sandbox-execution-guide

npx skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill sandbox-execution-guide --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

# Sandbox Execution Guide A skill for setting up and using sandboxed code execution environments for research computing. Covers containerized execution, security considerations, resource management, and integration with research workflows. ## Why Sandboxed Execution? Research code often requires: - Isolation from the host system for security - Reproducible environments across machines - Resource limits to prevent runaway computations - Multi-language support (Python, R, Julia, MATLAB) ## Docker-Based Sandboxes ### Creating a Research Container ```dockerfile # Dockerfile for a reproducible research environment FROM python:3.11-slim # System dependencies RUN apt-get update && apt-get install -y --no-install-recommends \ build-essential \ gfortran \ libopenblas-dev \ && rm -rf /var/lib/apt/lists/* # Create non-root user for security RUN useradd -m -s /bin/bash researcher USER researcher WORKDIR /home/researcher # Pin all dependencies COPY requirements.txt . RUN pip install --user --no-cache-dir -r requirements.txt # Copy project files COPY --chown=researcher:researcher . /home/researcher/project WORKDIR /home/researcher/project # Resource limits set at runtime, not build time CMD ["py

Run with CPU, memory, and time constraints
docker run \
Mount data as read-only, output directory as writable
Enter the reproducible environment
nix-shell shell.nix
Or use flakes for even better reproducibility
nix develop