repo-dive

name	repo-dive
description	Clone a GitHub repository to a temporary directory for efficient local code exploration. ALWAYS prefer this over WebFetch or `gh api` for reading repository code — local clone is faster and more complete. Use when a GitHub repository URL appears and the task involves code access: investigation, analysis, comparison, reference, learning, or review. Triggers: GitHub repo URL + any code exploration intent ("調査して", "コードを見て", "investigate", "analyze", "explore", "compare", "refer to").
argument-hint	<repo-url> [branch-or-pr]

Repo Dive

Clone a GitHub repository locally for efficient code exploration using Claude's native file tools.

Quick Start

/repo-dive https://github.com/owner/repo
/repo-dive https://github.com/owner/repo feature-branch
/repo-dive https://github.com/owner/repo/pull/123

Argument Parsing

Parse $ARGUMENTS to extract:

• Repo URL: https://github.com/owner/repo or owner/repo
• Optional target (second argument or embedded in URL):
  • Branch name: main, feature/xyz
  • PR URL: https://github.com/owner/repo/pull/123
  • PR shorthand: #123

If $ARGUMENTS is empty, ask the user for the repository URL.

Workflow

Step 1: Clone

Clone to /tmp/repo-dive/<owner>-<repo>:

# Default (shallow clone of default branch)
git clone --depth 1 https://github.com/owner/repo /tmp/repo-dive/owner-repo

# Specific branch
git clone --depth 1 --branch <branch> https://github.com/owner/repo /tmp/repo-dive/owner-repo

# PR — clone then checkout PR ref
git clone --depth 1 https://github.com/owner/repo /tmp/repo-dive/owner-repo
cd /tmp/repo-dive/owner-repo && gh pr checkout <number>

If the directory already exists, ask the user:

• Reuse the existing clone (skip clone step)
• Remove and re-clone (rm -rf then clone fresh)

Step 2: Explore

Use Claude's native file tools on the cloned directory:

• Glob for file discovery (/tmp/repo-dive/owner-repo/**/*.ts)
• Read for file contents
• Grep for content search
• Agent (subagent_type=Explore) for deep codebase exploration

For large-scale analysis, use the gemini-research skill targeting the cloned directory.

Do NOT use gh api for file contents — everything is local now.

Step 3: Report

Summarize findings to the user. Structure depends on the request but typically includes:

• Repository overview (language, framework, structure)
• Key findings relevant to the user's question
• Notable patterns, potential issues, or recommendations

Cleanup

Do NOT automatically delete the cloned repository after exploration. The user may want to continue investigating in follow-up messages. The /tmp directory is cleaned up on system restart.

If the user explicitly asks to clean up:

rm -rf /tmp/repo-dive/owner-repo

Notes

• For private repositories, git clone uses the gh CLI's auth context (credential helper). If clone fails, suggest gh auth status to verify authentication.
• Shallow clone (--depth 1) is the default. If the user needs git history (blame, log), re-clone without --depth 1.
• When exploring monorepos or very large repos, ask the user which subdirectory to focus on before diving in.