repl-interaction

name	repl-interaction
description	Patterns for controlling REPL sessions (Python, IPython, Node.js, Ruby, etc.) via terminalcp

REPL Interaction with terminalcp

This skill provides comprehensive patterns for managing and automating REPL (Read-Eval-Print Loop) sessions using terminalcp MCP. It enables AI agents to execute code interactively, explore APIs, and prototype solutions in real-time.

Overview

terminalcp enables:

• Interactive code execution in multiple languages
• Real-time output capture and parsing
• Persistent REPL sessions for stateful exploration
• Parallel REPL management for multi-language projects
• History and context preservation

The MCP server runs via npx @mariozechner/terminalcp@latest --mcp and provides full terminal emulation for REPL interaction.

Python REPL

Starting Python REPL

{
  "action": "start",
  "command": "python3 -i",
  "name": "python-repl"
}

Starting IPython (Enhanced REPL)

{
  "action": "start",
  "command": "ipython",
  "name": "ipython-repl"
}

Basic Code Execution

// Simple expression
{
  "action": "stdin",
  "id": "python-repl",
  "data": "2 + 2\r"
}

// Variable assignment
{
  "action": "stdin",
  "id": "python-repl",
  "data": "x = 42\r"
}

// Import module
{
  "action": "stdin",
  "id": "python-repl",
  "data": "import math\r"
}

// Use imported module
{
  "action": "stdin",
  "id": "python-repl",
  "data": "math.sqrt(16)\r"
}

Multi-Line Code Execution

// Start multi-line code
{
  "action": "stdin",
  "id": "python-repl",
  "data": "def greet(name):\r"
}

{
  "action": "stdin",
  "id": "python-repl",
  "data": "    return f'Hello, {name}!'\r"
}

// Empty line to finish
{
  "action": "stdin",
  "id": "python-repl",
  "data": "\r"
}

// Call the function
{
  "action": "stdin",
  "id": "python-repl",
  "data": "greet('Alice')\r"
}

Capturing REPL Output

{
  "action": "stdout",
  "id": "python-repl"
}

Returns the current terminal state including all executed code and output.

Stream Mode for Real-Time Monitoring

{
  "action": "stream",
  "id": "python-repl",
  "since_last": true
}

Returns only new output since last check, useful for monitoring long-running computations.

IPython Advanced Features

Magic Commands

// Time execution
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "%timeit sum(range(1000))\r"
}

// Run external script
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "%run script.py\r"
}

// Show history
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "%history\r"
}

// List variables
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "%whos\r"
}

// Edit code in external editor
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "%edit myfunction\r"
}

Object Introspection

// Get help
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "str.split?\r"
}

// Detailed help
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "str.split??\r"
}

// List methods
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "dir(str)\r"
}

// Type inspection
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "type(my_variable)\r"
}

Shell Commands in IPython

// Run shell command
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "!ls -la\r"
}

// Capture shell output
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "files = !ls *.py\r"
}

// Use Python variables in shell
{
  "action": "stdin",
  "id": "ipython-repl",
  "data": "!echo {my_var}\r"
}

Node.js REPL

Starting Node REPL

{
  "action": "start",
  "command": "node",
  "name": "node-repl"
}

Basic JavaScript Execution

// Simple expression
{
  "action": "stdin",
  "id": "node-repl",
  "data": "2 + 2\r"
}

// Variable declaration
{
  "action": "stdin",
  "id": "node-repl",
  "data": "const x = 42\r"
}

// Array methods
{
  "action": "stdin",
  "id": "node-repl",
  "data": "[1, 2, 3].map(n => n * 2)\r"
}

// Require module
{
  "action": "stdin",
  "id": "node-repl",
  "data": "const fs = require('fs')\r"
}

Multi-Line Code

// Start function definition
{
  "action": "stdin",
  "id": "node-repl",
  "data": "function greet(name) {\r"
}

{
  "action": "stdin",
  "id": "node-repl",
  "data": "  return `Hello, ${name}!`\r"
}

{
  "action": "stdin",
  "id": "node-repl",
  "data": "}\r"
}

// Call function
{
  "action": "stdin",
  "id": "node-repl",
  "data": "greet('Bob')\r"
}

Async/Await in REPL

// Define async function
{
  "action": "stdin",
  "id": "node-repl",
  "data": "async function fetchData() { return 'data' }\r"
}

// Call with await (Node 16+)
{
  "action": "stdin",
  "id": "node-repl",
  "data": "await fetchData()\r"
}

Node REPL Commands

// Show help
{
  "action": "stdin",
  "id": "node-repl",
  "data": ".help\r"
}

// Load file
{
  "action": "stdin",
  "id": "node-repl",
  "data": ".load script.js\r"
}

// Save session to file
{
  "action": "stdin",
  "id": "node-repl",
  "data": ".save session.js\r"
}

// Exit REPL
{
  "action": "stdin",
  "id": "node-repl",
  "data": ".exit\r"
}

Ruby REPL (irb)

Starting Ruby REPL

{
  "action": "start",
  "command": "irb",
  "name": "ruby-repl"
}

Ruby Code Execution

// Simple expression
{
  "action": "stdin",
  "id": "ruby-repl",
  "data": "2 + 2\r"
}

// String manipulation
{
  "action": "stdin",
  "id": "ruby-repl",
  "data": "'hello'.upcase\r"
}

// Array operations
{
  "action": "stdin",
  "id": "ruby-repl",
  "data": "[1, 2, 3].map { |n| n * 2 }\r"
}

// Define method
{
  "action": "stdin",
  "id": "ruby-repl",
  "data": "def greet(name); \"Hello, #{name}!\"; end\r"
}

// Call method
{
  "action": "stdin",
  "id": "ruby-repl",
  "data": "greet('Charlie')\r"
}

Lua REPL

Starting Lua REPL

{
  "action": "start",
  "command": "lua",
  "name": "lua-repl"
}

Lua Code Execution

// Simple expression
{
  "action": "stdin",
  "id": "lua-repl",
  "data": "print(2 + 2)\r"
}

// Table manipulation
{
  "action": "stdin",
  "id": "lua-repl",
  "data": "t = {1, 2, 3}\r"
}

{
  "action": "stdin",
  "id": "lua-repl",
  "data": "print(t[1])\r"
}

Advanced Patterns

API Exploration Workflow

// 1. Start IPython session
{
  "action": "start",
  "command": "ipython",
  "name": "api-explore"
}

// 2. Import library
{
  "action": "stdin",
  "id": "api-explore",
  "data": "import requests\r"
}

// 3. Inspect module
{
  "action": "stdin",
  "id": "api-explore",
  "data": "dir(requests)\r"
}

// 4. Get help on specific function
{
  "action": "stdin",
  "id": "api-explore",
  "data": "requests.get?\r"
}

// 5. Test API call
{
  "action": "stdin",
  "id": "api-explore",
  "data": "response = requests.get('https://api.github.com')\r"
}

// 6. Inspect response
{
  "action": "stdin",
  "id": "api-explore",
  "data": "response.json()\r"
}

// 7. Capture all output
{
  "action": "stdout",
  "id": "api-explore"
}

Data Analysis Session

// 1. Start IPython
{
  "action": "start",
  "command": "ipython",
  "name": "data-analysis"
}

// 2. Import pandas
{
  "action": "stdin",
  "id": "data-analysis",
  "data": "import pandas as pd\r"
}

// 3. Load data
{
  "action": "stdin",
  "id": "data-analysis",
  "data": "df = pd.read_csv('data.csv')\r"
}

// 4. Explore data
{
  "action": "stdin",
  "id": "data-analysis",
  "data": "df.head()\r"
}

{
  "action": "stdin",
  "id": "data-analysis",
  "data": "df.describe()\r"
}

// 5. Analyze
{
  "action": "stdin",
  "id": "data-analysis",
  "data": "df.groupby('category').mean()\r"
}

// 6. Get results
{
  "action": "stdout",
  "id": "data-analysis"
}

Multi-Language Development

// Start Python REPL for backend logic
{
  "action": "start",
  "command": "python3 -i",
  "name": "backend-repl"
}

// Start Node REPL for frontend logic
{
  "action": "start",
  "command": "node",
  "name": "frontend-repl"
}

// Test Python API
{
  "action": "stdin",
  "id": "backend-repl",
  "data": "def process_data(data): return data.upper()\r"
}

// Test JavaScript frontend
{
  "action": "stdin",
  "id": "frontend-repl",
  "data": "const formatData = (data) => data.toLowerCase()\r"
}

// List all active REPLs
{
  "action": "list"
}

Code Prototyping Workflow

// 1. Start REPL
{
  "action": "start",
  "command": "python3 -i",
  "name": "prototype"
}

// 2. Test idea
{
  "action": "stdin",
  "id": "prototype",
  "data": "def factorial(n): return 1 if n <= 1 else n * factorial(n-1)\r"
}

// 3. Test with examples
{
  "action": "stdin",
  "id": "prototype",
  "data": "factorial(5)\r"
}

// 4. Check edge cases
{
  "action": "stdin",
  "id": "prototype",
  "data": "factorial(0)\r"
}

{
  "action": "stdin",
  "id": "prototype",
  "data": "factorial(1)\r"
}

// 5. Refine implementation
{
  "action": "stdin",
  "id": "prototype",
  "data": "def factorial_iterative(n):\r    result = 1\r    for i in range(2, n + 1):\r        result *= i\r    return result\r\r"
}

// 6. Compare performance
{
  "action": "stdin",
  "id": "prototype",
  "data": "import time\r"
}

{
  "action": "stdin",
  "id": "prototype",
  "data": "start = time.time(); factorial(100); print(time.time() - start)\r"
}

// 7. Capture final output
{
  "action": "stdout",
  "id": "prototype"
}

Interactive Documentation Generation

// 1. Start IPython
{
  "action": "start",
  "command": "ipython",
  "name": "doc-gen"
}

// 2. Import module
{
  "action": "stdin",
  "id": "doc-gen",
  "data": "import mymodule\r"
}

// 3. Get function signatures
{
  "action": "stdin",
  "id": "doc-gen",
  "data": "import inspect\r"
}

{
  "action": "stdin",
  "id": "doc-gen",
  "data": "inspect.signature(mymodule.my_function)\r"
}

// 4. Get docstrings
{
  "action": "stdin",
  "id": "doc-gen",
  "data": "mymodule.my_function.__doc__\r"
}

// 5. Test examples from docs
{
  "action": "stdin",
  "id": "doc-gen",
  "data": "mymodule.my_function('test')\r"
}

// 6. Extract for documentation
{
  "action": "stdout",
  "id": "doc-gen"
}

Session State Management

// Save current REPL state to file
{
  "action": "stdin",
  "id": "python-repl",
  "data": "import dill\r"
}

{
  "action": "stdin",
  "id": "python-repl",
  "data": "dill.dump_session('session.pkl')\r"
}

// Later, restore state
{
  "action": "stdin",
  "id": "python-repl",
  "data": "dill.load_session('session.pkl')\r"
}

Best Practices

REPL Session Hygiene

1. Name sessions descriptively: Use name that reflects purpose (e.g., "api-test", "data-explore")
2. Clean up variables: Use del variable to free memory
3. Restart when needed: Stop and start fresh sessions to avoid state pollution
4. Use stream mode for long operations: Avoid blocking on compute-intensive tasks

Error Handling

// Catch and inspect errors
{
  "action": "stdin",
  "id": "python-repl",
  "data": "try:\r    risky_operation()\rexcept Exception as e:\r    print(f'Error: {e}')\r    import traceback\r    traceback.print_exc()\r\r"
}

Output Parsing

• Use stdout action after commands that produce output
• Parse REPL prompt patterns to extract results
• Use stream with since_last: true for incremental output
• Strip ANSI codes if needed for clean parsing

Managing Long-Running Operations

// Start long computation
{
  "action": "stdin",
  "id": "python-repl",
  "data": "result = expensive_computation()\r"
}

// Check progress with stream mode
{
  "action": "stream",
  "id": "python-repl",
  "since_last": true
}

// Don't block - continue other work
// Check back later with stdout
{
  "action": "stdout",
  "id": "python-repl"
}

Multi-Session Coordination

// List all active sessions
{
  "action": "list"
}

// Stop specific session
{
  "action": "stop",
  "id": "python-repl"
}

// Stop all sessions (clean shutdown)
// Iterate through list results and stop each

Common REPL Workflows

Quick Script Testing

1. Start REPL
2. Paste/type code snippets
3. Test with various inputs
4. Capture output
5. Refine code
6. Save final version to file

Library Evaluation

1. Start REPL
2. Import library
3. Explore API with dir() and help
4. Test key functions
5. Measure performance with %timeit
6. Document findings

Debugging Production Issues

1. Start REPL with production environment
2. Import relevant modules
3. Reproduce issue with test data
4. Inspect intermediate states
5. Test fixes
6. Verify solution

Interactive Configuration

1. Start REPL
2. Load configuration
3. Test different settings
4. Validate outcomes
5. Save working configuration

This skill provides comprehensive patterns for AI-driven REPL interaction using terminalcp's persistent session management.