One small step at a time
Model notes: Reduces runaway edits in agent mode
Fill in the variables (optional — they're inserted live below):
Implement only step {{step_number}} of the plan we agreed: "{{step_description}}".
Rules:
- Touch the fewest files possible. List the files you will change before changing them.
- Don't refactor, rename or reformat anything unrelated.
- After the change, tell me exactly how to run/verify it (command + expected result).
- If something in the plan looks wrong, stop and say so instead of improvising.
How to use it
- Fill the variables above, or replace each
{{variable}}after pasting. - Paste into your AI tool (ChatGPT, Claude, Gemini, Cursor, Claude Code, or a local model in Ollama/LM Studio).
- Review the output. Models make mistakes, especially with facts, numbers and code.
Get all 209 prompts + agent workflows
Every prompt as Markdown and JSON, 10 multi-step agent workflows, and 12 ready-made slash commands for Cursor and Claude Code. Instant zip download, one-time purchase.
More Vibe coding prompts
Vibe coding
Spec-first app kickoff
Any frontier model; works in Cursor, Claude Code, Windsurf
Variables: ideausersstack
You are a senior engineer starting a new project with me. Before writing any code, produce a short build spec.
Project idea: {{idea}}
Target users: {{users}}
Preferred stack (or "you choose"): {{stack}}
Write:
1. One-paragraph product summary.
2. MVP feature list (max 7), each with a one-line acceptance test.
3. Explicit out-of-scope list.
4. Data model (entities + key fields).
5. File/folder structure.
6. Step-by-step build plan in small commits, each independently runnable.
7. Open questions you need me to answer.
Do not write code yet. Wait for my "go". You are a senior engineer starting a new project with me. Before writing any code, produce a short build spec.
Project idea: {{idea}}
Target users: {{users}}
Preferred stack (or "you choose"): {{stack}}
Write:
1. One-paragraph product summary.
2. MVP feature list (max 7), each with a one-line acceptance test.
3. Explicit out-of-scope list.
4. Data model (entities + key fields).
5. File/folder structure.
6. Step-by-step build plan in small commits, each independently runnable.
7. Open questions you need me to answer.
Do not write code yet. Wait for my "go".
Open & fill variables → Vibe coding
Turn a screenshot into a component
Needs a vision-capable model
Variables: frameworkstyling
Recreate the UI in the attached screenshot as a {{framework}} component styled with {{styling}}.
- Match layout, spacing, hierarchy and colours as closely as possible; approximate fonts with system fonts.
- Use semantic HTML and accessible labels.
- Make it responsive: describe what changes below 640px.
- Use placeholder data in a typed props object, not hard-coded strings.
- Return a single file, then list anything you couldn't infer from the image. Recreate the UI in the attached screenshot as a {{framework}} component styled with {{styling}}.
- Match layout, spacing, hierarchy and colours as closely as possible; approximate fonts with system fonts.
- Use semantic HTML and accessible labels.
- Make it responsive: describe what changes below 640px.
- Use placeholder data in a typed props object, not hard-coded strings.
- Return a single file, then list anything you couldn't infer from the image.
Open & fill variables → Vibe coding
Add a feature without breaking things
Best with full-repo context (Cursor @codebase, Claude Code)
Variables: feature
I want to add: {{feature}}.
First, explore the codebase and tell me:
1. Where this naturally belongs and which existing patterns/utilities to reuse.
2. Every file you expect to modify or create.
3. Risks: what existing behaviour could break, and how we'll check it.
Then wait for approval. After approval, implement it with tests that follow the project's existing test style. I want to add: {{feature}}.
First, explore the codebase and tell me:
1. Where this naturally belongs and which existing patterns/utilities to reuse.
2. Every file you expect to modify or create.
3. Risks: what existing behaviour could break, and how we'll check it.
Then wait for approval. After approval, implement it with tests that follow the project's existing test style.
Open & fill variables → New prompts every week
One short email with the best new prompts and agent workflows we've tested. Free; unsubscribe any time.