git-repo-craft
Skill0 starsChanged 2 months agoMITView on GitHub
Automates GitHub repository creation, configuration, and metadata auditing. Mandates minimal yet meaningful repository descriptions, curated and standardized GitHub topics, clean README and license initialization, and active verification via the GitHub CLI (gh).
In this repository20 other files it gives its agents.
- AGENTS.mdAGENTS.md
- rules/AGENTS.mdAGENTS.md
- kurokoSkill
- tayoriSkill
- xerjSkill
- aesthetic-readme-craftSkill
- cli-builderSkill
- codebase-digestSkill
- git-release-craftSkill
- karakuriSkill
- python-clean-codeSkill
- rust-clean-codeSkill
- rust-style-pythonSkill
- systematic-code-verificationSkill
- compose-clean-codeSkill
- bash-clean-codeSkill
- tui-designSkill
- frontend-designSkill
- web-design-guidelinesSkill
- unslopSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. About 0 are common, 10 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈10
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 19
Minimal: No buzzword soup, no filler ("This is a repo that contains..."), no redundant repository name repetition.
- Very specific: line 29
No Redundancy: Never start with the repo name (e.g., karakuri — A tool for...).
- Distinctive: line 30
No Filler: Never use phrases like:
- Distinctive: line 51
Every repository must have 4 to 8 topics categorized across four dimensions:
- Very specific: line 64
Standard Abbreviations: Use recognized industry tags (cli, api, sdk, tui).
- Distinctive: line 70
When tasked with inspecting, cleaning, or updating an existing repository:
- Distinctive: line 78
If missing or empty: Analyze repository files (README, package manifests) to craft a minimal, punchy one-liner.
- Distinctive: line 79
If bloated or redundant: Strip filler words and repository name prefix.
- Distinctive: line 86
Check if topics are missing or suboptimal.
- Distinctive: line 107
When creating a new repository from scratch or an existing local directory:
- Very specific: line 110
Ensure .gitignore exists and matches the project language stack.
- Very specific: line 111
Ensure an open-source LICENSE (e.g., MIT, Apache-2.0) is present.
- Distinctive: line 112
Ensure a clean README.md with project title and brief overview exists.
- Distinctive: line 113
Verify initial commit is cleanly staged and committed.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
git-repo-craft Skill: Repository Creation, Editing & Metadata Crafting
Standardizes GitHub repository creation, editing, and metadata auditing. Guarantees that every repository has a minimal yet meaningful description, curated discoverable topics, clean initialization, and verified configuration.
1. The Core Philosophy: Minimal Yet Meaningful
A repository's GitHub metadata is its front door. Poor metadata kills discoverability and confuses visitors.
- Minimal: No buzzword soup, no filler ("This is a repo that contains..."), no redundant repository name repetition. Keep it under 80 characters or one crisp, punchy sentence.
- Meaningful: Immediately conveys WHAT the project is and WHAT PROBLEM it solves.
- Discoverable: 4–8 curated, kebab-case topics covering domain, language, purpose, and ecosystem.
2. Description Crafting Standards
Rules for Repository Descriptions:
- Maximum Length: Aim for 50–90 characters (hard limit: 120 characters).
- No Redundancy: Never start with the repo name (e.g.,
karakuri — A tool for...). The repo name is already visible above the description. - No Filler: Never use phrases like:
- ❌ "A repository for..."
- ❌ "This project is intended to..."
- ❌ "Collection of useful scripts and tools..."
- Action & Value Pattern:
[Adjective] [Core Entity] for [Primary Target / Use Case][Primary Action Verb] [Object] with [Key Differentiator]
Good vs. Bad Examples:
| Status | Description | Rationale |
| ❌ Bad | my-repo - A repository that contains various miscellaneous scripts and tools | Redundant name, filler phrasing, vague |
| ❌ Bad | A collection of python scripts for my daily tasks and work | Generic, zero meaning, filler phrasing |
| ✅ Good | Modular skills and automated workflows for AI coding agents | Minimal, precise domain, immediate value |
| ✅ Good | Fast Rust compiler offloader to Kaggle Cloud with zero local CPU load | Clear value proposition, concise |
| ✅ Good | Production standards, linting, and quality guardrails for Python | Punchy, informative, clean |
3. Curated Topics Matrix
Every repository must have 4 to 8 topics categorized across four dimensions:
| Dimension | Purpose | Examples |
| Domain | Primary field or area | ai-agents, cloud-computing, systems-programming, web-development |
| Language / Stack | Primary programming language | rust, python, bash, typescript, flutter |
| Purpose / Function | What the tool actually does | agent-skills, workflow-automation, developer-tools, cli, linter |
| Ecosystem | Surrounding framework / tool | antigravity, github-actions, docker, kaggle |
Topic Formatting Rules:
- Strictly Kebab-Case:
developer-tools, neverdeveloper_toolsorDeveloperTools. - All Lowercase:
ai-agents, neverAI-Agents. - No Punctuation or Spaces: Letters, numbers, and hyphens only.
- Standard Abbreviations: Use recognized industry tags (
cli,api,sdk,tui).
4. Repository Audit & Editing Workflow
When tasked with inspecting, cleaning, or updating an existing repository:
Step 1: Inspect Current Metadata
gh repo view <owner>/<repo> --json description,repositoryTopics,homepage,visibility,defaultBranchRefStep 2: Audit Description
- If missing or empty: Analyze repository files (README, package manifests) to craft a minimal, punchy one-liner.
- If bloated or redundant: Strip filler words and repository name prefix.
- Update description:
gh repo edit <owner>/<repo> -d "<Minimal yet meaningful description>"
Step 3: Audit & Apply Topics
- Check if topics are missing or suboptimal.
- Select 4–8 relevant topics from the codebase inspection.
- Apply topics:
(To remove obsolete topics: usegh repo edit <owner>/<repo> \ --add-topic topic-one \ --add-topic topic-two \ --add-topic topic-three \ --add-topic topic-four--remove-topic <old-topic>)
Step 4: Verify Updates
gh repo view <owner>/<repo> --json description,repositoryTopics5. Repository Creation Workflow (gh repo create)
When creating a new repository from scratch or an existing local directory:
Step 1: Pre-Flight Checklist
- Ensure
.gitignoreexists and matches the project language stack. - Ensure an open-source
LICENSE(e.g., MIT, Apache-2.0) is present. - Ensure a clean
README.mdwith project title and brief overview exists. - Verify initial commit is cleanly staged and committed.
Step 2: Initialize & Push to GitHub
# In the project directory:
gh repo create <repo-name> \
--public \
--description "<Minimal yet meaningful description>" \
--source=. \
--remote=origin \
--push(Use --private if internal or confidential).
Step 3: Immediately Set Curated Topics
gh repo edit \
--add-topic <domain-topic> \
--add-topic <language-topic> \
--add-topic <function-topic> \
--add-topic <ecosystem-topic>Step 4: Verify Final State
gh repo view --json description,repositoryTopics,urlConfirm the output displays the clean description and all topics without error.
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.

