accessibility-patterns
WCAG 2.2 AA compliance, ARIA patterns, keyboard navigation, screen reader optimization
Implement a feature with a bounded plan, targeted changes, and verification through the real execution path.
Claude Code
$ npx -y skills add vibeeval/vibecosystem --skill build --agent claude-code
Codex
$ npx skills add vibeeval/vibecosystem --skill build -a codex -y
Cursor
$ npx skills add vibeeval/vibecosystem --skill build -a cursor -y
Gemini CLI
$ npx skills add vibeeval/vibecosystem --skill build -a gemini-cli -y
GitHub Copilot
$ npx skills add vibeeval/vibecosystem --skill build -a github-copilot -y
OpenCode
$ npx skills add vibeeval/vibecosystem --skill build -a opencode -y
Windsurf
$ npx skills add vibeeval/vibecosystem --skill build -a windsurf -y
name: build description: Implement a feature with a bounded plan, targeted changes, and verification through the real execution path. user_invocable: true keywords: [build, greenfield, brownfield, tdd, refactor]
Use `/build [greenfield|brownfield|tdd|refactor] <goal>`. Infer the mode from the request and repository; ask only for missing information that changes the implementation. Honor supplied options, including diagnosis-only or no-commit requests.
Inspect repository instructions, working-tree changes, entry points and affected tests. Search filenames and symbols first, then read the implementation and relevant callers. Reuse a current plan or handoff after checking it against the code.
For a bounded task, state the intended change and implement it directly. For work spanning components, make a short plan with acceptance criteria, dependencies and verification. Preserve user changes, compatibility, state formats and error paths. In TDD mode establish a meaningful failing test before the fix; in refactor mode preserve observable behavior.
Run focused checks and exercise the public entry point, including built or installed artifacts when relevant. Broaden verification for affected boundaries or unresolved risk. Investigate demonstrated failures; report environment blockers rather than retrying indefinitely.
Review the diff and report behavior, checks and limitations. Commit or prepare a PR only within the requested workflow; publishing and destructive migrations require authorization.
Work in the parent context by default. For a concrete independent question Claude can use installed `scout`, `planner`, `spark`, `code-reviewer` or `verifier`. Give relevant paths and a short evidence-based output contract; reuse workers and avoid overlapping edits. Codex retains its parent `luna_worker` model and reasoning; Claude model metadata does not select a Codex model.
No additional skill, MCP, indexer, memory service or swarm is required. Use optional tools only when installed and useful.
A focused runtime for Claude Code. A bounded worker for Codex. Catalog (in repo): Default core: vibecosystem packages agents, skills and hooks for software work.
WCAG 2.2 AA compliance, ARIA patterns, keyboard navigation, screen reader optimization
axe-core integration, WCAG 2.2 AA checklist, keyboard navigation testing, screen reader testing, and ARIA pattern validation.
Steam-style achievement system with XP, levels, streaks, and skill trees. Gamifies the development workflow. 25 achievements…
Framework for measuring and tracking agent response quality over time. Detects regressions before they reach production. Use when…
Agent ve skill dosyalarinin yapisal dogrulamasi. Frontmatter kontrol, naming convention, zorunlu bolum kontrolu, tutarlilik…
You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or…
Ask which skill or flow fits your situation. A router over the skills in this repo.