Claude Code Reference
v2.1.114 · Commands, shortcuts, hooks, Python/AWS tips.
What's New
Everything recent across Claude Code, Claude models, new skills and agents.
New version 2.1.114 published on npm.
Claude Code skill for generating production-quality SVG+PNG technical diagrams. Supports 8 diagram types, 5 visual styles, and deep AI/Agent domain knowledge.
🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman
Most capable generally available model. Complex multi-step coding, long agentic workflows, 1M-token codebase reads.
An intuitive interface to create, orchestrate, and monitor Claude-based code agents for automated development workflows.
Claude Code skills based on The Minimalist Entrepreneur by Sahil Lavingia
A Claude Code skill that turns any codebase into a beautiful, interactive single-page HTML course for non-technical vibe coders.
Real-time visualization of Claude Code agent orchestration — see your agents think, branch, and coordinate as they work.
A collection of 130+ specialized Codex subagents covering a wide range of development use cases.
Graphs that teach > graphs that impress. Turn any code, or knowledge base (Karpathy LLM wiki), into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
Claude Autoresearch Skill — Autonomous goal-directed iteration for Claude Code. Inspired by Karpathy's autoresearch. Modify → Verify → Keep/Discard → Repeat forever.
Clone any website with one command using AI coding agents
A Claude skill that writes the accurate prompts for any AI tool. Zero tokens or credits wasted. Full context and memory retention
ARIS ⚔️ (Auto-Research-In-Sleep) — Lightweight Markdown-only skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation. No framework, no lock-in — works with Claude Code, Codex, OpenClaw, or any LLM agent.
Open-source AI agent command center for Claude Code agent teams. Built on OpenClaw.
Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Supercharge Claude Code with 11 AI agents, 36 commands & 15 skills — the claude-code plugin framework inspired by oh-my-zsh. 6-layer security hooks included. 5-min install.
Comparison between open source PI agent and closed source Claude Code agent
Quick Start
Start a session
claudeOpens interactive mode in your project directory.
One-shot question
claude -p "how does the auth module work?"Ask a question, get an answer, exit.
Continue last session
claude -cPicks up exactly where you left off.
Parallel worktree
claude -w feature-authWork on a feature branch without leaving your current work.
Plan mode (safe)
claude --permission-mode planClaude analyzes and plans but does not edit until you approve.
Pipe content
git diff main | claude -p "review this"Feed diffs, logs, or files directly into Claude.
Slash Commands
/help | Show help menu |
/config | Open settings interface |
/model | Switch between models (Opus, Sonnet, Haiku) |
/effort | Set effort level (low, medium, high, max) |
/clear | Clear conversation and start fresh |
/compact | Manually compact context window |
/resume | Resume previous sessions (open picker) |
/rename | Rename current session |
/plan | View current plan (in plan mode) |
/usage | View token usage and costs |
/version | Show Claude Code version |
/ide | Connect to VS Code or JetBrains IDE |
/thinking-toggle | Toggle extended thinking on/off |
/agents | View and create custom subagents |
/skills | Browse available skills |
/plugins | Browse and manage plugins |
/mcp | Add and configure MCP servers |
/hooks | Browse configured hooks |
/permissions | View and manage tool permissions |
/memory | View project memory (CLAUDE.md) |
/debug | Enable debug logging for troubleshooting |
/simplify | Review code for reuse, quality, and fix issues |
/batch | Plan and execute large-scale changes across 5-30 isolated worktree agents |
/loop | Run a prompt or command on a recurring interval |
/schedule | Create scheduled remote agents on cron schedule |
/claude-api | Build, debug, and optimize Claude API / Anthropic SDK apps |
/find-docs | Retrieve up-to-date docs for any library (via Context7) |
/init | Initialize CLAUDE.md for your project |
/review | Review a pull request |
/security-review | Perform security review of pending changes |
/insights | Generate report analyzing Claude Code sessions |
/install-github-app | Install Claude GitHub App for CI/CD |
/powerup | Interactive lessons with animated demos |
/remote-control | Start Remote Control session |
Keyboard Shortcuts
Global
Ctrl+C | Cancel current operation |
Ctrl+D | Exit Claude Code |
Ctrl+T | Toggle task list visibility |
Ctrl+O | Toggle verbose transcript (shows thinking) |
Chat Input
Enter | Submit message |
Shift+Enter | New line without submitting |
Ctrl+L | Clear prompt input |
Ctrl+S | Stash current prompt |
Ctrl+G | Open in external editor |
Ctrl+V | Paste image |
Ctrl+_ or Ctrl+Shift+- | Undo last action |
Shift+Tab | Cycle permission modes |
Cmd+P / Meta+P | Open model picker |
Meta+O | Toggle fast mode |
Option+T / Alt+T | Toggle extended thinking |
Ctrl+R | Open history search |
Up/Down Arrow | Navigate command history |
VS Code
Cmd+Esc / Ctrl+Esc | Toggle focus between editor and Claude |
Cmd+Shift+Esc / Ctrl+Shift+Esc | Open new conversation tab |
Option+K / Alt+K | Insert @-mention reference with line numbers |
Cmd+N / Ctrl+N | Start new conversation |
CLI Flags & Options
claudeclaudeclaude "prompt"claude "fix the auth bug"claude -pclaude -p "analyze this code"claude -cclaude -cclaude -r "name"claude -r "auth-refactor"claude --resumeclaude --resumeclaude -wclaude -w feature-auth--modelclaude --model claude-opus-4-6--effortclaude --effort high--permission-modeclaude --permission-mode plan--allowedTools--allowedTools "Read,Edit,Bash"--max-turnsclaude -p --max-turns 5--max-budget-usdclaude -p --max-budget-usd 5.00--output-format jsonclaude -p --output-format json "query"--bareclaude --bare--debugclaude --debugpipe inputcat file.py | claude -p "review this"Environment Variables
ANTHROPIC_API_KEY | API key for direct Anthropic access |
CLAUDE_CODE_USE_BEDROCK | Set to 1 to use AWS Bedrock |
CLAUDE_CODE_USE_VERTEX | Set to 1 to use Google Vertex AI |
CLAUDE_CODE_EFFORT_LEVEL | Default effort level (low/medium/high/max) |
MAX_THINKING_TOKENS | Max tokens for extended thinking |
CLAUDE_CODE_SIMPLE | Set to 1 for bare mode (no hooks/skills) |
Available Tools
Read files from disk
Create new files
Edit existing files (sends only the diff)
Run shell commands
Find files by pattern (faster than find)
Search file contents (faster than grep/rg)
Search the web
Fetch and parse web pages
Watch background scripts
Invoke subagents for parallel work
Create and manage task lists
Ask clarifying questions
Configuration Files
~/.claude/settings.json.claude/settings.json.claude/settings.local.jsonCLAUDE.md.claude/CLAUDE.md~/.claude/CLAUDE.md~/.claude/keybindings.json.claude/mcp.json.claude/skills/<name>/SKILL.md.claude/agents/<name>/.worktreeincludeHooks System
Hooks run shell commands in response to Claude events. Configure in settings.json. Exit 0 = allow, Exit 2 = block action.
SessionStart | Session begins. Cannot block. | |
UserPromptSubmit | You submit a prompt. Cannot block. | |
PreToolUse | Before tool executes. CAN block (exit 2). | Use matcher: "Bash" to intercept shell commands |
PostToolUse | After tool succeeds. Cannot block. | Great for auto-formatting after Edit/Write |
PostToolUseFailure | After tool fails. Cannot block. | |
Stop | Claude finishes responding. Cannot block. | |
Notification | Claude needs your attention. Cannot block. | |
SubagentStart | Subagent lifecycle start. | |
SubagentStop | Subagent lifecycle end. | |
FileChanged | Watched files change. |
Example: Auto-format Python on every edit
{
"hooks": {
"PostToolUse": [
{
"matcher": "Edit|Write",
"hooks": [
{
"type": "command",
"command": "ruff format $(jq -r '.tool_input.file_path') 2>/dev/null; ruff check --fix $(jq -r '.tool_input.file_path') 2>/dev/null; true"
}
]
}
],
"Notification": [
{
"matcher": "",
"hooks": [
{
"type": "command",
"command": "notify-send 'Claude Code' 'Needs your attention'"
}
]
}
]
}
}Extensions
Specialized subagents in .claude/agents/<name>.md. Invoke via the Agent tool.
Recommended · 50
Auto-discovered · dailyCoding · 19
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
affaan-m/everything-claude-code
Install ↓
git clone https://github.com/affaan-m/everything-claude-code.git /tmp/everything-claude-code && cp -r /tmp/everything-claude-code/.claude/agents/* .claude/agents/A Claude Code plugin that automatically captures everything Claude does during your coding sessions, compresses it with AI (using Claude's agent-sdk), and injects relevant context back into future sessions.
thedotmack/claude-mem
Install ↓
git clone https://github.com/thedotmack/claude-mem.git /tmp/claude-mem && cp -r /tmp/claude-mem/.claude/agents/* .claude/agents/from vibe coding to agentic engineering - practice makes claude perfect
shanraisshan/claude-code-best-practice
Install ↓
git clone https://github.com/shanraisshan/claude-code-best-practice.git /tmp/claude-code-best-practice && cp -r /tmp/claude-code-best-practice/.claude/agents/* .claude/agents/AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs
CherryHQ/cherry-studio
Install ↓
git clone https://github.com/CherryHQ/cherry-studio.git /tmp/cherry-studio && cp -r /tmp/cherry-studio/.claude/agents/* .claude/agents/A curated list of awesome skills, hooks, slash-commands, agent orchestrators, applications, and plugins for Claude Code by Anthropic
hesreallyhim/awesome-claude-code
Install ↓
git clone https://github.com/hesreallyhim/awesome-claude-code.git /tmp/awesome-claude-code && cp -r /tmp/awesome-claude-code/.claude/agents/* .claude/agents/Installable GitHub library of 1,400+ agentic skills for Claude Code, Cursor, Codex CLI, Gemini CLI, Antigravity, and more. Includes installer CLI, bundles, workflows, and official/community skill collections.
sickn33/antigravity-awesome-skills
Install ↓
git clone https://github.com/sickn33/antigravity-awesome-skills.git /tmp/antigravity-awesome-skills && cp -r /tmp/antigravity-awesome-skills/.claude/agents/* .claude/agents/🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade architecture, distributed swarm intelligence, RAG integration, and native Claude Code / Codex Integration
ruvnet/ruflo
Install ↓
git clone https://github.com/ruvnet/ruflo.git /tmp/ruflo && cp -r /tmp/ruflo/.claude/agents/* .claude/agents/Teams-first Multi-agent orchestration for Claude Code
Yeachan-Heo/oh-my-claudecode
Install ↓
git clone https://github.com/Yeachan-Heo/oh-my-claudecode.git /tmp/oh-my-claudecode && cp -r /tmp/oh-my-claudecode/.claude/agents/* .claude/agents/A powerful MCP toolkit for coding, providing semantic retrieval and editing capabilities - the IDE for your agent
oraios/serena
Install ↓
git clone https://github.com/oraios/serena.git /tmp/serena && cp -r /tmp/serena/.claude/agents/* .claude/agents/Install ↓
git clone https://github.com/gastownhall/beads.git /tmp/beads && cp -r /tmp/beads/.claude/agents/* .claude/agents/Install ↓
git clone https://github.com/addyosmani/agent-skills.git /tmp/agent-skills && cp -r /tmp/agent-skills/.claude/agents/* .claude/agents/Ghostty-based macOS terminal with vertical tabs and notifications for AI coding agents
manaflow-ai/cmux
Install ↓
git clone https://github.com/manaflow-ai/cmux.git /tmp/cmux && cp -r /tmp/cmux/.claude/agents/* .claude/agents/Clone any website with one command using AI coding agents
JCodesMore/ai-website-cloner-template
Install ↓
git clone https://github.com/JCodesMore/ai-website-cloner-template.git /tmp/ai-website-cloner-template && cp -r /tmp/ai-website-cloner-template/.claude/agents/* .claude/agents/232+ Claude Code skills & agent plugins for Claude Code, Codex, Gemini CLI, Cursor, and 8 more coding agents — engineering, marketing, product, compliance, C-level advisory.
alirezarezvani/claude-skills
Install ↓
git clone https://github.com/alirezarezvani/claude-skills.git /tmp/claude-skills && cp -r /tmp/claude-skills/.claude/agents/* .claude/agents/The best way to get AI coding agents to solve hard problems in complex codebases.
humanlayer/humanlayer
Install ↓
git clone https://github.com/humanlayer/humanlayer.git /tmp/humanlayer && cp -r /tmp/humanlayer/.claude/agents/* .claude/agents/Orchestration framework for running 20+ Claude Code agents in parallel: automated bug fixing, best-practices sweeps, lock-based coordination, and real-time tmux monitoring
Dicklesworthstone/claude_code_agent_farm
Install ↓
git clone https://github.com/Dicklesworthstone/claude_code_agent_farm.git /tmp/claude_code_agent_farm && cp -r /tmp/claude_code_agent_farm/.claude/agents/* .claude/agents/Build Claude Code–style deep agents in Python: tool-calling, sandboxed execution, multi-agent teams, skills, checkpoints, and unlimited context — all on Pydantic AI.
vstorm-co/pydantic-deepagents
Install ↓
git clone https://github.com/vstorm-co/pydantic-deepagents.git /tmp/pydantic-deepagents && cp -r /tmp/pydantic-deepagents/.claude/agents/* .claude/agents/Real-time visualization of Claude Code agent orchestration — see your agents think, branch, and coordinate as they work.
patoles/agent-flow
Install ↓
git clone https://github.com/patoles/agent-flow.git /tmp/agent-flow && cp -r /tmp/agent-flow/.claude/agents/* .claude/agents/Supercharge Claude Code with 11 AI agents, 36 commands & 15 skills — the claude-code plugin framework inspired by oh-my-zsh. 6-layer security hooks included. 5-min install.
sangrokjung/claude-forge
Install ↓
git clone https://github.com/sangrokjung/claude-forge.git /tmp/claude-forge && cp -r /tmp/claude-forge/.claude/agents/* .claude/agents/Research · 2
Install ↓
git clone https://github.com/NousResearch/hermes-agent.git /tmp/hermes-agent && cp -r /tmp/hermes-agent/.claude/agents/* .claude/agents/AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary
mvanhorn/last30days-skill
Install ↓
git clone https://github.com/mvanhorn/last30days-skill.git /tmp/last30days-skill && cp -r /tmp/last30days-skill/.claude/agents/* .claude/agents/Devops · 2
Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Panniantong/Agent-Reach
Install ↓
git clone https://github.com/Panniantong/Agent-Reach.git /tmp/Agent-Reach && cp -r /tmp/Agent-Reach/.claude/agents/* .claude/agents/Easily switch between alternative low-cost AI models in Claude Code/Agent SDK. For those comfortable using Claude agents and commands, it lets you take what you've created and deploy fully hosted agents for real business purposes. Use Claude Code to get the agent working, then deploy it in your favo
ruvnet/agentic-flow
Install ↓
git clone https://github.com/ruvnet/agentic-flow.git /tmp/agentic-flow && cp -r /tmp/agentic-flow/.claude/agents/* .claude/agents/Testing · 1
Install ↓
git clone https://github.com/darcyegb/ClaudeCodeAgents.git /tmp/ClaudeCodeAgents && cp -r /tmp/ClaudeCodeAgents/.claude/agents/* .claude/agents/General · 26
Bash is all you need - A nano claude code–like 「agent harness」, built from 0 to 1
shareAI-lab/learn-claude-code
Install ↓
git clone https://github.com/shareAI-lab/learn-claude-code.git /tmp/learn-claude-code && cp -r /tmp/learn-claude-code/.claude/agents/* .claude/agents/omo; the best agent harness - previously oh-my-opencode
code-yeongyu/oh-my-openagent
Install ↓
git clone https://github.com/code-yeongyu/oh-my-openagent.git /tmp/oh-my-openagent && cp -r /tmp/oh-my-openagent/.claude/agents/* .claude/agents/Install ↓
git clone https://github.com/HKUDS/nanobot.git /tmp/nanobot && cp -r /tmp/nanobot/.claude/agents/* .claude/agents/Install ↓
git clone https://github.com/wshobson/agents.git /tmp/agents && cp -r /tmp/agents/.claude/agents/* .claude/agents/A lightweight alternative to OpenClaw that runs in containers for security. Connects to WhatsApp, Telegram, Slack, Discord, Gmail and other messaging apps,, has memory, scheduled jobs, and runs directly on Anthropic's Agents SDK
qwibitai/nanoclaw
Install ↓
git clone https://github.com/qwibitai/nanoclaw.git /tmp/nanoclaw && cp -r /tmp/nanoclaw/.claude/agents/* .claude/agents/A visual, example-driven guide to Claude Code — from basic concepts to advanced agents, with copy-paste templates that bring immediate value.
luongnv89/claude-howto
Install ↓
git clone https://github.com/luongnv89/claude-howto.git /tmp/claude-howto && cp -r /tmp/claude-howto/.claude/agents/* .claude/agents/A powerful GUI app and Toolkit for Claude Code - Create custom agents, manage interactive Claude Code sessions, run secure background agents, and more.
winfunc/opcode
Install ↓
git clone https://github.com/winfunc/opcode.git /tmp/opcode && cp -r /tmp/opcode/.claude/agents/* .claude/agents/A Claude Code plugin that shows what's happening - context usage, active tools, running agents, and todo progress
jarrodwatts/claude-hud
Install ↓
git clone https://github.com/jarrodwatts/claude-hud.git /tmp/claude-hud && cp -r /tmp/claude-hud/.claude/agents/* .claude/agents/Claude Code skill implementing Manus-style persistent markdown planning — the workflow pattern behind the $2B acquisition.
OthmanAdi/planning-with-files
Install ↓
git clone https://github.com/OthmanAdi/planning-with-files.git /tmp/planning-with-files && cp -r /tmp/planning-with-files/.claude/agents/* .claude/agents/A collection of 100+ specialized Claude Code subagents covering a wide range of development use cases
VoltAgent/awesome-claude-code-subagents
Install ↓
git clone https://github.com/VoltAgent/awesome-claude-code-subagents.git /tmp/awesome-claude-code-subagents && cp -r /tmp/awesome-claude-code-subagents/.claude/agents/* .claude/agents/A curated collection of 1000+ agent skills from official dev teams and the community, compatible with Claude Code, Codex, Gemini CLI, Cursor, and more.
VoltAgent/awesome-agent-skills
Install ↓
git clone https://github.com/VoltAgent/awesome-agent-skills.git /tmp/awesome-agent-skills && cp -r /tmp/awesome-agent-skills/.claude/agents/* .claude/agents/Install ↓
git clone https://github.com/rowboatlabs/rowboat.git /tmp/rowboat && cp -r /tmp/rowboat/.claude/agents/* .claude/agents/Turn Claude Code into a full game dev studio — 49 AI agents, 72 workflow skills, and a complete coordination system mirroring real studio hierarchy.
Donchitos/Claude-Code-Game-Studios
Install ↓
git clone https://github.com/Donchitos/Claude-Code-Game-Studios.git /tmp/Claude-Code-Game-Studios && cp -r /tmp/Claude-Code-Game-Studios/.claude/agents/* .claude/agents/A collection of 130+ specialized Codex subagents covering a wide range of development use cases.
VoltAgent/awesome-codex-subagents
Install ↓
git clone https://github.com/VoltAgent/awesome-codex-subagents.git /tmp/awesome-codex-subagents && cp -r /tmp/awesome-codex-subagents/.claude/agents/* .claude/agents/Spec-driven development (SDD) for your team's workflow. Kiro style commands that enforce structured requirements→design→tasks workflow and steering, transforming how you build with AI. Support Claude Code, Codex, Opencode, Cursor, Github Copilot, Gemini CLI and Windsurf.
gotalab/cc-sdd
Install ↓
git clone https://github.com/gotalab/cc-sdd.git /tmp/cc-sdd && cp -r /tmp/cc-sdd/.claude/agents/* .claude/agents/A single hub to find Claude Skills, Agents, Commands, Hooks, Plugins, and Marketplace collections to extend Claude Code, Claude Desktop, Agent SDK and OpenClaw
davepoon/buildwithclaude
Install ↓
git clone https://github.com/davepoon/buildwithclaude.git /tmp/buildwithclaude && cp -r /tmp/buildwithclaude/.claude/agents/* .claude/agents/Shared starter template configuration and CLAUDE.md memory bank system for Claude Code
centminmod/my-claude-code-setup
Install ↓
git clone https://github.com/centminmod/my-claude-code-setup.git /tmp/my-claude-code-setup && cp -r /tmp/my-claude-code-setup/.claude/agents/* .claude/agents/A collection of my favorite custom Claude Code resources to make life easier.
glittercowboy/taches-cc-resources
Install ↓
git clone https://github.com/glittercowboy/taches-cc-resources.git /tmp/taches-cc-resources && cp -r /tmp/taches-cc-resources/.claude/agents/* .claude/agents/Collection of specialized AI subagents for Claude Code for personal use (full-stack development).
lst97/claude-code-sub-agents
Install ↓
git clone https://github.com/lst97/claude-code-sub-agents.git /tmp/claude-code-sub-agents && cp -r /tmp/claude-code-sub-agents/.claude/agents/* .claude/agents/Real-time monitoring for Claude Code agents through simple hook event tracking.
disler/claude-code-hooks-multi-agent-observability
Install ↓
git clone https://github.com/disler/claude-code-hooks-multi-agent-observability.git /tmp/claude-code-hooks-multi-agent-observability && cp -r /tmp/claude-code-hooks-multi-agent-observability/.claude/agents/* .claude/agents/A list of Claude Code Sub-Agents submitted by the community.
hesreallyhim/a-list-of-claude-code-agents
Install ↓
git clone https://github.com/hesreallyhim/a-list-of-claude-code-agents.git /tmp/a-list-of-claude-code-agents && cp -r /tmp/a-list-of-claude-code-agents/.claude/agents/* .claude/agents/Install ↓
git clone https://github.com/wbopan/cui.git /tmp/cui && cp -r /tmp/cui/.claude/agents/* .claude/agents/Open-source AI agent command center for Claude Code agent teams. Built on OpenClaw.
JohnRiceML/clawport-ui
Install ↓
git clone https://github.com/JohnRiceML/clawport-ui.git /tmp/clawport-ui && cp -r /tmp/clawport-ui/.claude/agents/* .claude/agents/Comparison between open source PI agent and closed source Claude Code agent
disler/pi-vs-claude-code
Install ↓
git clone https://github.com/disler/pi-vs-claude-code.git /tmp/pi-vs-claude-code && cp -r /tmp/pi-vs-claude-code/.claude/agents/* .claude/agents/Run Claude Agent (Claude Code) in a sandbox, control it via websocket
dzhng/claude-agent-server
Install ↓
git clone https://github.com/dzhng/claude-agent-server.git /tmp/claude-agent-server && cp -r /tmp/claude-agent-server/.claude/agents/* .claude/agents/CookHero是一个基于 LLM + RAG + Agent + 多模态的智能饮食与烹饪管理平台，支持智能菜谱查询、个性化饮食计划、AI 饮食记录、营养分析、Web 搜索增强，以及可扩展的 ReAct Agent / Subagent 工具体系，帮助厨房新手轻松成为“烹饪英雄”。
Decade-qiu/CookHero
Install ↓
git clone https://github.com/Decade-qiu/CookHero.git /tmp/CookHero && cp -r /tmp/CookHero/.claude/agents/* .claude/agents/Python Development
VS Code setting claudeCode.usePythonEnvironment: true activates your workspace Python env automatically.
Document your Python version, venv path, test command (pytest), linter (ruff), type checker (mypy), and key directories.
Claude understands these frameworks well. Describe your routes, models, and serialization patterns in CLAUDE.md.
Claude can run poetry/pipenv commands. Make sure they are in PATH.
Include data format, expected shape/columns, preprocessing steps in your prompts or CLAUDE.md.
Example CLAUDE.md for a Python/AWS project
# Project Context
## Stack
- Python 3.11, FastAPI, SQLAlchemy
- AWS: ECS Fargate, RDS Postgres, S3, Lambda
- Docker, Terraform (infra/)
- pytest, ruff, mypy
## Commands
- Run tests: `pytest tests/ -v`
- Lint: `ruff check src/`
- Type check: `mypy src/`
- Deploy: `./scripts/deploy.sh staging`
## Key Directories
- src/api/ — FastAPI route handlers
- src/models/ — SQLAlchemy models
- src/services/ — Business logic
- infra/ — Terraform configs
- tests/ — Test suites
## Conventions
- Use functional approach, no classes for services
- All API responses use Pydantic response models
- Write tests for all new endpoints
- Follow Conventional CommitsAWS & DevOps
Set CLAUDE_CODE_USE_BEDROCK=1 to route all Claude calls through your AWS account. No Anthropic API key needed.
export CLAUDE_CODE_USE_BEDROCK=1Use anthropics/claude-code-action@v1 with use_bedrock: true and AWS_ROLE_TO_ASSUME secret.
Claude can run aws CLI commands via Bash tool. Pre-approve with --allowedTools 'Bash(aws *)'.
Claude reads and writes Terraform/CDK. Put your backend config and state bucket info in CLAUDE.md.
Create a custom skill in .claude/skills/deploy/SKILL.md with your deploy steps. Claude follows it every time.
Use --max-budget-usd and --max-turns in CI pipelines to cap spending.
claude -p --max-budget-usd 2.00 --max-turns 5 "review this PR"Example: GitHub Actions PR review with AWS Bedrock
name: Claude PR Review
on:
pull_request:
types: [opened, synchronize]
jobs:
review:
runs-on: ubuntu-latest
permissions:
id-token: write
contents: read
pull-requests: write
steps:
- uses: aws-actions/configure-aws-credentials@v4
with:
role-to-assume: ${{ secrets.AWS_ROLE_TO_ASSUME }}
aws-region: us-east-1
- uses: anthropics/claude-code-action@v1
with:
use_bedrock: "true"
claude_args: "--model us.anthropic.claude-sonnet-4-6 --max-turns 3"
prompt: "Review this PR for bugs, security issues, and code quality."Power Tips
Start with --permission-mode plan. Review the plan, then switch to acceptEdits to execute. Safer for big changes.
Use claude -w feature-name to work on multiple branches simultaneously without conflicts.
If you have been chatting for a while, run /compact before starting a new big task. Frees context.
Include 'ultrathink' in your prompt to get maximum reasoning depth on complex debugging.
Use @filename.py in your prompt to reference files without pasting. Claude reads them automatically.
Pipe files, diffs, or logs into claude -p for automated analysis in CI/CD.
git diff main | claude -p "review this diff for security issues"Use /batch for changes spanning 5-30 files. Creates isolated worktree agents that work in parallel.
Create .claude/skills/<name>/SKILL.md for repeatable workflows (deploy, test, review). Claude follows them exactly.
Add a PostToolUse hook on Edit|Write to auto-run your formatter (prettier, black, ruff) after every file change.
Use /loop 5m 'check test status' to poll a long-running CI build without babysitting.
Daily Workflow Examples
Morning: review overnight PRs
claude -p "list open PRs on this repo and summarize changes"claude -c # continue to review specific PR/review 42 # deep review of PR #42Feature: add a new API endpoint
claude --permission-mode plan"add a GET /api/users/{id}/activity endpoint"# review plan → Shift+Tab to acceptEdits → Claude executesDebug: production error
cat /var/log/app/error.log | claude -p "what caused this and how to fix"claude -c # continue to implement the fixInfra: update Terraform
claude -w infra-update # isolated worktree"add a new RDS read replica in eu-west-1""run terraform plan and show me the diff"CI/CD: automated review
git diff main | claude -p --max-turns 3 --max-budget-usd 1.00 "review for bugs"