Every command, shortcut, and workflow tip for both tools in one place
Turn what you learned into a concrete stack decision.
01AI Agent ToolboxBuild autonomous AI agents that can read, write, and actWant the shortlist in your inbox?
Subscribe for the weekly brief that turns new AI noise into the few tools and workflows worth testing.
Curated bundles that help you move from this guide into a working stack.
Guide
OpenClaw Skills and Plugins
Extend your agent's capabilities with web search, browser automation, memory, and more
Guide
20 GitHub Repos Every AI Builder Should Star
The curated shortlist of repos that power AI agents, MCP servers, and Web3 tools
Guide
Claude Code + Agents: Advanced Configuration and Autonomous Workflows
CLAUDE.md architecture, subagents, hooks, MCP servers, and building Claude pipelines that run without you
These commands manage your OpenClaw installation, agents, and gateway.
| Command | What it does |
|---------|-------------|
| openclaw setup | Initialize OpenClaw configuration and workspace |
| openclaw onboard | Interactive wizard for gateway, workspace, and skills |
| openclaw gateway | Start the WebSocket gateway (runs your agents) |
| openclaw status | Show gateway status, connected agents, and channels |
| Command | What it does |
|---------|-------------|
| openclaw agent | Run a single agent turn (for testing) |
| openclaw agent --message "text" | Send a message to the agent and get a response |
| openclaw agents list | List all configured agents |
| openclaw agents create <name> | Create a new isolated agent |
| Command | What it does |
|---------|-------------|
| openclaw configure | Set up credentials, devices, agent defaults |
| openclaw models | Discover and configure LLM models |
| openclaw channels telegram setup | Configure Telegram bot connection |
| openclaw channels telegram allow <id> | Add a Telegram user to the allowlist |
| Command | What it does |
|---------|-------------|
| openclaw plugins list | Show installed plugins |
| openclaw plugins install <name> | Install a plugin |
| openclaw plugins enable <name> | Enable a disabled plugin |
| openclaw plugins disable <name> | Disable a plugin without removing it |
| Command | What it does |
|---------|-------------|
| openclaw cron list | Show all scheduled jobs |
| openclaw cron add | Create a new cron job |
| openclaw cron remove <id> | Delete a cron job |
| Command | What it does |
|---------|-------------|
| openclaw daemon start | Run the gateway as a background service |
| openclaw daemon stop | Stop the background service |
| openclaw daemon status | Check if the daemon is running |
| openclaw memory search <query> | Search the agent's memory |
Claude Code is Anthropic's terminal-based AI coding tool. These commands work inside a Claude Code session.
| Command | What it does |
|---------|-------------|
| /help | Show all available commands |
| /clear | Clear conversation history and start fresh |
| /compact | Compress conversation history to save context |
| /model | Switch between Claude models (Opus, Sonnet, Haiku) |
| /cost | Show token usage and cost for this session |
| Command | What it does |
|---------|-------------|
| /init | Create a CLAUDE.md file for the current project |
| /review | Review recent code changes |
| /pr-comments | Summarize and address PR review comments |
| Command | What it does |
|---------|-------------|
| /fast | Toggle fast mode (same model, faster output) |
| /status | Show current session status |
| /config | View or modify configuration |
These work inside the Claude Code terminal interface:
| Shortcut | Action |
|----------|--------|
| Enter | Send message (when input is single-line) |
| Shift+Enter | New line in input |
| Escape | Cancel current operation |
| Ctrl+C | Interrupt running tool or exit |
| Up arrow | Previous message in history |
| Down arrow | Next message in history |
| Tab | Accept autocomplete suggestion |
CLAUDE.md files are how you configure Claude Code's behavior for your project. They work at two levels:
~/.claude/CLAUDE.md applies to every project you open in Claude Code. Put your preferences here:
# Global Preferences
- Default to Bun instead of npm
- Keep responses concise
- Always use TypeScript
./CLAUDE.md in your project root applies only to that project:
# Project: My App
- Use Tailwind for styling
- Database: PostgreSQL via Drizzle ORM
- Run tests with: bun test
Project-level CLAUDE.md overrides user-level for that project. Claude Code reads both and merges them.
Your ~/.claude/ directory contains:
~/.claude/
├── CLAUDE.md # Global instructions
├── settings.json # Permissions and plugin config
├── settings.local.json # Local overrides (not synced)
├── commands/ # Custom slash commands
├── scripts/ # Utility scripts
├── hooks/ # Event hooks (status line, etc.)
└── plugins/ # Installed plugins
If you use the GSD (Get Shit Done) workflow system with Claude Code, these commands manage your project lifecycle:
| Command | What it does |
|---------|-------------|
| /gsd:new-project | Initialize a new project (questions → research → roadmap) |
| /gsd:new-milestone | Start a new milestone cycle |
| Command | What it does |
|---------|-------------|
| /gsd:discuss-phase N | Gather context for phase N through questions |
| /gsd:plan-phase N | Create a detailed plan for phase N |
| /gsd:execute-phase N | Build phase N from its plan |
| /gsd:verify-work N | Verify phase N passes all requirements |
| Command | What it does |
|---------|-------------|
| /gsd:quick | Execute a small task with atomic commits |
| /gsd:add-todo | Add a task to the backlog |
| /gsd:check-todos | List pending todos |
| Command | What it does |
|---------|-------------|
| /gsd:pause-work | Snapshot current state for later resume |
| /gsd:resume-work | Continue from a paused session |
| /gsd:progress | Check project progress and decide next action |
| Command | What it does |
|---------|-------------|
| /gsd:help | Show all GSD commands |
| /gsd:settings | Configure GSD workflow options |
| /gsd:debug | Start a systematic debugging session |
| /gsd:map-codebase | Analyze codebase structure |
The power combo: use Claude Code as your coding assistant and OpenClaw as your autonomous agent.
Claude Code (terminal)
└── Your CTO — reads code, builds features, commits
└── Uses: /gsd:quick, /gsd:execute-phase, /review
OpenClaw (background)
└── Your assistant — monitors, researches, alerts
└── Uses: Telegram, cron jobs, memory, web search
Claude Code can send messages to your Telegram (where OpenClaw also lives):
# From Claude Code, notify via Telegram
~/.claude/scripts/telegram-notify.sh --update "Feature X is done"
And OpenClaw can be configured to watch for specific events and react.
| Task | Use | |------|-----| | Write code | Claude Code | | Refactor a file | Claude Code | | Run tests | Claude Code | | Research a topic | OpenClaw | | Monitor prices | OpenClaw | | Scheduled reminders | OpenClaw | | Deploy an app | Claude Code | | Daily briefing | OpenClaw |
The key insight: Claude Code is a session-based tool (you start it, work, stop it). OpenClaw is always-on (daemon/gateway running in the background). Use each for what it's designed for.
Ready to discover the best tools for your projects? Check out 20 GitHub Repos Every AI Builder Should Star.