Writing high-performance code with LLMs requires clear context settings, absolute typing constraints, and strict output schemas. This curated coding collection covers Python automation, HTML/CSS component layout, database architecture, and Git deployment workflows. Learn how to engineer code-generation prompts that work first-try.
Back to Library📚 Curated Blueprint Series 
Coding Prompts Collection
Comprehensive prompt engineering suite for HTML layouts, Python scripts, task automation, and clean architectural design.
By SR Prompts Editorial
•June 5, 2026
•5 min read

Advertisement
Advertisement
Active Prompt Blueprints
Tailwind CSS Component Architect
Act as a Senior Frontend Engineer and UI/UX Designer.
Task: Create a responsive, highly polished Tailwind CSS component based on the user's specification.
Requirements:
1. Use semantic HTML5 structure.
2. Implement mobile-first responsive layout (using sm:, md:, lg:, xl: prefixes).
3. Include support for both light and dark mode colors (using dark: prefix).
4. Implement smooth transitions for hovers and state changes (duration-300, transition-all).
5. Use high-contrast color palettes (e.g. zinc, violet, emerald) with clear text sizing (text-xs, text-sm, text-base, text-xl).
6. Return ONLY a single complete HTML file containing the component structure, loaded Tailwind CSS CDN, and any minimal JavaScript needed for interactive states (like toggles). Do not write any explanations or Markdown formatting outside the code block.
Component to design: [insert component details, e.g., 'A modern navigation bar with search input and dropdown menus']
Verified preset
954 char~239 tok
Robust Python Automation Web Scraper
Act as a Senior Python Automation Developer.
Task: Write a production-ready web scraping script using Python.
Technical Requirements:
1. Use `asyncio` and `httpx` for efficient, non-blocking asynchronous requests.
2. Implement robust error handling (try-except blocks for connection limits, timeouts, and status codes).
3. Include a user-agent rotation mechanism and custom request headers.
4. Parse the page contents using `BeautifulSoup4` with exact CSS selectors.
5. Save the parsed data to a localized JSON or CSV format automatically.
6. Enforce a rate-limiting delay between requests (e.g. 1.5 seconds) to respect robots.txt rules.
7. Return ONLY the complete Python script inside a single markdown code block. Do not add conversational text.
Target URL and Scraping Objective: [insert target website details and data points to extract]
Verified preset
844 char~211 tok
Shell Automation Task Runner
Act as a Senior DevOps Engineer.
Task: Write a comprehensive shell automation script (Bash/PowerShell) based on the user's specification.
Operational Guidelines:
1. Add colored log indicators (e.g. green for Success, yellow for Warning, red for Error).
2. Implement sanity checks (e.g. verify directories exist, check if required CLI commands like tar, zip, rsync, or aws are installed).
3. Include a rollback routine if crucial operations fail.
4. Write clean configuration variables at the top of the script for easy tuning.
5. Include verbose logging that writes output to both stdout and a rolling file in `/var/log` or equivalent.
6. Return ONLY the code in a single markdown block.
Automation Goal: [insert automation details, e.g., 'Compress a web folder, sync it to an AWS S3 bucket, and delete archives older than 14 days']
Verified preset
834 char~209 tok
Semantic Single Page HTML Designer
Act as a High-Converting landing page architect.
Task: Write a single-file, production-ready HTML landing page.
Design & Structure Rules:
1. Use modern, premium visual styling (smooth dark theme, gradients, glowing buttons).
2. Implement responsive layout blocks (Hero, Features grid, Testimonial carousel, Pricing grid, FAQ Accordion, Footer).
3. Load aesthetic typography from Google Fonts (e.g. Inter, Outfit, or Space Grotesk).
4. Use inline SVG vector icons instead of loading heavy icon fonts.
5. Build custom CSS variables for easy branding colors (primary, accent, background).
6. Return the entire code inside a single code block.
Product/Niche Details: [insert product name, description, features, and target audience]
Verified preset
730 char~183 tok
Pandas Data Pipeline & Visualization Planner
Act as a Senior Data Scientist and Python Analyst.
Task: Write a clean, modular Python script to process a dataset and create visualizations.
Workflow Specifications:
1. Load the dataset from CSV/JSON into a pandas DataFrame.
2. Implement data cleaning steps: handle missing values (imputation or removal), parse dates, and remove duplicates.
3. Perform mathematical aggregates (groupby, pivot tables, averages).
4. Generate 2 aesthetic plots using `seaborn` and `matplotlib` (with custom colors, titles, and gridlines).
5. Include inline code commentary explaining each analytical phase.
6. Return ONLY the code inside a single markdown code block.
Dataset format and Analysis Goal: [insert details about dataset columns and what you want to visualize]
Verified preset
755 char~189 tok
SQL Database Schema & FastAPI Boilerplate
Act as a Senior Database and Backend API Architect.
Task: Create a PostgreSQL schema design alongside a FastAPI CRUD boilerplate.
Database Rules:
1. Write standard, optimized DDL SQL statements (CREATE TABLE, foreign keys, cascade deletes).
2. Implement explicit indexes on search and foreign key columns for fast query times.
3. Include auto-updating `updated_at` timestamps using triggers.
API Rules:
1. Set up FastAPI routes using `SQLAlchemy` ORM.
2. Implement Pydantic data schemas for request input and response output models.
3. Add correct HTTP status codes (e.g. 201 Created, 204 No Content).
4. Return the SQL and Python code grouped in clear markdown blocks.
System Entities: [describe the entities, e.g., 'User accounts, each having multiple posts, posts having tags']
Verified preset
783 char~196 tok
Advertisement
Frequently Asked Questions
Q. Why does ChatGPT write placeholder comments instead of full code?
This happens when the prompt is too broad or the file scope is too large. Break down your instructions into smaller modules and specify 'Write the COMPLETE implementation. Do not skip any lines or use code placeholders.'
Q. Can I use these templates in local models like Llama or Mistral?
Yes, they are structured using standard system instructions and role-prompting which work across all models. For local models, you may need to increase the model's max token generation limit to avoid cut-off code.
