containerization-research

What problem does it solve?

This Skill solves the problem of irreproducible research environments where results differ across machines, HPC nodes, and collaborator setups due to dependency drift and inconsistent runtime configurations.

Core Features & Use Cases

• Dockerfile-based environment freezing: Build pinned, layered container images for a consistent Python/runtime stack.
• HPC-ready deployment with Singularity/Apptainer: Run the same research environment safely on clusters without requiring root daemons.
• Multi-service research orchestration: Use Docker Compose to run analysis alongside supporting services (e.g., databases, tracking, notebooks) in a single reproducible workflow.

Use Case Example: You need to publish a “research compendium” (data + code + environment) so colleagues can exactly rerun your analysis and reproduce figures on both your workstation and an HPC cluster.

Quick Start

Use this skill to containerize your Python research stack by instructing the AI: “Generate a Dockerfile and a Singularity definition that pins numpy/pandas/scipy versions, sets up a non-root user, and provides commands to run the same analysis script locally and on SLURM.”