README Writer
Writes complete, professional README files with setup instructions, usage examples, and badges.
What it does
The README Writer prompt generates a complete README.md using a 13-section template that matches how modern open-source maintainers actually structure their docs. It leads with the problem solved — not the tech stack — puts a runnable quick start in the first screen, and covers everything a contributor needs: features, usage, configuration, API reference, architecture, contributing, roadmap, acknowledgements, license.
What the prompt generates
You paste in a short description of your project (name, what it does, stack, install command, one example) and Claude returns a full README.md ready to commit: H1 with one-line tagline, status badge placeholders in shields.io format, a 3–5 sentence 'what it does' paragraph, a 60-second quick start, 5–8 feature bullets starting with verbs, two or three realistic usage examples (not toy), a configuration table, an API reference or link, an optional architecture diagram, a contributing section with exact setup commands, a 3–6 item roadmap, acknowledgements, and the license block.
Who should use it
Open source maintainers who want consistent, scannable READMEs across projects. Solo developers launching side projects who want a professional first impression. Teams publishing internal libraries that new hires read on day one. CLI authors, SDK authors, library authors — any project where the README is the first thing a new user sees.
What it solves
Most README templates are too short (a paragraph and an install command) or too long (a wiki masquerading as a single file). This prompt hits the right level of detail for a repo README — enough for a new contributor to get productive in 60 seconds, not so much that the file becomes a wiki that nobody maintains.
How to install
Which tool are you using?
Not sure? Claude.ai is the website. Claude Code is the command-line tool you install separately. Cursor is a code editor that reads .cursorrules.
- 01
Open Claude.ai
Go to claude.ai and click 'New conversation'.
- 02
Copy the prompt
Click the Copy button on the prompt file below. The full 13-section README template is now on your clipboard.
- 03
Paste into the chat input
Paste with Cmd+V (Mac) or Ctrl+V (Windows). Do not send yet.
- 04
Replace the placeholder with your project info
Replace [PASTE YOUR PROJECT INFO HERE] with a short description: name, what it does, tech stack, install command, one example usage. 5–10 lines is enough.
Include anything specific: team conventions, license, audience (open source vs internal).
- 05
Send and review the result
Claude produces a README with all 13 sections. Review for accuracy — it will sometimes infer details. Ask follow-ups like 'make section 4 shorter' or 'add a monorepo note' rather than regenerating from scratch.
Save the final README.md to your repo root and commit. Pair with the Git Commit Messages skill for release notes later.
The claude prompt file
Copy the full contents below, or download the file directly.
You are a technical writer who specializes in open-source READMEs that actually get read end-to-end. You write for busy developers who will bounce if the quick start is not in the first screen of content. Generate a README.md for my project. Use these 13 sections in this exact order. Every section is required unless I explicitly say to omit it. ## 1. Project name + one-line tagline- Name as an H1- Tagline: one sentence that states the problem solved, not the tech used- No marketing language ## 2. Status badges- Build status, version, license at minimum- Test coverage, npm downloads, or stars if relevant- Use shields.io format- Leave URLs as placeholders [[BUILD_URL]] if I have not given them ## 3. What it does- One paragraph (3–5 sentences)- Describes the problem and how the project solves it- No jargon; assume a senior developer who has not seen this project before- No tech-stack bragging ## 4. Quick start- Install command(s), copy-pasteable, exactly as typed- A minimal "hello world" that actually runs and produces visible output- The reader should be able to verify the install works in under 60 seconds ## 5. Features- Bullet list, 5 to 8 items- Each feature in one line, starts with a verb ("Generate…", "Validate…")- No duplicates, no "and more" ## 6. Usage- 2 to 3 realistic examples, not toy examples- Each example: a short intro sentence + a runnable code block + a short note on what the output looks like- Cover the main use case, a secondary use case, and an integration example ## 7. Configuration- Table of env vars with: name, default, description, required flag- Config file examples if applicable- Note precedence if multiple configuration sources exist ## 8. API reference- Full reference inline if small (< 20 public APIs)- Otherwise link to /docs/api.md or a hosted site- For libraries: signature + one-sentence purpose per function ## 9. Architecture / How it works- Optional but highly recommended for non-trivial projects- One diagram or block list showing the major pieces- 2 to 4 sentences describing data flow ## 10. Contributing- How to set up a dev environment (exact commands)- How to run tests- PR rules (conventional commits, sign-off, formatting)- Link to CONTRIBUTING.md if it exists ## 11. Roadmap- 3 to 6 items of what is planned or explicitly not planned- Optional, but shows maturity ## 12. Acknowledgements- Credit to contributors, sponsors, libraries you stand on- Keep short and specific ## 13. License- License name + year + author- Link to LICENSE file # Hard Rules- Lead with the problem solved, not the tech stack- Every code block must be runnable exactly as written- No marketing fluff, no "blazing fast", no "powerful"- Include a screenshot or GIF placeholder () if the project has a UI- If a section truly does not apply, keep the heading and write one line explaining why it is omitted- Use sentence case for headings, not title case Project info:[PASTE YOUR PROJECT INFO HERE]Example output
What Claude does before and after you install this claude prompt.
Claude writes a 2000-word README with a screenshot on top, no quick start, and marketing language. Users bounce.
Claude writes a 400-word README with working install commands, a realistic usage example, and a clear configuration table. Users try it.
Customization tips
For libraries with multiple language bindings, extend the Usage section with one example per language. For CLIs, add a full --help output block. For SaaS open-source, add a 'Self-host vs hosted' section right after the quick start. For internal libraries, drop the Roadmap and Contributing sections.
Related resources
API Documentation Skill
Generates clean API docs from your code. OpenAPI-compatible output, examples included.
Git Commit Messages Skill
Generates conventional commit messages from your diff. Consistent, descriptive, team-ready.
CLAUDE.md Writer
Writes a complete CLAUDE.md file for your project based on your tech stack and team conventions.
Frequently asked questions
Does it add status badges automatically?
It suggests placeholder badges (build, version, license) in shields.io format. You fill in the actual URLs for your CI provider and npm / crates / pub registry.
Can it write READMEs for monorepos?
Yes. Generate a root README summarizing the repo and per-package READMEs separately. Pasting the monorepo overview + a specific package context gets the best results.
Will it hallucinate features I did not describe?
Not if you give it real project info. The prompt requires specific input — if you skip the info, Claude will ask clarifying questions rather than invent features.
Does it cover contributing guidelines?
Yes — section 10. Tell Claude what your actual workflow is (PR rules, sign-off requirements, commit format) and it will produce accurate contributing instructions.
Can it regenerate an existing README that is out of date?
Yes. Paste the old README along with your updated project info. Claude will keep what is still accurate and update what has changed.
Want more like this?
Browse the full RohanKit library — free resources for Claude and Cursor.