# Agent Firewall

> Input/output safety for AI agents: known-pattern injection and obfuscation scan, URL/IP, secrets.

## Facts
- Page: https://tashan.sh/capability/pkg-agent-firewall-mcp
- tashan id: pkg:agent-firewall-mcp
- Source: https://github.com/mlawsonking/MCP
- npm: https://www.npmjs.com/package/agent-firewall-mcp
- Type: npm
- Category: security
- tashan score: 53.0 / 100
- Adoption: 36.0
- Upkeep: 55.0
- Freshness: 85.0
- Evidence coverage: 100% of the inputs this score can use
- Health: active
- Instruction depth: thin
- npm downloads: 363/week
- License: MIT
- Official: no

## Install

```sh
claude mcp add agent-firewall -- npx -y agent-firewall-mcp
```

## Security audit
- Known advisories: 0
- Install-time script: none declared
- Build provenance: not attested

Permissions are read from DECLARED dependencies only. Nothing is executed, so an empty result means "nothing declared", never "nothing possible".

---
Measured 2026-09-13 by tashan (https://tashan.sh) from public evidence. Scorer s5.