.md)
- Principles plus machine-readable token files
- Copy one file or the whole suite
It is not something you add because of a fine or an invoice.
It is not a premium feature.
It is baseline competence inside the workflow.
This repository is a technical toolkit for web interfaces and physical printables. It combines WAI / WCAG 2.1 AA practice with four additional constraints that checklists often skip:
1. Physiological contrast (avoid pure black-on-white visual vibration)
2. Executive-function chunking (short measures, hard paragraph caps, empty space)
3. Dual medium (web code ≠ print geometry)
4. Multi-channel state (never hue alone)
It is for any surface a human has to read, tap, or print. Not for one brand, one shop, or one product line.
Quick Start for AI (Agents & LLMs)
If you are instructing an AI agent (like a GitHub Agent, Copilot, or Cursor) to build an accessible UI, provide them with the master context file:
1. Point your agent to read agent-instructions.md.
2. Or, copy the contents of agent-instructions.md into your agent's system prompt / .cursorrules file.
- **Direct Link for Remote Agents**: Provide the raw GitHub URL so web-browsing agents (ChatGPT, Claude, Gemini) can ingest the master rules in one turn:
https://raw.githubusercontent.com/gabrieldcpaiva/not-optional/main/agent-instructions.md
- **1-Line CLI Drop-In**: Add curl snippets for terminal and IDE agents (Cursor, Windsurf, Claude Code, Aider):
Agent Skills Registry Command
For systems adhering to the agentskills.io specification:
*To rebuild this file after modifying rules, run ./build-agent-context.sh.*
Quick Start for Humans
To understand what these rules look like in practice, see the concrete baseline examples:
- **Web:** examples/web-baseline.html (Demonstrates tokens, focus rings, micro-chunking, and semantic HTML).
- **Print:** examples/print-baseline.html (Demonstrates physical geometry, safe margins, and vector text).
Dual-medium boundary
- **Web:** semantic HTML, :focus-visible, WCAG 2.1 AA, ARIA only when a native element cannot express the state.
- **Print:** vector text, ≥12pt body, ≥12.7mm safe margins, high-contrast ink-saving layers. No ARIA, no focus logic, no live regions in PDFs.
What this is
- Nine agent skills (skills/*/SKILL.md)
- Principles plus machine-readable token files
- Copy one file or the whole suite
- 来源
- github · https://github.com/gabrieldcpaiva/not-optional
- 按领域浏览
- 开源项目
- Traction
- 1 ★
- 收录于
- 7 天前
更多开源项目
所属领域与标签
每个标签都能点进去看同类的全部条目。数字是当前库里的真实条数。