Skip to main content
Glama

Student Leave MCP App With Github Copilot Agent

mcp-student-leave-copilot-ab (-student-leave--ab)

A lightweight MCP server in Node.js, integrated with Github Copilot, to manage student leave requests and approvals.

Related MCP server: MCP Leave Manager Server

Getting Started

Create a project

mkdir mcp-student-leave-copilot-ab

cd mcp-student-leave-copilot-ab

npm init -y

npm i @modelcontextprotocol/sdk zod

node server.js

Student Leave MCP server (stdio) and Github Copilot agent ready!

Create server.js

Add npm scripts (optional)

{
  "type": "module",
  "scripts": {
    "start": "node server.js"
  }
}

Wire it up in Github Copilot (VS Code)

Github Copilot looks for MCP servers in its mcp.json

/Users/abhijeetgiram/Library/Application Support/Code/User/mcp.json

{
  "servers": {
    "mcp-student-leave-copilot-ab": {
      "command": "node",
      "args": [
        "/Users/abhijeetgiram/Workspace/Personal/mcp-student-leave-copilot-ab/server.js"
      ],
      "alwaysAllow": ["list_students", "get_student_details", "approve_leave"],
      "disabled": false
    }
  }
}

Use it from Github Copilot

In a Github Copilot chat, just ask things like:

  • “Show me the student list.” → Github Copilot calls list_students

  • “Get details of E002.” → Github Copilot calls get_student_details with { id: "S002" }

  • “Approve leave L1002 for E002.” → Github Copilot calls approve_leave with { studentId: "S002", leaveId: "L2002" }

Model Context Protocol (MCP)

  • It’s a new open standard by Anthropic that defines how AI models, tools, and apps can talk to each other.

  • Think of it like a “plugin protocol” for AI assistants.

  • MCP Server = Your code (like the student leave manager we built).

  • MCP Client = Something that talks to the server (Github Copilot, Cline, Claude, or any AI agent).

  • They communicate over:

    • stdio (local processes, great for Github Copilot in VS Code)

    • or HTTP/WebSocket (remote deployment, so web apps / React can call it).

  • MCP turns your Node.js script into a “tool plugin” for AI assistants.

Mock data

{ "students": [ { "id": "S001", "name": "Rahul Sharma", "class": "10A", "leaveBalance": 12 }, { "id": "S002", "name": "Priya Desai", "class": "9B", "leaveBalance": 15 }, { "id": "S003", "name": "Ananya Gupta", "class": "11C", "leaveBalance": 10 } ], "leaves": [ { "id": "L2001", "studentId": "S001", "from": "2025-09-01", "to": "2025-09-02", "days": 2, "status": "pending" }, { "id": "L2002", "studentId": "S002", "from": "2025-09-05", "to": "2025-09-07", "days": 3, "status": "pending" } ] }

Miscellaneous

pwd

/Users/abhijeetgiram/Workspace/Personal/mcp-student-leave-copilot-ab

/Users/abhijeetgiram/Library/Application Support/Code/User/settings.json

To create a file on Mac using the command line, use the touch command. For example, to create a file named

touch mcp.json

You can also create and edit a file using nano:

nano mcp.json

Available Tools

3 tools
approve_leaveApprove leaveC

Approve a pending leave and update balance.

Inputs: leaveId (required), studentId (required).

get_student_detailsGet student details by IDA

Return one student and their leave requests.

Inputs: id (required).

list_studentsGet student listA

Return all students with remaining leave balance.

Inputs: no named parameters.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.6/5.0

    Scored across 3 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose: listing all students, fetching a single student's details, and approving a leave. There is no overlap or ambiguity between them.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun snake_case pattern: list_students, get_student_details, approve_leave. The naming style is uniform and predictable.

    Tool Count5/5

    Three tools is well-scoped for a focused student leave copilot, each serving a distinct function without redundancy. The count is within the ideal 3-15 range.

    Completeness4/5

    The core workflow of viewing students and approving leave is covered, but there is no reject or batch approval capability. Minor gaps exist, but they are workaroundable by iterating through students.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers