Ever Wondered How Powerful a Well-Written Prompt Can Be?

Mohammad Shariya — Sat, 24 May 2025 14:57:15 +0000

Lately, I’ve been experimenting with writing clear, structured prompts for AI tools—and honestly, the results have been impressive!

I realized that when I tell the AI exactly what I want, the output becomes much more useful and aligned. So I started following a simple prompt-writing formula that’s been giving me surprisingly consistent and smart results.

And I thought—why not share my little experience with the community?

Here's how I structure prompts that actually work:

System Instruction
→ Tell the AI what “role” it’s supposed to play. (e.g., a teacher, developer, assistant)
Context
→ Give it the background or situation—so it understands the purpose.
Few-shot Examples
→ Provide 1–2 short examples to guide the kind of output you expect.
User Task
→ Clearly define what the AI needs to do.
Requirements
→ Mention tone, style, or any must-haves to guide the final response.

Example:

System: You are a research assistant.  
Context: Helping to summarize academic papers for social media.  
Few-shot Examples: "Key findings", "Why this matters", "Future work".  
User Task: Summarize a paper on machine learning in healthcare.  
Requirements: Use simple language, bullet points, include 1 key takeaway.

When I follow these five steps, the prompts become much more effective, and the results feel genuinely tailored. It's become my go-to style whenever I work with AI tools now.

How do you write prompts?

Do you follow a specific structure, or have any tips of your own?
Drop a comment—I'd love to learn from your approach too!

Writing Better Git Commit Messages — What I Learned from Experience

Mohammad Shariya — Fri, 23 May 2025 15:29:00 +0000

Is writing a good commit message really that important?

After working on projects for quite some time, I’ve realized that a clear, to-the-point commit message makes teamwork a whole lot easier — and it helps me too when reviewing or debugging code later!

Lately, I’ve been more intentional about how I write my commits, and honestly, the entire process has become smoother and more enjoyable.

Why do good commit messages matter?

Easier team collaboration: No more guessing games about who changed what.
Faster debugging: Tools like git blame and git bisect are only useful when the messages give actual clues.
Quicker code reviews: A good message helps the reviewer understand what to focus on instantly.

The format I follow (and it really works!)

I stick to a simple, consistent format:

<type>(<scope>): <short, clear summary>

type: feat, fix, docs, style, refactor, test
scope: the part of the project (like auth, ui)
summary: under 50 characters, written in the imperative mood (e.g., “Add”, “Fix”, “Refactor”)

If needed, I also add:

Body: What I did and why I did it (not just what)
Footer: Issue numbers (e.g., Closes #123) or notes about breaking changes

A few real examples I use:

feat(api): add user-profile endpoint
fix(ui): prevent navbar overflow on mobile
docs(readme): update installation instructions
refactor(auth): extract token logic into service
test(payment): add unit tests for Stripe webhook handler

What I avoid now:

❌ Vague messages like "stuff" or "bugfix"
❌ Non-descriptive ones like "updated files"

Since I started following this approach, things have become more efficient — both for me and my team. Less confusion, fewer meetings, and faster progress. 😄