# Repo Init

> >

- **Type:** Skill
- **Install:** `agentstack add skill-aberson-claude-skills-repo-init`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [aberson](https://agentstack.voostack.com/s/aberson)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [aberson](https://github.com/aberson)
- **Source:** https://github.com/aberson/claude-skills/tree/main/repo-init

## Install

```sh
agentstack add skill-aberson-claude-skills-repo-init
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Repo Init

## Overview

Use this skill to take a bare local project directory all the way to a fully scaffolded GitHub repo with a README and issues.

**Core principle:** Read the plan first -> git init -> push -> README -> issues, in that order.

## Steps

### 1. Confirm inputs with the user

Confirm these inputs before any git or gh command runs:
- Repo visibility: **public or private?**
- Repo name (suggest the directory name as default)
- Target account/org (default: the authenticated `gh` user)

### 2. Check prerequisites

```bash
gh auth status          # must show active account with repo scope
ls -la     # confirm .git does NOT exist
```

If `.git` already exists, skip step 4 (git init + first commit) and jump to step 5
(create the GitHub repo and push). Do not re-run `git init` on an existing repo.

### 3. Read the plan document

Before writing any git commits or the README, read `plan.md` (or equivalent) in
the project directory. This is required — the README and issues are derived from it.

### 4. Initialize git and make the first commit

```bash
cd 
git init
git add .
git commit -m "Initial commit:  project scaffold

Co-Authored-By: Claude Sonnet 4.6 "
```

### 5. Create the GitHub repo and push

```bash
gh repo create  --private --source=. --remote=origin --push
# or --public if requested
```

### 6. Write the README

Generate `README.md` from the plan. A good README includes:
- One-paragraph description of what the project does
- Stack table (tool + why)
- Prerequisites
- Setup steps (numbered, runnable)
- Key design decisions (brief)
- Data store or project structure diagram (if relevant)

Do **not** duplicate the full plan — summarize and surface what a new contributor needs to get started.

Commit and push:
```bash
git add README.md
git commit -m "Add README

Co-Authored-By: Claude Sonnet 4.6 "
git push
```

### 7. Create GitHub issues from the build steps

Read the build order section of the plan. Create one issue per step using `gh issue create`.

The plan typically labels build steps under a "Build Order" or "Steps" heading;
if no such section exists, ask the user how to slice the work before any issues are filed.

**Issue body structure for each:**
- `## Build step N of M`
- `### Files` — list of files to produce
- `### What to build` — concrete description drawn from the plan
- `### Quality gates` — test/lint/typecheck commands
- `### Notes` — key constraints or gotchas from the plan

Create issues **one at a time** with separate `gh issue create` calls (not batched in a single shell command) to avoid quoting failures.

```bash
gh issue create --title "Step N: " --body ""
```

## Constraints

- Always confirm visibility, name, and account with the user before any `git` or `gh` command runs.
- Always read `plan.md` (or equivalent) before writing the README or creating issues; both are derived from it.
- Always call `gh issue create` once per issue; never stack multiple heredocs in one shell command.
- Always include a `Co-Authored-By: Claude` trailer in the scaffold commit and the README commit.
- Skip `git init` when `.git` already exists in the project directory.

## Limitations

- Handles initial repo creation only; updating issues, labels, or milestones later needs separate tooling.
- Requires the `gh` CLI authenticated with `repo` scope; org repos under restricted-policy may still reject the `gh repo create` call.
- Does not configure branch protection, default reviewers, labels, milestones, or GitHub Projects.
- Treats `plan.md` as the single source for README and issues; multi-plan or wiki-style sources are out of scope.

## Common Mistakes

**Batching issue creates in one shell command**
- Problem: Multi-heredoc quoting fails with EOF errors
- Fix: One `gh issue create` call per issue, separate Bash tool calls

**Writing the README before reading the plan**
- Problem: Generic README that misses key design decisions
- Fix: Always read plan.md first

**Skipping the inputs check**
- Problem: Creating a public repo when the user wanted private (or wrong account)
- Fix: Confirm visibility, name, and account before any git or gh commands

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [aberson](https://github.com/aberson)
- **Source:** [aberson/claude-skills](https://github.com/aberson/claude-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-aberson-claude-skills-repo-init
- Seller: https://agentstack.voostack.com/s/aberson
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.