FrancyJGLisboa

Gitreverse — Development skill for Claude Code

Development community

Reverse engineer any repo into it's original prompt.

How to install Gitreverse

This entry records only its repository, not the path inside it, so there is no exact command to give. Open FrancyJGLisboa/gitreverse and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Gitreverse does

Reverse engineer any repo into it's original prompt.

Alternatives in Development

Root Cause Tracing — Use when errors occur deep in execution and you need to trace back to find the original trigger 102.5k ★
Master Routing — reverse-skill PRIMARY 快路径 29.6k ★
Generic Extract Skill — You are an Autonomous 3D Technical Artist Agent 14.1k ★

README

GitReverse

https://github.com/user-attachments/assets/f0cdb7b2-c6f0-4483-8a01-153170479f2e

Turn a **public GitHub repository** into a **single synthetic user prompt** that someone might paste into Cursor, Claude Code, Codex, etc. to vibe code the project from scratch.

The app pulls **repo metadata**, a **root file tree** (depth 1), and the **README**, then uses an LLM to produce one short, conversational prompt grounded in that context.

Paste a GitHub URL or `owner/repo` on the home page. You can also open **`/owner/repo`** (e.g. `/vercel/next.js`) for a shareable link that runs the same flow.

GitHub-style **`/owner/repo/tree/...`** URLs on this site **redirect to `/owner/repo`** so they do not 404. The reverse flow still uses the whole repo for now; **subfolder-aware** context (scoped to that path) is planned for a later change.

Stack

Next.js 16 (App Router), React 19, TypeScript, Tailwind CSS 4, GitHub API, Supabase (optional), Stripe (optional).

Configuration

Copy `.env.example` to `.env.local` and fill in at least one LLM API key.

Quick LLM (required)

The quick reverse endpoint supports four providers. Set **`GITREVERSE_QUICK_LLM`** to pin one, or leave it unset (`auto`) to let the app use whichever key it finds first:

Provider	Key env var	Model env var	Default model
Grok (xAI)	`XAI_API_KEY`	`XAI_MODEL`	`grok-3`
OpenRouter	`OPENROUTER_API_KEY`	`OPENROUTER_MODEL`	`google/gemini-2.5-pro`
Azure OpenAI	`AZURE_OPENAI_API_KEY` + `AZURE_OPENAI_BASE_URL`	`AZURE_OPENAI_MODEL`	`gpt-5.4`
Google AI Studio	`GOOGLE_GENERATIVE_AI_API_KEY`	`GOOGLE_AI_STUDIO_MODEL`	`gemini-2.5-pro`

In `auto` mode the order of preference is: Grok → OpenRouter → Azure → Google.

Azure quick reverse uses `gpt-5.4` by default with `AZURE_OPENAI_REASONING_EFFORT=medium`. Title generation also uses Azure and defaults to `gpt-5.4-mini` with reasoning disabled.