Claude Code Plugins

Community-maintained marketplace

Feedback

Execute JavaScript or Python code in a secure Hyperlight VM sandbox with full isolation. Use this when users request running code, performing calculations, testing algorithms, or demonstrating code behavior with actual output.

Install Skill

Shared

Installs to .agents/skills, used by Codex, Amp, Warp, Cursor, OpenCode, and more.

CodexAmp
Warp
CursorOpenCode
Cline
Gemini CLI
GitHub Copilot
Personal

Available across projects.

$npx skills-installer add @jsturtevant/local-code-interpreter-tool/execute-code --client shared
Project

Writes to .agents/skills.

$npx skills-installer add @jsturtevant/local-code-interpreter-tool/execute-code -p --client shared
Note: Review the skill instructions before using it.

SKILL.md

name execute-code
description Execute JavaScript or Python code in a secure Hyperlight VM sandbox with full isolation. Use this when users request running code, performing calculations, testing algorithms, or demonstrating code behavior with actual output.

Execute Code Skill

Execute JavaScript or Python code in a secure Hyperlight VM sandbox with full isolation.

This skill is standalone and shareable - it can be copied to any .claude/skills/ directory and works independently of the original repository.

Requirements

  • Linux with KVM support (/dev/kvm)
  • Python 3.8+
  • Git (for cloning hyperlight-nanvix)
  • Rust toolchain (installed automatically by setup script)

When to Use

Use this skill when:

  • The user asks to run, execute, or test code
  • You need to verify calculations or algorithm correctness
  • You want to demonstrate code behavior with actual output

Standalone Installation

To use this skill in a new project or share it with others:

1. Copy the Skill

# Copy the entire execute-code directory to your project
cp -r /path/to/execute-code ~/.claude/skills/
# Or for project-specific use:
cp -r /path/to/execute-code /your/project/.claude/skills/

2. Run Setup

cd ~/.claude/skills/execute-code
./scripts/setup.sh

The setup script will:

  • Create a virtual environment in .venv
  • Install Rust nightly toolchain (if not present)
  • Clone and build hyperlight-nanvix from source
  • Configure everything for standalone use

3. Activate and Use

source .venv/bin/activate
python3 scripts/run.py --lang javascript --code 'console.log("Hello!");'

Quick Start

JavaScript

python3 scripts/run.py --lang javascript --code 'console.log(2 + 2);'

Python

python3 scripts/run.py --lang python --code 'print(2 + 2)'

Usage Notes

  • JavaScript: Use console.log() to see output
  • Python: Use print() to see output
  • Code runs in VM-isolated sandbox (hyperlight-nanvix)
  • The skill is self-contained with its own virtual environment

Examples

Calculate Fibonacci

python3 scripts/run.py --lang javascript --code '
const fib = n => n <= 1 ? n : fib(n-1) + fib(n-2);
for (let i = 0; i < 10; i++) console.log("fib(" + i + ") =", fib(i));
'

Prime Numbers

python3 scripts/run.py --lang python --code '
primes = [n for n in range(2, 50) if all(n % i for i in range(2, int(n**0.5)+1))]
print(f"Primes under 50: {primes}")
'

Execute from File

python3 scripts/run.py --lang python --file my_script.py

Manual Installation (Alternative)

If you prefer to install dependencies manually:

# 1. Install Rust nightly
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
rustup install nightly

# 2. Clone and build hyperlight-nanvix
git clone https://github.com/hyperlight-dev/hyperlight-nanvix.git
cd hyperlight-nanvix
pip install maturin
maturin develop --features python

Directory Structure

When installed standalone, the skill has this structure:

execute-code/
├── SKILL.md           # This file
├── REFERENCE.md       # Detailed API reference
├── scripts/
│   ├── run.py         # Main execution script
│   └── setup.sh       # Automated setup script
├── .venv/             # Virtual environment (created by setup)
└── vendor/            # hyperlight-nanvix source (created by setup)
    └── hyperlight-nanvix/

See Also

  • REFERENCE.md - Detailed API and configuration options
  • scripts/run.py - The self-contained execution script
  • scripts/setup.sh - Automated setup for standalone use