christianestay/claude-code-base-project
Developer tools이 페이지는 AI로 자동 번역되었습니다. 기술적인 세부 사항은 원문을 확인하세요. 원문 보기
Production-ready Claude Code template with 4 specialized agents, 12 Open Specs skills, self-improvement loop, and anti-hallucination layers. Built for Turborepo + React + Supabase monorepos
개요
A production-ready template for Claude Code that turns it into a team of specialized agents with persistent memory, domain-specific skills, and automated workflows. Built for monorepos. — all 12 skills follow the Agent Skills open standard, making them portable to GitHub Copilot, OpenAI Codex CLI, Cursor, and 20+ other compatible tools. Claude Code out of the box is a general-purpose assistant. This template transforms it into an opinionated engineering team that: - — evaluates scope, risks, and affected workspaces before writing a single line - — every feature, form, query, and test follows documented conventions - — a self-improvement loop captures lessons and prevents repeated errors - — runs type-check, lint, and tests before marking anything as done - — uses subagents and progressive disclosure to stay within token limits Four specialized agents, each with persistent memory across sessions: Every agent starts by reading tasks/lessons.
README
원문 README의 일부입니다. 원문 보기.
Claude Code Project Template v3.4
A production-ready template for Claude Code that turns it into a team of specialized agents with persistent memory, domain-specific skills, and automated workflows. Built for Turborepo + React/Vite + Supabase monorepos.
Open Specs compliant — all 12 skills follow the Agent Skills open standard, making them portable to GitHub Copilot, OpenAI Codex CLI, Cursor, and 20+ other compatible tools.
Why this template?
Claude Code out of the box is a general-purpose assistant. This template transforms it into an opinionated engineering team that:
- Plans before coding — evaluates scope, risks, and affected workspaces before writing a single line
- Follows consistent patterns — every feature, form, query, and test follows documented conventions
- Learns from mistakes — a self-improvement loop captures lessons and prevents repeated errors
- Verifies its own work — runs type-check, lint, and tests before marking anything as done
- Manages context efficiently — uses subagents and progressive disclosure to stay within token limits
What’s inside
.claude/
├── agents/ # 4 specialized agents with persistent memory
│ ├── implementer.md # Builds features following project conventions
│ ├── code-reviewer.md # Reviews code for quality, UI, forms, a11y
│ ├── researcher.md # Investigates packages/APIs with anti-hallucination
│ └── test-engineer.md # Writes and runs tests (Vitest + Testing Library)
│
├── skills/ # 12 skills (Open Specs compliant)
│ ├── scaffold-feature/ # Generate feature file structure
│ │ ├── SKILL.md
│ │ └── references/CODE-PATTERNS.md
│ ├── supabase-patterns/ # Validated Supabase patterns
│ │ ├── SKILL.md
│ │ └── references/ADVANCED.md, TANSTACK.md
│ ├── ui-patterns/ # UI: layouts, tables, feedback, pages
│ │ ├── SKILL.md
│ │ └── references/LAYOUTS.md, DATA-TABLES.md, FEEDBACK.md, PAGES.md
│ ├── forms-validation/ # Zod + React Hook Form + shadcn/ui
│ │ ├── SKILL.md
│ │ └── references/PATTERNS.md, WIZARD.md
│ ├── code-quality/ # ESLint, Prettier, tsconfig, Vite configs
│ │ ├── SKILL.md
│ │ └── references/CONFIGS.md
│ ├── commit/SKILL.md # Auto-bump patch + conventional commits
│ ├── deploy-check/SKILL.md # Pre-merge verification checklist
│ ├── new-feature/SKILL.md # Full feature pipeline
│ ├── evaluate/SKILL.md # Scope, effort, and risk assessment
│ ├── review/SKILL.md # Code quality review
│ ├── research/SKILL.md # Technical research with source verification추천 도구
다른 키워드를 입력하거나 필터를 제거해 보세요.
설치
npx skillfish add christianestay/claude-code-base-project