
What You Can Do Via Chat
Manage your AI agent through chat. No SSH needed for normal operations. Learn what you can do just by talking to your agent.
Your AI agent isn't just for conversation — it's your interface to the entire system. Almost everything can be done through chat. No SSH needed for normal operations.
Just Ask Your Agent — Your AI agent has full access to its own system. It can edit files, run commands, install skills, restart itself, and more. Think of it as your system administrator that happens to be an AI. Just describe what you want in natural language.
Configuration
Change AI model
Switch between Claude, GPT-4, and other models.
/model
switch to Claude Opus
use GPT-4 instead
Edit agent personality
Your agent can edit its own SOUL.md file.
make yourself more professional
be more casual and funny
update your personality to focus on marketing
Add API keys
Your agent can write to auth files securely.
add my OpenRouter key: sk-xxx
update the Anthropic API key
save this OpenAI key for me
Skills & Updates
Install new skills
Browse and install capabilities from clawhub.
install the weather skill
show me available skills
add email support
Update OpenClaw
Your agent can update itself.
update yourself
check for updates
upgrade to the latest version
Restart gateway
Restart services when needed.
restart yourself
reboot the gateway
restart after config changes
Automation
Set up reminders
One-time or recurring notifications.
remind me tomorrow at 9am to call John
set a reminder for my meeting in 2 hours
remind me every Monday at 8am to review goals
Create cron jobs
Schedule recurring tasks.
run a backup every night at midnight
check my email every hour
post daily updates at 5pm
Daily Operations
Search the web
Real-time information from the internet.
what's the latest news on AI?
search for restaurants near me
find articles about React 19
Manage files
Create, edit, and organize workspace files.
create a new markdown file
edit my todo list
organize my project folder
Run commands
Execute shell commands and scripts.
run git status
check disk space
list running processes
Backup workspace
Your agent can create and manage backups.
backup my workspace
create a tar archive of everything
zip up my project files
View system status
Check health and performance.
/status
show me system info
how are you doing?
Pro Tips
🗣️ Be conversational
You don't need to memorize commands. Just say what you want: "Can you update yourself?" works just as well as "/update".
🔄 Let it figure out the details
Your agent knows how to accomplish tasks. Say "install the weather skill" and it will run the right commands, handle errors, and let you know when it's done.
📝 Ask for documentation
Your agent can read its own config files, check available skills, and explain how things work. Just ask "What skills do I have installed?" or "Show me my config".
🛠️ It can fix itself
If something's not working, tell your agent. It can check logs, restart services, and troubleshoot issues. "You're not responding quickly — can you check your status?"
✅ The Bottom Line
Your AI agent is designed to be self-managing. You interact with it through chat, and it handles the technical details. SSH access is there if you need it, but for 95% of tasks, just ask your agent. It's faster, easier, and safer.