# Ops Docker Optimization

> Techniques for creating tiny, fast, and secure production-grade images.

- Skill: `jcorpac/ops-docker-optimization` (Agent Skill)
- Install (CLI): `npx skillmds@latest add jcorpac/ops-docker-optimization`
- Raw SKILL.md: https://api.skillmd.com/api/skills/jcorpac/ops-docker-optimization/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- Author: jcorpac (https://skillmd.com/u/jcorpac)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/jcorpac/ops-docker-optimization

---


# Docker Optimization

Production images should be minimal, fast to pull, and secure.

## Multi-stage Builds
- **Pattern**: Use a heavy build image (with compilers/tools) and copy only the compiled artifacts to a tiny runtime image (e.g., `alpine`).
- **Benefit**: Reduces image size by up to 90% and removes build-time vulnerabilities.

## Layer Reduction
- **Chaining**: Chain `apt-get install` commands and clean up in the same `RUN` layer to avoid bloated images.

## Best Practices
- **Alpine & Distroless**: Use the smallest possible OS base.
- **Cache Mounting**: Use `--mount=type=cache` in BuildKit to persist pip/npm caches across builds.