readme-wizard
Skill46 starsChanged 8 months agoMITView on GitHub
Generate a polished, professional README.md for any project — from scratch or by improving an existing one. Use this skill whenever the user mentions README, wants to improve their repo's first impression, asks about shields.io badges, wants a star history chart, contributor avatars, documentation tables, project structure trees, or mermaid architecture diagrams in their README. Also use it when someone wants their GitHub repo to "look more professional", "look legit", or make a better first impression — even if they never say the word "README". Covers badge generation, README rewrites for new versions, fork customization, and adapting README structure to different project types (libraries, apps, docs repos, monorepos). Does NOT cover: CONTRIBUTING.md, CHANGELOG, API documentation, CI/CD setup, GitHub profile READMEs, or generating docs for individual components.
In this repository2 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- good-morningSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 3 are distinctive, and 7 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
7
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 15
Run scripts/scan_project.sh <project-directory> to collect structured JSON metadata:
- Very specific: line 30
Read references/readme-best-practices.md before writing.
- Very specific: line 34
Use assets/readme-template.md as the base structure.
- Very specific: line 36
When rendering the project tree, keep it close to how people browse repos: list directories before files, alphabetize within each group, …
- Distinctive: line 42
Small utilities/scripts: Keep it short — hero + what it does + usage is enough
- Very specific: line 46
Read assets/badges.json for the full badge catalog organized by category (status, social, extras) with {{PLACEHOLDER}} markers for dynamic …
- Distinctive: line 48
Golden rule: Only include badges for things that actually exist.
- Very specific: line 53
No placeholder text remains ({{PROJECT_NAME}}, TODO, etc.)
- Distinctive: line 54
All badges point to real shields.io URLs for things that actually exist
- Very specific: line 61
Only read assets/diagrams.md if the project has multiple components or a clear data flow (APIs, monorepos, content pipelines).
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
README Wizard
Generate or improve a project's README by scanning the repo and producing a polished, professional result based on real project data.
Workflow
1. Scan the project
Run scripts/scan_project.sh <project-directory> to collect structured JSON metadata:
- Project name and description
- License type
- Git remote (owner and repo name)
- Package manager (npm, yarn, pnpm, pip, cargo, go, etc.)
- CI setup (provider and workflow files)
- Social links (YouTube, Discord, Twitter, LinkedIn, Bluesky, Twitch)
- Directory structure (top 2 levels)
The reference script checks local files first, then uses the GitHub API to look up the homepage URL and crawls that homepage for additional social links. Your own first version can stay local-file only and add this enrichment later if you want it.
Handling missing data: The scan will return empty strings for anything it can't find. Never fabricate metadata — if a field is empty, skip the related section or badge entirely. For example: no CI workflows means no build badge; no social links means no social section.
2. Read the best practices guide
Read references/readme-best-practices.md before writing. It covers structure, tone, project-type adaptation, and common pitfalls. This context makes a real difference in output quality — the guide explains the reasoning behind each section so you produce READMEs that feel intentional rather than formulaic.
3. Build the README
Use assets/readme-template.md as the base structure. Replace {{PLACEHOLDER}} markers with actual project data from the scan.
When rendering the project tree, keep it close to how people browse repos: list directories before files, alphabetize within each group, and add trailing / markers for directories.
Adapt — don't copy blindly. The template is a starting point. Drop sections that don't apply (e.g., no Contributing section for a personal notes repo) and adjust the tone to match the project:
- Libraries/frameworks: Focus on installation, API, and examples
- Applications: Focus on setup, configuration, and screenshots
- Docs/learning repos: Focus on structure and navigation
- Small utilities/scripts: Keep it short — hero + what it does + usage is enough
4. Add badges
Read assets/badges.json for the full badge catalog organized by category (status, social, extras) with {{PLACEHOLDER}} markers for dynamic values. Use style=for-the-badge for all badges.
Golden rule: Only include badges for things that actually exist. A badge for a nonexistent CI workflow or a guessed npm package is worse than no badge at all.
5. Validate the output
Review the generated README against the assertions in evals/evals.json. Verify:
- No placeholder text remains (
{{PROJECT_NAME}},TODO, etc.) - All badges point to real shields.io URLs for things that actually exist
- Sections are proportional to the project — a simple script doesn't need a 200-line README
- Social badges only appear if social links were found in the scan
- The tone is concise and direct, not marketing fluff
6. Optionally add a diagram
Only read assets/diagrams.md if the project has multiple components or a clear data flow (APIs, monorepos, content pipelines). Skip entirely for simple projects. Generate the diagram from the project's actual structure — the templates in the file are starting points to adapt, not copy.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

