# CyberAgents AI Security Firewall

> CyberAgents AI Security Firewall is a freemium MCP server for security. Hosted MCP server that screens AI agent traffic for prompt injection and risky tool calls. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Freemium
- **Starting price:** $49 a month
- **Platforms:** API, Web
- **Website:** https://cyberagents.app/open
- **Repository:** https://github.com/ghazwanalemara/sovereign-agent-skills
- **MCP Registry name:** io.github.GhazwanAlemara/cyber-agents
- **Categories:** [security](https://ai-powered.com/categories/security)
- **Listing updated:** 2026-11-19

## About

The README positions this as one of several live remote MCP servers in a larger skills suite, and the registry entry describes it as a service for filtering prompt injection in real time, inspecting MCP payloads and scoring trust between agents. The product page adds a middleware SDK that watches agent behaviour and blocks hijacked instructions and data exfiltration, plus a scanner that checks MCP tool definitions.

The page also lists features beyond the core filter: autonomous pull requests for supply-chain fixes, ephemeral GVisor sandboxing for untrusted code, and SOC2/HIPAA audit evidence generation. These are described as product claims and are not independently verified here. Access is through a streamable-http endpoint with a free Community tier.

It suits teams running AI agents that call tools or other agents and want a security layer in front of them. Developers who only need a local, self-hosted scanner may prefer the companion open-source CLI repositories the README links to.

## Features

- **Prompt injection filtering:** Screens agent inputs in real time for malicious instructions and prompt hijacking attempts.
- **MCP tool validation:** Checks MCP tool calls and payloads before they reach the agent's tools.
- **Agent-to-agent trust scoring:** Scores identity and trust when one agent calls another over A2A.
- **Autonomous code remediation:** Detects supply-chain vulnerabilities and opens pull requests with suggested fixes.
- **Sandboxed code execution:** Runs untrusted code in ephemeral GVisor containers, according to the product page.
- **Compliance evidence logging:** Generates SOC2 and HIPAA evidence for agent actions and data access, per the site.

## Good for

- Filter untrusted user or web content before it reaches an AI agent
- Validate MCP tool calls and payloads in an agent pipeline
- Check trust when autonomous agents delegate work to other agents
- Produce audit trails of agent tool calls for compliance reviews

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http cyber-agents https://cyberagents.app/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://cyberagents.app/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "cyber-agents": {
      "url": "https://cyberagents.app/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "cyber-agents": {
      "type": "http",
      "url": "https://cyberagents.app/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "cyber-agents": {
      "serverUrl": "https://cyberagents.app/mcp"
    }
  }
}
```

### Grok Bot (Grok Bot desktop app → Connect apps → Add a custom connector)

```sh
https://cyberagents.app/mcp
```

Enter a Name (for example "CyberAgents AI Security Firewall"), choose Type: Remote HTTPS, paste this as the Server URL and choose Add plugin.

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "cyber-agents": {
      "type": "http",
      "url": "https://cyberagents.app/mcp"
    }
  }
}
```

## FAQ

### Is there a free plan?

Yes. The product page lists a Community tier at $0/month with up to 3 repositories, 10,000 agent actions per month, basic auto-remediation PRs and community support.

### How much does the Pro plan cost?

The product page lists Pro at $49/month, with unlimited repositories, the real-time prompt injection SDK, automated PR auto-merge and priority support.

### How is it connected to an AI client?

The registry entry lists a remote streamable-http endpoint at cyberagents.app/mcp. The README says an API key is claimed from the open page and set as CYBERAGENTS_API_KEY.

### Is CyberAgents AI Security Firewall free?

CyberAgents AI Security Firewall has a free plan, with paid plans for more usage or features. Paid plans start at $49 a month.

### How do I add CyberAgents AI Security Firewall to Claude Code?

Run this command in your terminal: claude mcp add --transport http cyber-agents https://cyberagents.app/mcp

### What are the best alternatives to CyberAgents AI Security Firewall?

The closest alternatives to CyberAgents AI Security Firewall are APEX Agent Safety & Guardrails, mcp-audit, xyz MCP by CyberXYZ, Diogenes Agent Guardrails & Sandbox MCP and SLF Guardian.

## Where this comes from

- Description: written from the project's README, the project's website and its MCP Registry entry.
- Status: checked automatically by AIPoweredBot on 19 November 2026.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [APEX Agent Safety & Guardrails](https://ai-powered.com/mcp/apex-agent-safety-and-guardrails): APEX Agent Safety & Guardrails is a paid MCP server for security.
- [mcp-audit](https://ai-powered.com/mcp/mcp-audit-4): mcp-audit is an MCP server for security.
- [xyz MCP by CyberXYZ](https://ai-powered.com/mcp/xyz-mcp-by-cyberxyz): xyz MCP by CyberXYZ is an MCP server for security.
- [Diogenes Agent Guardrails & Sandbox MCP](https://ai-powered.com/mcp/diogenes-agent-guardrails-and-sandbox-mcp): Diogenes Agent Guardrails & Sandbox MCP is a freemium MCP server for security.
- [SLF Guardian](https://ai-powered.com/mcp/slf-guardian): SLF Guardian is a freemium MCP server for security.
- [KillerSkills](https://ai-powered.com/mcp/killerskills): KillerSkills is a freemium MCP server for security.
- [Redthread](https://ai-powered.com/mcp/redthread): Redthread is a freemium MCP server for security.
- [Supernova MCP Gateway & Shield](https://ai-powered.com/mcp/supernova-mcp-gateway-and-shield): Supernova MCP Gateway & Shield is an open-source MCP server for security.

---
Source: https://ai-powered.com/mcp/cyberagents-ai-security-firewall
