Best AI Prompts for Developers
These AI prompts for developers cover the work around the code as well as the code itself: planning a feature, designing the schema, tracking down a bug, reviewing a pull request and writing it all down. Each one asks for your stack, your code and your constraints first, so the answer fits your project instead of a generic tutorial. Most of them list what they assumed and mark gaps instead of inventing APIs, file paths or test results.
The collection has seven groups. Planning prompts turn ideas into specs, review system designs, design schemas and APIs, and record architecture decisions. Coding prompts write code from a spec, build accessible React components, write SQL queries and regular expressions, and include a patient mentor for beginners. Reading and refactoring prompts explain unfamiliar code, port it to a new language or framework and plan safe refactors. Debugging prompts find root causes, explain error messages and track down slow code. Testing and review prompts write unit tests, review code and accessibility, and check your own code for security holes. Agent prompts brief Claude Code, Cursor and other coding agents, set up a rules file for your repo and check what the agent wrote. Shipping prompts write READMEs, commit messages, CI pipelines and postmortems.
Most prompts work in ChatGPT, Claude and Gemini, and the rest are written for a coding agent that can read your repo. Fill in the blanks, paste it in and check the result before you ship it. Save your favorites to your Musebox library and remix them.
Planning and design
Feature Spec — From Idea to Buildable Ticket
Turn a feature idea into a spec a developer can build from: the problem, user stories, acceptance criteria, edge cases, data and API changes, and the open questions to settle first.
You're a staff engineer who has seen too many features built from a one-line ticket. Turn my feature idea into a spec a ...
System Design Review — Trade-offs Before You Build
Get a design reviewed before you build it: the weak points, failure modes, scaling limits and simpler alternatives, with each risk ranked and the questions a senior reviewer would ask.
You're a principal engineer running a design review. Read my proposed design, find the weak points before they're built,...
Database Schema Design — Tables, Keys and Migrations
Design a database schema from your requirements: tables, columns and types, keys and constraints, indexes for your real queries, and a migration plan that flags steps that could lock a table or lose data.
You're a database engineer who designs schemas that are still easy to change a year later. Design the schema for my feat...
API Design — Endpoints, Errors and Versioning
Design an API that other developers can use without asking you questions: endpoints, request and response shapes, status codes, error format, pagination, auth and a versioning plan.
You're an API designer who has maintained public APIs for years. Design an API for my feature that's consistent, predict...
Architecture Decision Record — Write Down the Why
Write an architecture decision record (ADR) from your notes: the context, the options you weighed, the decision and its consequences, so the next developer knows why and doesn't reopen it.
You're a tech lead who writes things down so the team doesn't argue about them twice. Turn my notes into an architecture...
Writing code
Senior Full-Stack Software Engineer, Coding Mentor & App Builder
This is a master “coding mentor” prompt that turns ChatGPT into a senior full-stack developer who can build any app with you from scratch or by editing your existing project. It forces ChatGPT to: Infer or choose a tech stack, Give you a clear plan, Output full, copy-paste-ready files, Walk you through everything step-by-step like a beginner, Briefly explain the concepts it uses so you actually learn as you go.
You are a senior full-stack software engineer and coding mentor helping me build and extend software applications. I wi...
Code From Spec — Small, Tested and Idiomatic
Write code from a spec in your stack and style: a short plan, the implementation, tests for the edge cases, and a list of every assumption made, instead of a wall of code you have to reverse-engineer.
You're a senior developer pairing with me. Write the code for this task the way an experienced engineer on my team would...
React Component — Accessible, Typed and Ready for Every State
Build a React component that's typed, accessible and handles loading, empty and error states, matched to your styling and existing components, with usage examples and tests.
You're a senior front-end engineer who ships components other developers enjoy reusing. Build the component I describe s...
SQL Query Writer — Correct, Readable and Indexed
Write a SQL query from a plain-English question and your schema, with an explanation of how it works, the edge cases it handles, and what would keep it fast at scale.
You're a database engineer who writes SQL other people can read. Write a query that answers my question correctly agains...
Regex Builder — Pattern, Explanation and Test Cases
Build a regular expression from examples of what should and shouldn't match. You get the pattern, a piece-by-piece explanation, test cases, and a warning when regex is the wrong tool.
You're a developer who writes regular expressions people can maintain. Build a pattern for my case and prove it works. ...
Reading and refactoring code
Code Explainer — What It Does and Why It's Written That Way
Paste code you didn't write and get it explained at your level: what it does, how data flows through it, the non-obvious parts, the assumptions it makes and the questions to ask its author.
You're a senior developer who's good at explaining other people's code. Walk me through this code so I understand what i...
Code Migration — Port Code to a New Language or Framework
Port code to another language, framework or major version without changing what it does: a mapping of old to new, idiomatic target code, the gaps with no direct equivalent, and tests that check both versions match.
You're a senior engineer who has led several migrations that shipped without regressions. Port my code to the target I n...
Refactor Plan — Safe Steps, Same Behavior
Plan a refactor you can ship safely: what's wrong with the code now, the target design, small steps that keep behavior unchanged, tests to add first, and how to roll each step back.
You're a senior engineer known for refactors that never break production. Plan how to improve this code in small, safe s...
Codebase Analysis & Refactoring Opportunities
This prompt walks through a full review of your Next.js / React codebase to spot practical refactoring opportunities, not theoretical ones. It checks structure, routing, state management, and patterns so you can see where things are getting too big, too complex, or duplicated. It looks for DRY violations, messy components, and missed Next.js and React best practices, then suggests concrete ways to split things up, clean them, or move logic into better places. You also get a prioritized action plan, so you know what to fix first for the biggest impact without breaking existing functionality.
# Codebase Analysis & Refactoring Opportunities ## Objective Perform a comprehensive analysis of this codebase to ident...
Debugging and performance
Bug Hunt — From Symptom to Root Cause
Track down a bug methodically: ranked hypotheses, the quickest test for each, the root cause once found, a minimal fix and a regression test so it stays fixed.
You're a senior engineer who debugs by evidence, not by guessing. Help me find the root cause of this bug and fix it pro...
Structured Python Debugging Assistant
A highly structured and effective Python debugging prompt template to help users isolate, understand, and resolve errors in their code using best practices
You are an expert Python debugging assistant. Help identify and resolve issues in the following code by reasoning throug...
Error Message Explainer — What Broke and Where to Look
Paste an error or stack trace and get it in plain language: what it means, the line that matters, the most likely causes in your code, and the first things to check.
You're a patient senior developer. Explain this error so I understand what broke, then tell me where to look first. Err...
Performance Investigation — Measure Before You Optimize
Find out why something is slow before you change it: a measurement plan, likely bottlenecks ranked, the profiling tools to use, and fixes ordered by impact and effort.
You're a performance engineer who measures before touching anything. Help me find out why this is slow and what's worth ...
Testing, review and security
Unit Tests — Edge Cases First
Write unit tests for existing code in your test framework: a test plan, tests for the happy path, edge cases and failures, and a list of behaviors that look like bugs.
You're a developer who writes the tests everyone else forgets. Write unit tests for my code that catch real bugs, not ju...
Code Review — Findings Ranked by Severity
Get a thorough code review of a diff or file: bugs, security issues, edge cases and readability problems, each ranked by severity with a suggested fix, plus what's good about the change.
You're a senior engineer doing a careful code review. Review my change the way a thoughtful teammate would: catch real p...
Security Review — Check Your Code for Common Holes
Review your own code for common security flaws: injection, broken auth and access control, exposed secrets, unsafe input handling and risky dependencies, with each finding rated and fixed.
You're an application security engineer reviewing code my team owns. Find the security weaknesses in it and show how to ...
WCAG Accessibility Audit & Compliance Analysis
Comprehensive accessibility audit prompt for Next.js/React codebases. Scans for WCAG 2.2 Level AA compliance issues including semantic HTML, keyboard navigation, ARIA usage, form accessibility, color contrast, dynamic content announcements, and responsive/motion considerations. Outputs prioritized findings with specific code fixes, affected WCAG criteria, and user impact assessments.
# Accessibility Audit & WCAG Compliance Analysis ## Objective Perform a comprehensive accessibility audit of this codeb...
Working with AI coding agents
Coding Agent Brief — A Task Claude Code or Cursor Can Finish
Turn a vague coding task into a brief an AI coding agent can finish without drifting: current state, desired state, files in scope, constraints, acceptance checks and when to stop and ask.
You're a tech lead who writes briefs for AI coding agents such as Claude Code, Cursor and Codex. Turn my rough task into...
Project Rules File — A CLAUDE.md or AGENTS.md for Your Repo
Write a rules file that gives AI coding agents the context they keep missing: stack, commands, structure, conventions and the things they must never do, kept short enough to actually be read.
You're a senior engineer setting up a repository for AI coding agents. Write the project rules file the agent reads at t...
Codebase Tour — Get Oriented in an Unfamiliar Repo
Get oriented in a codebase you didn't write: what it does, how it's organized, how a request flows through it, where to make common changes and the risky areas to avoid on day one.
You're a senior engineer onboarding me to a codebase I didn't write. Give me a guided tour so I can make my first change...
AI Code Check — Review Agent-Written Code Before You Merge
Review code an AI agent wrote before you merge it: does it match the request, what did it invent or change outside scope, which tests are real, and what to verify by hand.
You're a skeptical senior engineer reviewing code an AI coding agent produced. Your job is to catch what agents typicall...
Docs, Git and shipping
README Writer — Setup That Works on the First Try
Write a README a new developer can follow on the first try: what the project does, prerequisites, setup and run commands, configuration, common problems and how to contribute.
You're a developer advocate who tests every README on a clean machine. Write a README for my project that gets someone f...
Commit Message and PR Description — What Changed and Why
Turn a diff into a clear commit message and pull request description: what changed, why, how it was tested, the risks and what reviewers should look at first.
You're a senior engineer whose pull requests get reviewed fast because they're easy to read. Write the commit message an...
CI Pipeline — Fast Checks on Every Pull Request
Set up a continuous integration pipeline for your repo: jobs for install, lint, type check, tests and build, with caching, clear failure messages and secrets handled safely.
You're a DevOps engineer who builds pipelines developers don't hate. Write a CI pipeline for my repository that catches ...
Incident Postmortem — Blameless and Specific
Write a blameless postmortem from your incident notes: summary, impact, timeline, root cause and contributing factors, what went well, and action items with owners and dates.
You're a site reliability engineer who runs blameless postmortems. Turn my incident notes into a postmortem the team can...
Related collections
Best AI Prompts for Product Managers
These AI prompts for product managers cover the work from the first customer interview to the post-launch review: discovery, strategy, prioritization, specs, experiments and launches. Most ask for your own evidence first, such as interview notes, backlog estimates, baselines or results, so the plan, PRD or readout fits your product instead of reading like a template. The new prompts tell the AI to use only what you give it, mark its assumptions and show the math where there is some. Check the result with your team anyway. The collection has six groups. Discovery prompts plan customer interviews and usability tests, synthesize notes, map jobs to be done, size an opportunity and analyze customer segments. Strategy prompts draft a one-page strategy, team OKRs, a North Star metric and a competitive teardown. Roadmap prompts score a backlog with RICE, triage feature requests, build a now, next and later roadmap and help you say no. Delivery prompts write a PRD, user stories, a feature spec, an edge case review and a rollout plan. Experiment prompts plan an A/B test, write up the result and investigate a metric that moved. The last group plans a launch, writes release notes and a launch email, sends stakeholder updates and runs the post-launch review. Most prompts work in ChatGPT, Claude and Gemini. Fill in the blanks, paste it in and edit the result. AI can help you prepare and make sense of research, but it can't talk to your customers for you, so the discovery prompts start from real conversations and data. Keep confidential roadmaps and customer details out unless your company's AI policy allows them. Save your favorites to your Musebox library and remix them.
Best AI Prompts for Finance Professionals
These AI prompts for finance professionals cover the work of an FP&A, accounting or finance operations team: budgets and forecasts, variance analysis and reporting, the month-end close and audit, cash and working capital, business cases, and explaining the numbers to everyone else. Each new prompt asks for your own figures first, works from what you give it, shows its calculations and labels its assumptions and any illustrative numbers, so you can check the work the way you'd check a junior analyst's. The collection has six groups. Planning prompts build a driver-based budget, update the rolling forecast, model scenarios, cost a hiring plan and forecast 13 weeks of cash. Reporting prompts explain variances, write the monthly management report and the board deck finance section, define KPIs and calculate SaaS metrics. Close prompts build a close checklist, work through reconciliations, document journal entries and track audit requests. Cash prompts review AR aging, chase late invoices, check a payment run for duplicates and fraud, and measure working capital. Decision prompts review capital requests, model a price change, find cost savings, compare build against buy and compare supplier quotes. The last group explains numbers to non-finance colleagues, reviews spreadsheet models, documents internal controls and builds Excel formulas. Most prompts work in ChatGPT, Claude and Gemini. Fill in the blanks, paste it in and check the result against your source data before anyone relies on it. They're written not to give investment, tax or accounting advice or to state what GAAP, IFRS or tax rules require; confirm those with your controller, auditors or advisers. Use only AI tools your company has approved for financial data. At a public company, unreleased results usually count as material non-public information, so they go only into tools approved for them. Save your favorites to your Musebox library and remix them.
Best AI Prompts for Marketers
These AI prompts cover the marketing work that eats your week: planning content, writing emails and ads, doing SEO and figuring out who you're selling to. Each one asks for your product, audience and goal up front, so the output is specific enough to publish or present after a light edit. The collection is organized into four groups. Content and social prompts plan a week of posts or a 30-day calendar and turn one blog post into a set of social posts. Email, ad and landing page prompts write welcome sequences, cart recovery flows, launch and win-back emails, search and social ads, and landing page copy. SEO prompts handle blog outlines, content briefs, title tags and meta descriptions, product descriptions, internal links, alt text, schema markup and keyword ideas. Strategy and analytics prompts build customer personas, audience segments, brand voice guides, competitor messaging teardowns, demand tests and campaign reports. Every prompt works with ChatGPT, Claude and Gemini. Copy one, fill in your details and paste it into the assistant you use. Save the ones you rely on to your Musebox library and remix them to fit your brand.