# steinwall

The steinwall plugin for Claude Code and Codex enforces spending caps, stops dangerous shell commands, and scans plugins before installation.

- plugin · `type=plugin` · Claude Code, Codex · ★ 1596 repo stars
- Author: Maximilian Steinmetz
- License: FSL-1.1-MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/hashgraph-online/awesome-codex-plugins/tree/HEAD/plugins/maxxstm/steinwall
- Updated: 2026-10-29
- Security check: automated, 2026-11-16 (text and 3 script files) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/plugins/steinwall

## What it does

SteinWall runs entirely on your local machine using lifecycle hooks. It intercepts agent actions before execution takes place. Cost management includes preset caps of $25 per day and $300 per month. Once spending reaches the ceiling, further prompts and tool runs stay blocked until you adjust the limit. It also guards against critical mistakes, rejecting operations like destructive root deletion, force pushes, credentials leaks, and running downloaded scripts straight into a shell.

Stuck loops get broken up automatically. When an identical command fails three times without codebase modifications, SteinWall injects guidance, turning into a hard block at five attempts. The built-in scanner inspects MCP packages, subagents, and plugins for prompt injections or obfuscated instructions. Node.js 20 or newer on your `PATH` is required for execution.

## Install

### Claude Code (native platform)

**Plugin marketplace**
In the agent chat (not a shell):
```text
/plugin marketplace add hashgraph-online/awesome-codex-plugins
```
In the agent chat (not a shell):
```text
/plugin install steinwall@awesome-codex-plugins
```

## When to use it
- enforce daily and monthly spending caps on agent sessions
- stop destructive actions like `git push --force` or reading `.env` credentials
- break repetitive failure loops before token allowances run out
- inspect new MCP servers and plugins for prompt injections before adding them
- record project state into a handoff file when approaching usage thresholds

## Example prompts
- /steinwall:setup
- /steinwall:budget --daily 30 --monthly 400
- /steinwall:scan ./my-plugin

## Similar
- [Agent Guardrail](https://affpapa.org/en/nevibecode/mcp/agent-guardrail.md) — MCP server · 534 installs
- [nah](https://affpapa.org/en/nevibecode/plugins/nah.md) — plugin · ★ 494 repo stars
- [guard-pack](https://affpapa.org/en/nevibecode/plugins/guard-pack.md) — plugin · ★ 586 repo stars
- [config-guard](https://affpapa.org/en/nevibecode/plugins/config-guard.md) — plugin · ★ 586 repo stars
- [SatGate – Economic Firewall for AI Agents](https://affpapa.org/en/nevibecode/mcp/satgate-economic-firewall-for-ai-agents.md) — MCP server · ★ 11 repo stars
- [Mandare](https://affpapa.org/en/nevibecode/mcp/mandare.md) — MCP server · ★ 2 repo stars

From the same set: [hashgraph-online/awesome-codex-plugins](https://affpapa.org/en/nevibecode/by/hashgraph-online/awesome-codex-plugins.md) — 391 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
