docker-validation_skill

Overview

This skill validates Dockerfiles and Docker Compose files against best practices, security standards, and modern Compose syntax. It produces a prioritized report with concrete fixes, and can generate validation scripts and CI checks on request. Use it to make Docker configurations production-ready, secure, and efficient.

How this skill works

I inspect project files to locate Dockerfiles and Compose manifests, run available linters (Hadolint, DCLint) and Docker CLI checks, and perform manual rule-based analysis. The checks cover syntax, multi-stage build correctness, layer optimization, secret exposure, runtime user, Compose version deprecation, and service-level recommendations. Results are classified by severity and delivered as a concise remediation report; optional scripts or CI snippets can be created.

When to use it

• Validate one or more Dockerfiles before a release
• Audit Docker Compose files for modern syntax and security
• Verify correct multi-stage builds and artifact copying
• Find and remove secrets accidentally baked into images
• Prepare Docker configurations for CI/CD and production deployments

Best practices

• Pin base image versions; avoid :latest
• Use multi-stage builds so final image is minimal and free of build tools
• Run processes as a non-root user and specify USER in final stage
• Prefer COPY over ADD, clean package caches, and combine RUN steps
• Remove obsolete Compose version fields and validate with docker compose config
• Manage secrets externally (Compose secrets, environment files, or secret stores)

Example use cases

• Scan a repository to produce a validation report listing critical, high, medium, and low issues
• Convert an existing Dockerfile to a multi-stage build and verify artifact transfers
• Update docker-compose.yml to remove deprecated version field and pin service images
• Create a bash validation script and GitHub Actions job to run checks in CI
• Run a security audit to ensure no secrets, non-root runtime, and recommended base images