The Universal Adapter: Driving the Model Context Protocol (MCP) Client with Workload Automation

Introduction
The Model Context Protocol (MCP) Client integration is an innovative, open-standard capability designed to enhance HCL Automation Orchestration users. By standardizing how your central orchestration engine connects to external tools, infrastructure, and operational utilities, this integration allows enterprises to securely discover and invoke specialized data sources and command layers without the overhead of building bespoke plugins for every internal IT asset.

The Model Context Protocol (MCP) Client can be downloaded from Automation Hub to instantly bring plug-and-play architectural flexibility to your Workload Automation setup.

​Overview of Model Context Protocol (MCP) Client integration
The Model Context Protocol (MCP) Client integration for Workload Automation provides a cutting-edge, open-standard solution for connecting your automation workflows directly to external infrastructure, tools, and operational utilities. Developed as an open-source specification to prevent enterprise application fragmentation, MCP standardizes how orchestrators interact with specialized data sources and command layers. As a secure MCP Client, this integration allows Workload Automation to dynamically discover and invoke tools exposed by any compliant MCP Server, providing a plug-and-play framework that eliminates the need to develop custom, bespoke plugins for every internal or niche IT asset.

Key features:
The Model Context Protocol (MCP) Client integration is built to maximize architectural flexibility and standardization across hybrid IT ecosystems:

• Unified Plug-and-Play Framework: Invoke deterministic diagnostics, database utilities, and cloud operations across any infrastructure exposing an MCP interface without writing custom integration code.
• Dynamic Tool Discovery: Connect to local or remote MCP Servers at task definition time, automatically parsing available primitives (tools, resources, and prompts) directly from the Universal Orchestrator Console.
• Type-Safe Parameter Passing:Pass precise arguments and variables from the orchestrator to remote tool endpoints, ensuring exact execution across infrastructure boundaries. 
• Structured Property Capture: Automatically capture text-based and structured outputs returned by the MCP Server, exposing them natively as job properties for downstream routing and decision-making.
• Secure Transport and Auth Support: Establish stateful, secure sessions using modern transport standards (such as SSE or local I/O)  with flexible authentication options (Basic, Bearer, or API keys). 
• Native Ecosystem Compatibility: Fully integrated with Workload Automation, allowing you to seamlessly interleave open protocol actions with traditional deterministic IT jobs.

Use Cases:
The Model Context Protocol (MCP) Client integration functions as a universal adapter for modern, automated operations:

• Automated Infrastructure Health Monitoring: Instantly invoking specialized diagnostic tools on internal servers to triage anomalies and aggregate log files immediately after a failure event occurs.
• No-Code Tool Chain Stitching: Rapidly linking proprietary, in-house database scripts or security scanners into main production flows without waiting for dedicated vendor plug-in development.
• Dynamic Security Vulnerability Analysis: Triggering an MCP Server (such as the HCL AppScan MCP Server) to parse code repositories, pull SAST/DAST results, and return structured risk reports directly into the operational workflow.
• Cross-System Incident Triage: Coordinating between external monitoring systems and ticketing databases via standardized tool calls to assemble an un-attended diagnostic summary.

Example workflow: Automated Infrastructure Health Triage
The Model Context Protocol (MCP) Client integration serves as an automated first responder for enterprise environments. In an Automated Infrastructure Health Monitoring scenario, a high-priority scheduled batch process fails late at night, triggering an alert inside Workload Automation.

Instead of waiting for an engineer, the orchestrator instantly executes an MCP Client task mapped to an internal operational MCP Server. It invokes a tool called diagnose_failed_jobs, passing the list of failed job IDs as  arguments. The MCP Server connects to the underlying logs, queries the monitoring framework, and aggregates a structured summary. Workload Automation reads this  output property and triggers a sequential task invoking a second tool, suggest_remediation. The prescriptive corrective action is automatically parsed and forwarded directly to the incident management system, significantly compressing the mean time to repair (MTTR) before human intervention is required.

Getting started with Model Context Protocol (MCP) Client plugin
This section describes the properties and configurations of the Model Context Protocol (MCP) Client integration.
The MCP Client Plug-in allows enterprise workload automation jobs to interact with external AI tool servers through the open Model Context Protocol (MCP). By utilizing this plug-in, workload automation (WA) jobs can invoke tools provided by any MCP-compliant server and capture the tool's text output as structured job output properties.

Core Concepts

• Tool: A callable function exposed by the MCP server,with a name, description, and input schema that defines its expected arguments.

Request Lifecycle
The plug-in creates a Streamable HTTP connection to the MCP server, performs an MCP handshake to negotiate capabilities, invokes the requested tool, and maps the tool's results to the job output properties before closing the connection.
Specify the following information:

Connection Info
Use this section to connect to Model Context Protocol (MCP) and establish a connection to the works correctly. A connection object contains all the parameters needed to access a specific specific MCP server:

MCP server:
The MCP server depends on the following details:

• Server URL: This is the base URL of the target MCP server (for example, http://my-mcp-server:8080).
• Endpoint: This refers to the Streamable HTTP endpoint path that is appended to the server URL. If this is omitted, the plug-in defaults to the standard MCP path, which is /mcp (for example, http://my-mcp-server:8080/mcp).
• Auth Token:: This provides a static Bearer token for authenticated MCP servers. The token is stored in an encrypted format in the WA metadata database. Leave this field empty if you are using public servers that do not require authentication.
• Timeout (seconds): This sets the maximum time, in seconds, that the plug-in waits for the server to respond. It applies to both the initial connection and individual tool invocation requests. The value must be between 1 and 3600 seconds, and the default is set to 30 seconds.

Test Connection
Click the Test Connection button to ensure that your connection to Model Context Protocol (MCP) is working correctly.

Action
Jobs are defined with plugin type MCP and action callTool.

Tool Name:
This refers to the specific tool that invokes on the MCP server. The Workload Automation job editor features a searchable dropdown menu that connects to the server in real-time, allowing the operator to select from the tools currently available on the server without leaving the job editor.

Arguments (JSON):
This field is for the input arguments needed by the tool, provided as a JSON object string (for example, `{"city": "London", "unit": "celsius"}`). If the tool does not require any arguments, you can leave this field empty.

​Use Client Transport Preference:
When enabled, this option enforces the use of the transport protocol specified in the connection settings, overriding the agent's default preference. This is especially useful when working with agents that support multiple protocols but require a specific binding for optimal performance.

job details :
The following YAML snippet illustrates a Workload Automation job definition using the Model Context Protocol (MCP). plug-in:
===============================================================
kind: JobDefinition
apiVersion: workloadautomation.hcl.com/v1
metadata:
  name: mcp-weather-check
  description: "Invoke a weather tool on an MCP server"
spec:
  taskType: mcp
  taskAction: callTool
  connection:
    type: MCP Server Connection
    name: my-mcp-connection
  parameters:
    toolName: get_current_weather
    arguments: |
      {"city": "London", "unit": "celsius"}

In conclusion:
The Model Context Protocol (MCP) Client integration transforms Workload Automation into an incredibly versatile, future-proof command center. By adopting a universal interface standard for data and tool connectivity, it strips away the engineering overhead of building custom middleware—allowing your enterprise to expand its automation perimeter with unprecedented speed, security, and architectural simplicity.

---

Savitha Gururaj, Senior Technical Writer & Project Lead
13+ Years Experience. Specializing in API Documentation (REST/JSON/XML), DITA-XML, Content Strategy, and Agile Technical Communication. Extensive experience with FrameMaker, Oxygen XML, MadCap Flare, and Confluence. Passionate about transforming complex engineering systems into intuitive content solutions.

---

Ernesto Carrabba, Product Manager, HCL Software
Ernesto Carrabba is the Product Manager for HCL Clara, HCL HERO and HCL Workload Automation. Ernesto is a very dynamic product manager with experience in building and launching IoT products, combined with a master's degree in mechanical engineering and study researches on Augmented and Virtual Reality.

---