Generate and Refine Code Prompts

A curated library of battle-tested prompt templates, inspired by the community awesome-chatgpt-prompts collection. Role-based prompts include an Expert Developer prompt for code review (identify bugs, suggest performance improvements, recommend better patterns), a Code Reviewer prompt formatted into red/yellow/green severity tiers, a Technical Writer prompt for turning technical concepts into accessible documentation, and a System Architect prompt weighing scalability, reliability, maintainability, performance, and cost with explicit trade-off analysis.

Task-specific prompts include Debug This Code (problem identification, root cause, fix, and prevention), Explain Like I'm 5, Code Refactoring (readability first, remove duplication, single responsibility, meaningful names), Write Tests (happy-path scenarios, edge cases, error conditions, boundary values, arrange-act-assert structure), and API Documentation generation formatted as OpenAPI/Swagger or Markdown. Analysis prompts cover Code Complexity Analysis (cyclomatic complexity, coupling, cohesion, dependencies, technical debt), Performance Analysis (time and space complexity, I/O bottlenecks, algorithmic issues, quick wins), and a Security Review across input validation, authentication and authorization, data protection, and injection vulnerabilities, classified by severity. Creative prompts cover Brainstorm Features (ranked by impact-to-effort ratio) and a Name Generator across descriptive, evocative, acronym, and metaphorical categories. Transformation prompts cover migrating code between languages and converting between data formats, both requiring exact functionality or data preservation and edge-case handling.

It closes with four named prompt-engineering techniques - Chain of Thought, Few-Shot Learning, the Persona Pattern, and Structured Output for JSON-schema responses - and a six-point prompt-improvement checklist covering clear objective, context, format, examples, constraints, and success criteria.

Use it when a ready-to-use prompt template is needed - a role-based persona, a task-specific pattern (debugging, refactoring, test-writing, documentation), an analysis prompt, or a prompt-engineering technique - rather than writing a prompt from scratch.

Input: a task or persona needing a prompt template, with placeholders like [CONCEPT], [PRODUCT], [SOURCE]/[TARGET] to fill in. Output: a ready-to-use prompt template, filled in with the specifics of the current task. A Few-Shot Learning template, for instance, structures the fill-in as numbered input/output example pairs followed by the actual input to complete, rather than a single instruction.

Built on and credits the community awesome-chatgpt-prompts GitHub repository, with prompts.chat and Learn Prompting listed as further-reading resources for prompt engineering.

Anyone who wants a ready-made prompt for a common task - code review, debugging, documentation, brainstorming - or a reusable pattern like Chain of Thought, Few-Shot, or Persona to structure a new one.