MCP for DevOps, NetOps, and SecOps: Real-World Use Cases and Future Insights

MCP for DevOps, NetOps, and SecOps: Real-World Use Cases and Future
Insights

In the previous post on MCP for DevOps: Architecture and Components, we discussed what MCP is and is not. We dove into a few architectural components and gently touched on use cases. Now, let’s explore a few possible use cases for MCP in DevOps/NetOps/SecOps.

I have cherry-picked a few customer and partner use cases I’ve personally worked with and found appropriate for our discussion. My list will not be exhaustive, but it should give you a solid view of practical uses for MCP. Let your mind ponder the possibilities in your environment. 😃

In the YouTube series on MCP for DevOps, we will leverage some use cases to build a working implementation with MCP, tools, and Cisco products.

Recap – Model Context Protocol (MCP)

If you didn’t catch part 1 in this blog series on MCP for DevOps: Architecture and Components, check it out. But for now, here’s a quick level-set on MCP.

As illustrated in Figure 1, the Model Context Protocol (MCP) provides a uniform way to integrate an AI model into tools and services.

Figure 1. MCP with LLMs and Tools

MCP Overview

It is:

• A lightweight communication protocol designed specifically for AI agents and applications.
• Built to connect those agents to tools, APIs, databases, and file systems.
• Structured as a client/server architecture—simple and predictable.
• Plumbing

It is not:

• A messaging protocol for agent-to-agent communication.
• An LLM, database, AI assistant, or agent.
• A general-purpose integration platform.
• A replacement for your existing APIs or data bus.

Common MCP Use Cases

As mentioned above, MCP integrates AI applications, tools, data sources, APIs, etc. However, MCP, being a protocol, does not work alone. A client and server must use the protocol and complete the pairing.

When AI applications and agents integrate the MCP SDK for client use and create an MCP server to work on behalf of local or remote tools, the following typical use cases can facilitate a low-toil/high-reward outcome.

• Automating Routine Tasks:
  MCP can handle repetitive chores such as generating reports, managing GitHub repos, building Ansible playbooks, and managing CI/CD pipelines.

• Unified Data and Action Management:
  Think of MCP as your AI application or agent’s centralized hub for interacting with diverse systems such as observability solutions from Splunk, orchestration systems such as Cisco NSO, and AI security platforms such as Cisco AI Defense.

• Enhanced Context and Decision-Making:
  MCP-powered AI applications and agents provide richer context by accessing data from multiple sources, leading to faster, smarter decisions.

• Compliance and Security:
  MCP interactions across your systems can be secure, compliant, and auditable when used with standardized security protocols, processes, and tools.

As illustrated in Figure 2, the MCP Client (AI application, assistant, or agent) can use MCP Servers to integrate with multiple automation, observability, security, and collaboration systems by calling those through APIs, data sources, etc.

Figure 2. MCP with Tools, Services, Platforms

Unified Automation with MCP

DevOps Use Cases

• CI/CD Automation:
  AI applications using MCP can automate entire CI/CD pipelines, seamlessly managing builds, tests, deployments, and notifications through Cisco Webex.

• Efficient Code Management:
  GitHub MCP integration enables an AI application or agent to manage branches, review pull requests, triage issues, and scan for vulnerabilities.

• Infrastructure Automation:
  With MCP Server integrations for Terraform and Ansible, your AI agent can quickly and accurately provision infrastructure or modify settings.

• Streamlined Incident Response:
  Cisco Webex integrated with MCP helps your AI application or agent actively engage in troubleshooting and incident management, significantly reducing response times.

DevOps Scenario:
Imagine asking your AI application (Chat interface or even your IDE):

“Create a new release branch, run tests, deploy to staging, and send a notification to Cisco Webex.”

As illustrated in Figure 3, your AI application seamlessly orchestrates actions via GitHub, Docker, and Jenkins using MCP and sends updates through Cisco Webex.

Figure 3. MCP-Powered CI/CD Pipeline

Pipeline Automation with MCP

NetOps Use Cases

• Dynamic Network Management:
  MCP enables AI-driven management of network configurations using natural language, leveraging Cisco APIs or Infrastructure-as-Code (IaC) tools.

• Automated Network Monitoring:
  With MCP, you can use an AI application or agent to monitor network performance, detect anomalies, and automatically remediate issues via Cisco solutions like ThousandEyes, Meraki Dashboard, and many more.

• Cloud Infrastructure Automation:
  MCP allows you to use AI to manage cloud-based networking infrastructure, leveraging Kubernetes APIs and Cisco network controllers for intelligent automation.

NetOps Scenario:

“Add a new OSPF IPv6 route for the 2001:db8:cafe::1/64 network at Data Center A.”

As illustrated in Figure 4, using MCP, your AI application uses an MCP Server to interact with Cisco APIs or even NETCONF/RESTCONF to make OSPF routing updates. It immediately updates the NetOps team via Cisco Webex.

Figure 4. AI-driven Traffic Management using MCP

Network Automation with MCP

SecOps Use Cases

• Proactive Threat Response:
  AI agents using MCP swiftly detect and mitigate threats by adjusting firewall settings with Cisco Secure Firewall and automatically isolating compromised endpoints using Cisco Secure Endpoint.

• Automated Vulnerability Management:
  MCP integrations enable AI to identify vulnerabilities and generate immediate infrastructure or host configuration fixes through Ansible playbooks and Terraform providers.

• Real-time Incident Orchestration:
  With MCP, AI orchestrates comprehensive incident responses, isolating threats, deploying patches, and alerting teams via Cisco Webex.

As illustrated in Figure 5, the following scenario can be realized using MCP:

SecOps Scenario:
Upon receiving a notification that the system identified malware, your AI assistant uses various tools via MCP to immediately:

1. Isolates the infected device using Cisco Secure Endpoint APIs
2. Applies fixes through Ansible
3. Updates firewall policies
4. Informs your security team via Cisco Webex

Figure 5. Incident Management using MCP

Security Incident Automation with MCP

I have not scratched the surface of what is possible using AI, MCP, and an endless array of future MCP servers.

Future Outlook

MCP’s ecosystem continues to expand, promising deeper integrations with Cisco solutions and broader industry adoption. Expect more sophisticated cross-domain orchestration, streamlined cloud-hosted services, and AI-driven proactive optimizations. MCP is setting the stage for smarter, faster, and more secure tool-based operations.

Things to consider:

While MCP is great for AI applications interacting with external tools and data sources, today, it isn’t built for production-grade agent-to-agent composition, deployment, discovery, connectivity, or lifecycle management of agents. MCP is not yet built to manage the dynamic discovery of MCP Servers and the tools they represent.

It is also a Wild Wild West show on MCP Servers. Everyone is creating them. That is great as it shows interest in MCP and how easy it is to leverage the MCP SDK, indicating that MCP provides direct value. However, I caution you to carefully evaluate the MCP servers you leverage in your enterprise use cases. Downloading and using an unknown MCP Server that anyone can publish could cause harm if you don’t understand the tools, resources, etc., the MCP Server is built to use.

A few of the many possible security implications for MCP use include:

• Privilege escalation threats
• Observability into what each tool call is doing
• Dependency on additional code and packages for proper end-to-end encryption and trust

There is a good blog post on MCP security considerations at the community.cisco.com site: Overview of MCP and Its Security Architecture.

In the future, we will see services and tools that validate the code/image of a given MCP Server as we do with app stores, container images, etc. Until there is a standardized and well-understood way to ensure you are not using a harmful MCP Server, I would be extra vigilant about researching and truly understanding what the server is doing on your behalf.

What is next? We will continue this series on MCP for DevOps by getting into the hands-on side of MCP use. Stay tuned for some YouTube videos and more blogs on specific MCP Clients and MCP Servers that are great for Dev/Net/SecOps.

Share: