ic3software/iccube-dev repository overview

docker run -d --name iccube-ai --restart unless-stopped \
  -p 8080:8080 \
  -e APP_DEFAULT_LLM_MODEL=openai::gpt-5.6-luna \
  -e OPENAI_API_KEY=your-api-key \
  ic3software/iccube-dev:ai-server-0.1.0

/opt/icCube-agent
   /config
      application.yaml
      logback.xml
      /agents
         ...
      /skills
         ...
   /bin
   /lib 

/home/ic3/icCube-agent-data

<aiComponentConfiguration active="true">

    <url>/icCube/ai/*</url>

    <!-- <filter>White List</filter> -->
    <filter>HTTP Form Authentication</filter>

    <!--
        The remote AI server (aka., agents) address where 
        the AI requests are actually processed.
    -->
    <server>http://localhost:8080</server>

    <!--
        How long to wait to establish the initial TCP connection 
        with the AI Server.

        Default: 10s
    -->
    <connectTimeoutS>10</connectTimeoutS>

    <!--
        How long to wait for the AI Server to actually respond 
        with the data once the connection is made.

        Default: 60s
    -->
    <timeoutS>300</timeoutS>

</aiComponentConfiguration>

<mcpComponentConfiguration active="true">

    <url>/icCube/mcp/*</url>

    <!-- <filter>White List</filter> -->
    <filter>MCP Authentication</filter>
    <filter>HTTP Form Authentication</filter>

    <!--
        The address of icCube used by AI Server when connecting to its MCP Server:

            e.g., E.g., http://icCube-address:8282/icCube/mcp

        When not defined, localhost:httpComponentConfiguration.portNumber is used:

            e.g., E.g., http://192.168.1.8:8282/icCube/mcp
    -->
    <address/>

</mcpComponentConfiguration>