# God Mode Claude

> Drop this CLAUDE.md into your project. Watch Claude Code mass level up. One file. Zero config.

- **Type:** MCP server
- **Install:** `agentstack add mcp-0rnot-god-mode-claude`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [0rnot](https://agentstack.voostack.com/s/0rnot)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [0rnot](https://github.com/0rnot)
- **Source:** https://github.com/0rnot/god-mode-claude

## Install

```sh
agentstack add mcp-0rnot-god-mode-claude
```

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

## About

# ⚡ GOD MODE CLAUDE

### One file. Drop it in. Claude Code mass levels up.

```
Before: "Claude, build a login page"  →  gets spaghetti, breaks half your app
After:  "Claude, build a login page"  →  clean code, tests pass, docs updated
```

[](https://opensource.org/licenses/MIT)
[](https://docs.anthropic.com/en/docs/claude-code)
[](#why-under-200-lines)

---

## What is this?

A single `CLAUDE.md` file that makes Claude Code mass dramatically better at its job.

Not a collection. Not a framework. **One file.**

```bash
curl -o CLAUDE.md https://raw.githubusercontent.com/0rnot/god-mode-claude/main/CLAUDE.md
```

Drop it into your project root. Done.

---

## What changes?

| Without | With God Mode |
|:--------|:-------------|
| Claude edits files it shouldn't | ✅ Hard file boundaries enforced |
| Writes code, doesn't test it | ✅ Forced verification loop — tests before marking done |
| Guesses when confused | ✅ Asks clarifying questions when 
🧠 The Thinking Rules — How Claude reasons about your code

```
1. Read before you write — always check existing code patterns first
2. If confidence 

🛡️ The Safety Rules — What Claude must never do

```
1. Never modify: generated files, lock files, .env, migrations, vendor/
2. Before destructive operations: always ask for explicit confirmation
3. Never use bare except/catch — always handle specific errors
4. Never commit secrets, API keys, or credentials
```

⚡ The Quality Rules — How Claude writes better code

```
1. Write tests for every new function — run them before marking done
2. Keep functions under 50 lines, files under 300 lines
3. Use the project's existing patterns — don't introduce new paradigms
4. Prefer explicit over clever — readable beats concise
```

📦 The Output Rules — How Claude communicates

```
1. Show only changed lines, not entire files
2. Use Conventional Commits: feat/fix/docs/refactor(scope): description
3. One logical change per commit — no "fix everything" commits
4. Don't repeat code the user just showed you
```

---

## Why under 200 lines?

Anthropic's own engineers found that **CLAUDE.md files over 200 lines degrade Claude's performance**. The model starts ignoring or "forgetting" rules. God Mode stays lean by design — every line earns its place.

> "The best CLAUDE.md is short, opinionated, and enforces verification."
> — Anthropic internal best practice

---

## Customize It

The file has clearly marked sections. Swap in your stack:

```markdown
## Tech Stack          ← Change this
## Commands            ← Change this
## Rules               ← Keep most of these
## Off-Limits          ← Add your generated/vendor paths
```

---

## FAQ

**Q: Does this work with Cursor / Windsurf / other AI editors?**
A: It's optimized for Claude Code, but most rules are universal. Rename to `.cursorrules` or equivalent.

**Q: Will this conflict with my existing CLAUDE.md?**
A: Merge the rules you like. The [Safety Rules](#the-rules-inside) and [Quality Rules](#the-rules-inside) are the highest-impact additions.

**Q: What if my project is huge?**
A: Use `.claude/rules/` directory for topic-specific rules. Keep the root `CLAUDE.md` as the master controller under 200 lines.

---

## Contributing

Found a rule that improves Claude Code's output? Open a PR. But:
- ✅ Must be battle-tested in a real project
- ✅ Must fit within the 200-line budget (something else gets cut)
- ❌ No language/framework-specific rules (keep it universal)

---

**One file. Massive difference.**

**⭐ Star if Claude Code just got better for you.**

*Made by [0rnot](https://github.com/0rnot)*

## Source & license

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

- **Author:** [0rnot](https://github.com/0rnot)
- **Source:** [0rnot/god-mode-claude](https://github.com/0rnot/god-mode-claude)
- **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:** yes
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** yes
- **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/mcp-0rnot-god-mode-claude
- Seller: https://agentstack.voostack.com/s/0rnot
- 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%.