Daily Briefing Bot

Sends a morning AI briefing to your Telegram every weekday at 8 AM. Reads your calendar, unread emails summary, and news feed.

git clone https://github.com/openclaw-ai/example-daily-briefing
cd example-daily-briefing
cp .env.example .env  # Add OPENCLAW_API_KEY and TELEGRAM_BOT_TOKEN
pip install -r requirements.txt
# Schedule with cron or systemd

Automated Code Reviewer

GitHub Action that reviews every PR and posts a comment with security findings, style issues, and suggestions.

# .github/workflows/ai-review.yml
name: AI Code Review
on: [pull_request]
jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with: {fetch-depth: 0}
      - run: pip install openclaw
      - run: |
          git diff origin/${{ github.base_ref }}...HEAD > pr-diff.txt
          openclaw run "Review this PR diff for security, bugs, and style." \
            --file pr-diff.txt \
            --output review.md
      - uses: actions/github-script@v7
        with:
          script: |
            const fs = require('fs')
            const body = fs.readFileSync('review.md', 'utf8')
            github.rest.issues.createComment({...context.repo, issue_number: context.issue.number, body})

Document Processing Pipeline

Watches a folder for new PDFs, extracts structured data to JSON, and posts a summary to Slack.

git clone https://github.com/openclaw-ai/example-doc-pipeline
# Processes: PDF -> JSON extraction -> Slack notification

Data Anomaly Monitor

Runs every hour, checks your metrics CSV for anomalies, and sends an alert if anything unusual is detected.

git clone https://github.com/openclaw-ai/example-anomaly-monitor

Advanced Sample Projects

Multi-Agent Research Pipeline

Orchestrates three specialized agents: a researcher, a fact-checker, and a writer. Each runs as a separate OpenClaw task passing its output to the next:

#!/usr/bin/env bash
# multi-agent-research.sh
TOPIC="$1"

echo "=== Phase 1: Research ==="
openclaw run "Research the topic: $TOPIC. Produce 10 key facts with sources." \
  --model gpt-4o > /tmp/research.md

echo "=== Phase 2: Fact-check ==="
openclaw run "Fact-check the following claims. Flag any that are uncertain or false." \
  --file /tmp/research.md --model claude-3-5-sonnet > /tmp/checked.md

echo "=== Phase 3: Write article ==="
openclaw run "Write a 600-word blog post based on these verified facts." \
  --file /tmp/checked.md --model gpt-4o > "article-$TOPIC.md"

echo "Done. Article saved to article-$TOPIC.md"

Self-Healing Deploy Script

Deploys your application and, on failure, asks OpenClaw to diagnose and fix the issue:

#!/usr/bin/env bash
set -uo pipefail

deploy_output=$(./deploy.sh 2>&1) || {
  echo "Deployment failed. Asking AI to diagnose..."
  FIX=$(echo "$deploy_output" | openclaw run \
    "This deploy script failed. Diagnose the error and provide the exact shell \
     commands to fix it. Be concise." --stdin --model gpt-4o)
  echo "=== Suggested Fix ==="
  echo "$FIX"
  exit 1
}

Config File Templates

Download ready-to-use config.yaml templates for common setups:

TemplateUse CaseDownload
Minimal (OpenAI)Single user, fast startminimal.yaml
Multi-providerSwitch between OpenAI, Claude, Ollamamulti-provider.yaml
Local / OfflineOllama only, no cloudlocal.yaml
CI/CD ServerAutomated pipelines, no interactivityci.yaml
High SecuritySandboxed, read-only, no execsecure.yaml

Community-Contributed Projects

These projects were built and open-sourced by OpenClaw community members:

  • openclaw-notion-sync — Sync OpenClaw task summaries to Notion databases automatically
  • openclaw-jira-bot — Auto-create and update Jira tickets from code diffs and error logs
  • openclaw-obsidian — Obsidian plugin: run OpenClaw tasks on your notes directly from the editor
  • openclaw-docker-healthcheck — Container monitoring that uses OpenClaw to explain and suggest fixes for container failures

Want to submit your project? Post it in GitHub Discussions: Show and Tell →

Starter Templates

Copy these minimal scripts to get started without writing from scratch:

Daily Briefing (5 lines)

#!/usr/bin/env bash
# daily-briefing.sh — add to crontab: 0 8 * * 1-5
TASKS=$(head -5 ~/todo.txt 2>/dev/null)
openclaw run "Write a 3-sentence morning briefing for a developer. Tasks today: $TASKS" \
  --model gpt-4o-mini | openclaw telegram send

File Organizer

#!/usr/bin/env bash
# organise-downloads.sh
openclaw run \
  "List every file in ~/Downloads and suggest which of these folders each file belongs in:
   work/, personal/, installers/, docs/.
   Output a shell mv command for each file." \
  --model gpt-4o-mini

Weekly Git Summary

#!/usr/bin/env bash
# weekly-git-summary.sh
git log --since="7 days ago" --oneline | \
  openclaw run "Write a one-paragraph summary of what changed this week based on these commit messages."

Tips for Running Sample Projects

  • Start with --dry-run — run openclaw run --dry-run first to see what the agent plans without executing side effects
  • Use a dedicated venv — sample projects may have different dependency requirements; isolate them per project
  • Set max_steps conservatively — for new projects, start with max_steps: 10 and increase as you understand the task scope
  • Log everything initially — set log_level: debug when first running a sample to understand the agent's reasoning chain
  • Add to your own projects — each sample includes a config.yaml you can copy and modify; see Config Examples for more patterns

Share your projects with the community in GitHub Discussions Show & Tell.