Tooling Agent Skill Multi Repository Name Conventions — Git skill for Claude Code

Repository Name Conventions

A comprehensive Git repository naming convention, packaged as a skill that Claude Code, Cursor, GitHub Copilot, Gemini CLI, and any other `AGENTS.md`-reading tool (Codex, Grok, etc.) can all load — plus scripts to validate and generate compliant names, and hooks that enforce the convention at repo-creation time.

What this is

Most naming conventions live in someone's head, a wiki page nobody reads, or a Slack thread from two years ago. This repo turns one into something executable: a machine-checkable pattern, a script that tells you exactly why a candidate name fails, and hooks that stop a bad name before the repo even exists — while staying readable enough that a human can also just read `reference/conventions.md` and follow it by hand.

Why repository naming conventions matter

• Discoverability. acme-fin-payments-service-go-microservice-go_ledger-tier1 tells you the owning domain, the criticality, the language, and the shape of the thing before you've opened it. A hundred repos named after inside jokes do not.
• Automation. CODEOWNERS routing, branch-protection tiers, security scanning policy, and cost allocation can all be driven mechanically off a structured name (every *-tier1 repo needs two reviewers) instead of a spreadsheet someone forgets to update.
• Onboarding cost. A new engineer (or a new AI agent) can guess what a repo does, and what convention a new one should follow, without reading a README first.
• Grep-ability at scale. Once you have dozens or hundreds of repos — which happens fast with AI-agent tooling (skills, MCP servers, hooks, commands are all small, numerous, and fast-moving) — an inconsistent namespace stops being a style nitpick and starts being an actual cost.

The pattern, at a glance

Core (always required):

---_[.]

Full (adds optional enterprise/freelance/lifecycle segments):