A ready-to-use .claude/ configuration folder for any repository. Ships with skills, agents, and settings aligned to Claude Code best practices (March 2026).
git clone https://github.com/your-org/claude-code-bootstrap.git my-project
cd my-project
rm -rf .git && git initThen open Claude Code and say: "plan repo" to plan your stack, then "initialize repo" to configure.
# Copy the .claude/ folder into your repo
cp -r path/to/claude-code-bootstrap/.claude/ your-repo/.claude/
cp path/to/claude-code-bootstrap/CLAUDE.md your-repo/CLAUDE.md
cp path/to/claude-code-bootstrap/agents.md your-repo/agents.mdThen open Claude Code in your repo and say: "initialize repo" to merge the template with your existing setup.
plan repo → initialize repo → (build features) → update practices
↑ |
| spec developer ←─────┘
| (plan in one session, execute in another)
└──────────────────────────────────────────────────────┘
- Plan first:
plan reporesearches current options and recommends the best stack for your project, then generates README, design guardrails, and tools reference. - Initialize:
init-reporeads the plan and configures.claude/with skills, agents, settings, and hierarchical CLAUDE.md files. - Build features: Use
spec developerfor big features -- it interviews you, explores the codebase, and generates a detailed plan saved to/tasks. Execute in a fresh session. - Stay current:
update practicesfetches latest best practices and updates your config. Safe to run anytime.
| Skill | Trigger | Description |
|---|---|---|
| plan-repo | "plan repo" | Research and recommend best tech stack, generate README, design guardrails, tools reference |
| init-repo | "initialize repo" | Build or rebuild the .claude/ folder with best practices |
| update-practices | "update practices" | Fetch latest best practices and update config |
| spec-developer | "spec developer" | Interview-driven feature spec saved to /tasks |
| code-review | "code review" | Full codebase review with severity-ranked findings |
| security-scan | "security scan" | OWASP Top 10, secrets detection, dependency audit |
| performance-review | "performance review" | Bottleneck analysis with impact-ranked fixes |
| dependency-audit | "dependency audit" | Outdated, vulnerable, and unused dependency detection |
| test-scaffold | "scaffold tests" | Generate test files for untested modules |
| doc-sync | "sync docs" | Align documentation with current code |
| mermaid-diagram | "mermaid diagram" | Generate data flow / architecture diagrams |
See agents.md for the full agent registry.
| Agent | Purpose |
|---|---|
| architect | Phase-based planning, tech stack decisions, file structure |
| reviewer | Code review for correctness and maintainability |
| security | Vulnerability detection and security analysis |
| performance | Bottleneck identification and optimization |
| explorer | Codebase exploration, research, and context gathering |
- Phase-based planning: Foundation → Core → Polish → Ship. No timelines.
- Hierarchical CLAUDE.md: Root → subfolder, loaded top-down. Only relevant files load.
- Subagent-first: Always offload research, exploration, and log analysis to subagents. Include a "why" in every subagent prompt.
- Plan/execute separation: Plan in one session, execute in another. Save plans to
/tasks. - Date-aware practices: Always checks the current date when fetching best practices.
- Tools reference:
.claude/references/tools.mdlists all CLI tools so Claude can detect and install missing ones. - Design guardrails:
.claude/references/design-guardrails.mdenforces UI/design SLA for frontend projects.
Say "update practices" in Claude Code. The skill fetches the latest best practices from official and community sources, then updates your config. Safe to run anytime.
See instructions.md for complete documentation on every skill, agent, and configuration option.
MIT