Add Copilot Chat MCP support

.gitignore

@@ -24,6 +24,9 @@ env/
 
 # IDE
 .vscode/
+!.vscode/
+.vscode/*
+!.vscode/mcp.json
 .idea/
 *.swp
 *.swo
@@ -84,4 +87,4 @@ site/
 *.log
 *.tmp
 temp/
-tmp/
+tmp/

.vscode/mcp.json

@@ -0,0 +1,9 @@
+{
+  "servers": {
+    "shortkit-ml": {
+      "type": "stdio",
+      "command": ".venv/bin/python",
+      "args": ["-m", "shortcut_detect.mcp_server"]
+    }
+  }
+}

README.md

@@ -177,13 +177,22 @@ Custom conditions can be registered via `@register_condition("name")`. See [Cond
 
 ## MCP Server
 
-ShortKit-ML ships an [MCP](https://modelcontextprotocol.io/) server so AI assistants (Claude, Cursor, etc.) can call detection tools directly from chat — no Python script required.
+ShortKit-ML ships an [MCP](https://modelcontextprotocol.io/) server so AI assistants (Claude, Cursor, GitHub Copilot in VS Code, etc.) can call detection tools directly from chat — no Python script required.
 
 ```bash
 uv pip install -e ".[mcp]"
 ```
 
-Full setup instructions for **Claude Code** and **Claude Desktop**, available tools, file-based input, and troubleshooting: [**criticaldata.github.io/ShortKit-ML/mcp-server**](https://criticaldata.github.io/ShortKit-ML/mcp-server/).
+### GitHub Copilot in VS Code
+
+1. After installing the MCP extra, open VS Code and go to Settings (Ctrl+,).
+2. Search for "GitHub Copilot Chat: MCP" and open the JSON settings.
+3. Copy the server configuration from the repo's `.vscode/mcp.json` file into the `github.copilot.chat.mcp` object in your settings.
+4. Ensure the `"command"` path points to your virtual environment's Python (e.g., update `.venv/bin/python` to the full path if needed).
+5. Restart VS Code or reload the GitHub Copilot Chat extension.
+6. The ShortKit-ML tools will now be available in Copilot Chat sessions.
+
+Full setup instructions for **Claude Code**, **Claude Desktop**, and **GitHub Copilot in VS Code**, available tools, file-based input, and troubleshooting: [**criticaldata.github.io/ShortKit-ML/mcp-server**](https://criticaldata.github.io/ShortKit-ML/mcp-server/).
 
 ## Paper Benchmark Datasets
 

docs/mcp-server.md

@@ -1,6 +1,6 @@
 # MCP Server
 
-ShortKit-ML ships a [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server that exposes shortcut detection as callable tools. AI assistants like Claude can run full analyses — generate data, run detectors, compare methods — directly from chat, with no Python script required.
+ShortKit-ML ships a [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server that exposes shortcut detection as callable tools. AI assistants like Claude and GitHub Copilot in VS Code can run full analyses — generate data, run detectors, compare methods — directly from chat, with no Python script required.
 
 ## Prerequisites
 
@@ -18,8 +18,6 @@ uv run python -m shortcut_detect.mcp_server
 
 The server should start without errors. Press `Ctrl+C` to stop it.
 
----
-
 ## Option 1 — Claude Code (recommended)
 
 Claude Code reads a `.mcp.json` file from the project root. This file is already committed to the repo, so no manual setup is needed after installing the package.
@@ -215,3 +213,37 @@ To reuse a previous run:
 ```
 Get the summary for session_id "my-run-1"
 ```
+
+---
+
+## GitHub Copilot in VS Code
+
+GitHub Copilot Chat in VS Code can consume the same MCP server through Agent mode. This repository includes a checked-in `.vscode/mcp.json` template so the ShortKit-ML server is ready to use once your `.venv` is installed.
+
+### 1. Install dependencies
+
+```bash
+uv pip install -e ".[mcp]"
+```
+
+### 2. Open the workspace in VS Code
+
+Make sure the repo root is open so VS Code can read `.vscode/mcp.json`.
+
+### 3. Start Copilot Chat in Agent mode
+
+Open Copilot Chat, switch the mode dropdown to **Agent**, and open the tools list. You should see the `shortkit-ml` MCP server and the ShortKit-ML tools.
+
+### 4. Try a prompt
+
+```text
+Generate 200 synthetic samples with a linear shortcut and run probe, statistical, and hbac on them.
+```
+
+### 5. Customize the Python path if needed
+
+Edit `.vscode/mcp.json` and point `command` at your workspace `.venv/bin/python` path. Use the same `shortcut_detect.mcp_server` module entry point as Claude Code and Claude Desktop.
+
+### Manual verification
+
+The repository-level config and tests are in place, but the Copilot Chat smoke test itself still needs to be confirmed in VS Code by opening the tools panel and running a real prompt.

output/test_group_labels.csv

@@ -0,0 +1,201 @@
+group
+0
+1
+0
+0
+0
+1
+0
+0
+0
+1
+0
+0
+0
+0
+1
+0
+1
+1
+1
+0
+1
+0
+1
+1
+1
+1
+1
+1
+1
+1
+0
+0
+1
+1
+1
+0
+1
+0
+0
+0
+0
+0
+1
+1
+1
+1
+1
+0
+1
+1
+0
+1
+0
+1
+0
+1
+1
+0
+0
+0
+0
+0
+0
+0
+0
+1
+1
+0
+1
+1
+1
+1
+0
+1
+0
+1
+1
+1
+0
+1
+0
+1
+0
+1
+0
+0
+1
+0
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+0
+0
+1
+1
+1
+1
+1
+1
+1
+1
+0
+1
+0
+1
+1
+0
+1
+0
+1
+1
+0
+1
+0
+1
+0
+0
+1
+1
+0
+1
+1
+1
+0
+0
+0
+0
+0
+0
+0
+0
+0
+1
+0
+1
+1
+1
+0
+0
+0
+0
+1
+0
+0
+0
+0
+0
+1
+0
+1
+0
+1
+0
+0
+1
+1
+1
+0
+1
+0
+0
+1
+1
+0
+0
+1
+1
+1
+0
+0
+0
+0
+0
+0
+1
+0
+0
+0
+1
+0
+0
+1
+0
+0
+0
+0
+0
+1
+1
+1
+0
+0

output/test_group_labels.json

@@ -0,0 +1 @@
+[0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0]