Build Your First OpenClaw Workflow: A Hands-On Tutorial

2026-02-02 · Nia

Theory is great, but you learn by doing. In this hands-on tutorial, we'll build a complete OpenClaw workflow from scratch—an email monitoring system that alerts you to important messages.

What We're Building

By the end of this tutorial, you'll have:

Email monitoring that checks your inbox every 30 minutes

Smart filtering that identifies urgent emails

Telegram alerts for messages that need attention

Daily digest summarizing all emails

Time Required: 30 minutes

Skill Level: Beginner

Prerequisites

Before starting, ensure you have:

OpenClaw installed and running (installation guide)
Telegram configured as a channel
Gmail account (or any IMAP-compatible email)

Step 1: Set Up Email Access

First, let's give OpenClaw access to your email.

Create an App Password (Gmail)

Go to Google Account Security

Enable 2-Factor Authentication if not already enabled

Go to App Passwords

Create a new app password for "OpenClaw"

Copy the 16-character password

Store Credentials Securely

# Create config directory
mkdir -p ~/.config/gmail

Store credentials (replace with your values)
echo "your-email@gmail.com" > ~/.config/gmail/email
echo "xxxx-xxxx-xxxx-xxxx" > ~/.config/gmail/app_password

Secure the files
chmod 600 ~/.config/gmail/*

Step 2: Create the Email Checker Script

Create a new file in your workspace:

mkdir -p ~/openclaw/workspace/tools

Create tools/email_checker.py:

#!/usr/bin/env python3
"""
Email Checker for OpenClaw
Checks Gmail inbox and returns summary of new/important emails
"""

import imaplib
import email
from email.header import decode_header
from pathlib import Path
from datetime import datetime, timedelta

Load credentials
EMAIL = Path.home() / ".config/gmail/email"
PASSWORD = Path.home() / ".config/gmail/app_password"

def get_credentials():
return (
EMAIL.read_text().strip(),
PASSWORD.read_text().strip()
)

def check_emails(hours_back=24, max_emails=20):
"""Check for new emails in the last N hours."""
username, password = get_credentials()

# Connect to Gmail
mail = imaplib.IMAP4_SSL("imap.gmail.com")
mail.login(username, password)
mail.select("inbox")

# Search for recent emails
since_date = (datetime.now() - timedelta(hours=hours_back)).strftime("%d-%b-%Y")
_, messages = mail.search(None, f'(SINCE "{since_date}")')

email_ids = messages[0].split()[-max_emails:]  # Get last N

results = []
for email_id in reversed(email_ids):  # Most recent first
_, msg_data = mail.fetch(email_id, "(RFC822)")
msg = email.message_from_bytes(msg_data[0][1])

# Decode subject
subject, encoding = decode_header(msg["Subject"])[0]
if isinstance(subject, bytes):
subject = subject.decode(encoding or "utf-8")

# Get sender
sender = msg.get("From", "Unknown")

# Get date
date = msg.get("Date", "Unknown")

# Check if unread (simplified check)
_, flags_data = mail.fetch(email_id, "(FLAGS)")
is_unread = b"\\Seen" not in flags_data[0]

results.append({
"subject": subject,
"from": sender,
"date": date,
"unread": is_unread
})

mail.logout()
return results

def is_urgent(email_data):
"""Determine if an email is urgent based on keywords."""
urgent_keywords = [
"urgent", "asap", "important", "action required",
"deadline", "immediate", "critical", "emergency"
]

subject_lower = email_data["subject"].lower()
return any(keyword in subject_lower for keyword in urgent_keywords)

def format_summary(emails):
"""Format emails into a readable summary."""
if not emails:
return "📭 No new emails in the last 24 hours."

unread = [e for e in emails if e["unread"]]
urgent = [e for e in emails if is_urgent(e)]

summary = f"📧 Email Summary\n\n"
summary += f"Total: {len(emails)} | Unread: {len(unread)} | Urgent: {len(urgent)}\n\n"

if urgent:
summary += "🚨 Urgent:\n"
for e in urgent:
summary += f"  • {e['subject'][:50]}\n    From: {e['from'][:30]}\n"
summary += "\n"

if unread:
summary += "📬 Unread:\n"
for e in unread[:5]:  # Show top 5
summary += f"  • {e['subject'][:50]}\n    From: {e['from'][:30]}\n"

return summary

if __name__ == "__main__":
emails = check_emails()
print(format_summary(emails))

Make it executable:

chmod +x tools/email_checker.py

Test It

python3 tools/email_checker.py

You should see a summary of your recent emails.

Step 3: Create an OpenClaw Skill

Skills tell OpenClaw how to perform specific tasks. Create skills/email/SKILL.md:

# Email Management Skill

Check Emails

When asked to check emails:
Run python3 tools/email_checker.py
Report the summary
Offer to read specific emails if any are urgent

Mark as Read

When asked to mark an email as read:
Currently not supported - ask user to mark in Gmail

Compose Email

When asked to compose/draft an email:
Ask for recipient, subject, and key points
Draft the email
Save to workspace/drafts/
Notify user for review

Step 4: Set Up Automatic Checking

Now let's make OpenClaw check emails automatically using cron jobs.

Option A: Using HEARTBEAT.md

Add to your HEARTBEAT.md:

## Email Check (every heartbeat)

If it's been more than 30 minutes since last email check:
Run email checker script
If urgent emails exist, alert immediately
Log check time to memory/heartbeat-state.json

Option B: Using Cron Jobs

Create a scheduled task:

# Through OpenClaw

"Schedule an email check every 30 minutes that alerts me to urgent emails"

Or manually via the cron tool:

name: "📧 Email Check" schedule: kind: every everyMs: 1800000 # 30 minutes payload: kind: systemEvent text: "Check emails and alert if any urgent messages"

sessionTarget: main

Step 5: Add Daily Digest

Create a daily summary that runs every evening.

Add to HEARTBEAT.md or create a cron job for 6 PM:

## Daily Email Digest (6 PM)

At 6 PM:
Run email checker for last 24 hours
Categorize: Work, Personal, Newsletters, Other
Summarize key threads
List action items
Send digest to Telegram

Step 6: Test the Complete Workflow

Let's test everything works together.

Manual Test

Send yourself a test email with "URGENT" in the subject, then:

"Check my emails"

OpenClaw should detect and highlight the urgent email.

Automation Test

Wait for the scheduled check (or trigger it manually):

"Run the email heartbeat check now"

Verify Logging

Check that state is being tracked:

cat memory/heartbeat-state.json

Extending the Workflow

Now that the basic workflow works, you can extend it:

Add Priority Rules

def get_priority(email_data):
"""Assign priority based on sender and content."""
vip_senders = ["boss@company.com", "client@important.com"]

sender = email_data["from"].lower()

if any(vip in sender for vip in vip_senders):
return "high"
elif is_urgent(email_data):
return "high"
else:
return "normal"

Add Response Drafting

## Auto-Draft Responses

For routine emails (meeting confirmations, etc.):
Identify email type
Generate appropriate response
Save to drafts folder
Notify user for review

Add Email Threading

Group related emails together:

def group_by_thread(emails):
"""Group emails by conversation thread."""
threads = {}
for e in emails:
# Simple thread detection by subject
subject = e["subject"].lower()
subject = subject.replace("re:", "").replace("fwd:", "").strip()

if subject not in threads:
threads[subject] = []
threads[subject].append(e)

return threads

Troubleshooting

"Authentication Failed"

Verify app password is correct
Ensure 2FA is enabled on Gmail
Check if "Less secure app access" is needed (shouldn't be with app passwords)

"No Emails Found"

Check the date range (default is 24 hours)
Verify the IMAP server is accessible
Try mail.select("INBOX") instead of mail.select("inbox")

"Skill Not Found"

Ensure skill directory is in the right location
Check SKILL.md syntax
Restart OpenClaw to reload skills

What You've Learned

Congratulations! You've built a complete workflow that:

✅ Connects to external services (Gmail)

✅ Processes data (email parsing)

✅ Makes decisions (urgency detection)

✅ Automates tasks (scheduled checks)

✅ Communicates results (Telegram alerts)

This same pattern applies to any integration: connect, process, decide, automate, communicate.

Next Steps

Add more email providers (Outlook, Yahoo)
Integrate with calendar for meeting-related emails
Build a web dashboard for email analytics
Add sentiment analysis for customer emails

For more OpenClaw tutorials, see automation with cron jobs and 10 real-world use cases.

Build more with OpenClaw. Find resources at docs.openclaw.ai.