Most developers use ChatGPT wrong.
They type vague questions like write a Python script to monitor my server and get mediocre, half-working code that needs hours of fixing.
The secret? Structured prompts. Specific inputs, outputs, error handling, deployment context — all in one prompt.
Here are 5 prompts I use daily that generate clean, working Python scripts in seconds.
Prompt #1 — Automated File Organizer
The problem: Downloads folder with 500 random files. Finding anything takes forever.
The prompt:
Write a Python script that automatically organizes files in a folder by their extension.
Create subfolders (Images, Documents, Videos, Archives, Code, Other) and move files accordingly.
Add logging so I can see what was moved. Include error handling for locked files.
What you get: 60 lines of clean Python. Handles edge cases, logs every move, skips files it can't touch.
Run it as a cron job every night. Never think about your downloads folder again.
Prompt #2 — System Monitor with Telegram Alerts
The problem: Server goes down at 3AM. You find out at 9AM from angry users.
The prompt:
Write a Python script that monitors system resources:
- CPU usage, RAM, disk space
- Sends alert via Telegram when thresholds exceeded (CPU >85%, disk >80%)
- Logs metrics to CSV every 5 minutes
- Runs as a background daemon Include setup instructions for Telegram bot.
What you get: A complete monitoring daemon with Telegram integration, CSV logging, and threshold alerts. Your phone becomes a server dashboard.
Prompt #3 — Automated Email Reporter
The problem: Every Monday, manually copying data into an email report. Every. Single. Week.
The prompt:
Write a Python script that:
- Reads data from a CSV file
- Generates a summary report with key stats
- Sends it via Gmail SMTP automatically every Monday at 8AM
- Uses environment variables for credentials (never hardcoded) Include full setup instructions.
What you get: Set it up once, forget about it. Report lands in inboxes every Monday automatically.
Prompt #4 — Web Scraper with Pagination
The problem: Need data from a website that has 50 pages. Manual copy-paste would take hours.
The prompt:
Create a Python web scraper using Beautiful Soup and requests that:
- Scrapes product listings (name, price, URL) from a website
- Handles pagination automatically
- Saves results to CSV with timestamp
- Respects robots.txt and adds random delays between requests
- Handles connection errors with retry logic
What you get: A respectful, production-ready scraper that handles real-world edge cases — timeouts, missing data, pagination breaks.
Prompt #5 — Database Backup with Auto-Cleanup
The problem: Manual backups that nobody does consistently. Until something breaks.
The prompt:
Write a Python script that automatically backs up a SQLite database:
- Creates timestamped backup files
- Compresses them with gzip
- Keeps only the last 7 backups (deletes older ones)
- Sends a Telegram notification on success or failure
- Designed to run as a cron job
What you get: Set-and-forget backup system. Runs nightly, cleans itself up, notifies you either way.
Why These Prompts Work
Notice the pattern in every prompt:
- Numbered requirements — ChatGPT handles lists better than paragraphs
- Specific constraints — last 7 backups, CPU >85%, every Monday at 8AM
- Deployment context — designed for cron, runs as daemon
- Security requirements — use environment variables, never hardcode
The more specific you are, the better the output.
Pro Tips
- Add make it production-ready to any prompt for better error handling
- Add explain each section with comments if you're learning
- Use GPT-4 for longer, more complex scripts — GPT-3.5 struggles with multi-requirement prompts
Want More?
I compiled 50 prompts like these into a single PDF, organized into 5 categories:
- Python Automation (10 prompts)
- Bash & Shell Scripts (10 prompts)
- Workflow & No-Code Automation (10 prompts)
- DevOps & Infrastructure (10 prompts)
- AI & ChatGPT Integration (10 prompts)
👉 Get the full pack here — $8, instant download.
What do you automate most often? Drop it in the comments — I might add it to the next version.
Top comments (0)