---
title: "Network Security"
description: "Nexus OS controls network access at multiple levels. Three network modes are available: Mode Description Use Case none No network access Pure computation agents restricted Allowlist only API calling agents open Full acce"
resource: https://www.aiagents.nexus/docs/manual/security/network-security
generated: { by: "process:nexus-agent-assets", at: 2026-09-07T09:13:03Z }
status: stable
---

# Network Security

Nexus OS controls network access at multiple levels.

## Network Policies

Three network modes are available:

| Mode | Description | Use Case |
|------|-------------|----------|
| `none` | No network access | Pure computation agents |
| `restricted` | Allowlist only | API-calling agents |
| `open` | Full access | Web scraping agents |

## Allowlist Configuration

```yaml
agents:
  researcher:
    sandbox:
      network: restricted
      allowed_hosts:
        - "*.openai.com"
        - "api.anthropic.com"
        - "api.tavily.com"
      blocked_ports:
        - 22    # SSH
        - 3306  # MySQL
```

## TLS Verification

All outbound HTTPS connections verify TLS certificates. Self-signed certificates are rejected unless explicitly allowed:

```yaml
agents:
  internal-agent:
    sandbox:
      tls_verify: false  # Only for internal services
```

## Rate Limiting

Network requests can be rate-limited per agent:

```yaml
agents:
  scraper:
    sandbox:
      rate_limit: 10/s    # Max 10 requests per second
      burst_limit: 50     # Allow bursts up to 50
```

## Dashboard Monitoring

The dashboard shows network activity per agent:

```bash
naos dashboard
# Navigate to Agents > [agent] > Network tab
```