mcp_Server

This MCP server is the Spring Boot integration for the Model Context Protocol (MCP) that enables seamless interaction between a Spring-based application and the Goose AI Agent. The setup assumes you build a Spring Boot application that includes the MCP client integration and then run Goose to coordinate AI-assisted capabilities within your app ecosystem. The workflow typically involves building a JAR for your Spring Boot extension, launching it as an MCP-enabled service, and using Goose sessions to drive AI tasks or agent behaviors inside the running extension. The README guides you through building the JAR, installing Goose, and running a Goose session, which together enable you to leverage MCP for model-driven interactions within your Spring application.

Prerequisites:

• Java Development Kit (JDK 17+ recommended; verify compatibility with your Spring Boot project)
• Gradle (or the Gradle wrapper included in your project)
• Goose AI Agent and its client tooling as documented by Goose
• Basic familiarity with running Java JARs and Spring Boot projects

Installation steps:

1. Ensure prerequisites are installed

• JDK: confirm with

java -version

• Gradle (or use the project wrapper):

./gradlew help

• Goose: follow the official Goose docs to install and configure the AI Agent.

2. Build the Spring Boot/MCP project

./gradlew clean build

The resulting JAR will typically be located under target/ (for example: target/mcp_feature-0.0.1-SNAPSHOT.jar).

3. Run the MCP-enabled Spring Boot extension

# Example command if you expose the JAR directly
java -jar target/mcp_feature-0.0.1-SNAPSHOT.jar

4. Install and start Goose sessions (as per Goose docs)

• Follow Goose setup instructions to install the client.
• Start a session and, if needed, pass in the extension JAR path or configuration so Goose can interact with MCP.

5. Optional WSL considerations (as described in the README):

• Create and use a run-extension.sh script to launch the JAR inside WSL, ensure Java is installed, and make the script executable before starting Goose with your extension.

Tips and common issues:

• Ensure the JAR name in the run command matches the built artifact (target/mcp_feature-0.0.1-SNAPSHOT.jar in the example).
• If you run into Java memory issues, adjust JVM options via JAVA_OPTS (e.g., -Xms512m -Xmx2g).
• When using WSL, follow the steps to create a small launcher script (run-extension.sh) and make it executable before starting Goose sessions.
• Goose integration relies on MCP primitives; ensure the Goose client is properly installed and reachable from the Spring application environment.
• If your project uses a different artifact name or version, update the mcp_config accordingly.
• The mcp_config assumes a Java-based MCP server; if you containerize the app, you can adapt to docker-based execution instead.