Free tool · Developer
AI developer prompt generator
Turn a short development request into a structured brief for coding agents such as Claude Code, Codex or Cursor: goal, requirements, stack conventions, security, testing, SEO, deliverables and acceptance criteria.
Developer Prompt Generator
Prompt
Structured developer prompt
Updates as you type. Nothing leaves your browser unless you use Refine with AI.
Refine with AI sends the fields above to OpenAI to make the brief more specific to your stack. Do not include passwords, keys or confidential client data.
What it creates
The brief a senior developer would write
Coding agents do what the prompt says and fill gaps with assumptions. “Add a comparison page” produces something; a brief that states the stack, the constraints, the security expectations and how the result will be checked produces something you can merge. The generator builds that brief as you type, with conventions for the technology you choose.
PROJECT and GOAL
The type of project, technology and versions, and the task in your words.
CURRENT STATE
What exists today. If you leave it empty, the agent is told to inspect the repository and summarize the relevant structure before changing anything.
REQUIREMENTS and IMPLEMENTATION
Your requirements and restrictions as lists, plus stack conventions: WordPress hooks and plugins instead of core edits, Magento modules with plugins and di.xml, Laravel Form Requests and policies, React components with TypeScript, and more.
SECURITY
Input validation, output escaping, authorization, nonces or CSRF tokens, prepared statements and secret handling, specific to the stack.
TESTING and SEO
Test types and cases for the stack, and SEO requirements (URLs, redirects, metadata, server-rendered content) only when the task affects public pages.
OUTPUT and ACCEPTANCE CRITERIA
What the agent must deliver (summary, changed files, migrations, test results) and testable criteria derived from your requirements.
Why it matters
Most agent mistakes start as missing context
AI coding agents are fast, but they do not know which files must not be touched, which framework conventions your team follows, or what “done” means for this ticket. Without that, they refactor unrelated code, edit core files, skip permission checks or declare success without running tests.
A structured prompt reduces rework. Asking for a plan first catches misunderstandings before code is written; explicit restrictions protect upgrade paths; acceptance criteria give both the agent and the reviewer the same checklist. It also documents the intent of the change for whoever reviews the pull request.
This is how HUB works with coding agents on client projects: senior engineers write the specification and review every change, and the agents do the typing.
How to use it
Brief, refine, review
- 01Describe the task and constraints
Pick the project type and stack, write the task in one or two sentences, and list requirements and restrictions one per line.
- 02Refine with AI if the task is complex
Refine with AI makes the brief more specific and lists open questions. Answer those before you start.
- 03Copy it into your agent
Paste the prompt into Claude Code, Codex, Cursor, Copilot or any other assistant, or save it with the ticket as a Markdown file.
- 04Review the result against the criteria
Check the plan, read the diff and run the tests. The acceptance criteria are your review checklist.
Questions
About the prompt
generator
Built an application with AI tools already? Score its production readiness with the AI Code Health Check.
Which coding agents does it work with?
Any that accept a text instruction: Claude Code, OpenAI Codex, Cursor, GitHub Copilot, Windsurf and others. The format is plain text with clear section headings.
Is my task description sent anywhere?
No, the basic prompt is built in your browser. Refine with AI sends the form fields to OpenAI through HUB’s server, without storing or caching them. Do not include passwords, keys or confidential client data.
Why does the prompt ask for a plan before coding?
A short plan lets you catch a wrong approach before the agent changes dozens of files. Untick the option for small, well-defined tasks.
Can HUB run the agents for us?
Yes. HUB’s AI Coding & Development service combines coding agents with senior engineers who own the specification, review and release.
Need help with what
the tools found?
HUB engineers implement technical SEO, performance, security and AI-readiness fixes on WordPress, Magento, custom PHP and modern JavaScript stacks.