Trademarks of their respective owners. KissMySkills is not affiliated with or endorsed by them.
How it works
Ask it something hard.
📎 ai-skill-for-docker.txtLOADED ✓
Write a Dockerfile for my Node 20 Express API. It should be small, cache npm installs and not run as root.
CLAUDE · WITH THIS SKILLYou get: a two-stage Dockerfile in a code block -> a deps stage on node:20-alpine running npm ci with a cache mount and omit=dev -> a runtime stage that copies node_modules, sets NODE_ENV, switches to USER node and uses exec-form CMD -> a .dockerignore that keeps node_modules, .git and .env out of the context -> one-line reasons for the two stages and the built-in node user, plus a stated assumption about the entry file and a TypeScript variant -> the next action: the docker build and docker run commands to verify on port 3000.
Questions before you buy? A person answers, usually the same day: hello@kissmyskills.com
// the file itself
Read it before you buy it.
The first lines of the file your AI will read, exactly as delivered. Nothing rewritten for the page.
Use the name Once the file is loaded, talk to it by name: “Serge, check this page.” That is what the name is for. It also keeps several skills apart in one chat.
// the skill your AI runs on
You are a senior engineer who has used Docker daily for years, optimising for small reproducible images and builds that hit the cache. You have been activated to write the Dockerfile, the compose stack and explain why a container exits.
KissMySkills.com | Install: paste this whole file into Claude Projects (Project instructions), a custom GPT (Instructions), a Gemini Gem, or your AI's system prompt. Then just ask questions the way you normally would. You are a senior engineer who has used Docker daily for years in local development, CI and production. You optimise for small reproducible images, builds that hit the cache, containers that run as non-root, and compose stacks a new teammate brings up with one command. Tone: direct, practical, step-by-step when a sequence matters. · Core objects: images, layers, containers, volumes, bind mounts, networks (bridge, host, none, user-defined with DNS), registries (Docker Hub, GHCR, ECR), contexts, the Engine daemon. · Dockerfile traps: FROM pinned by tag or digest, RUN and cache invalidation, COPY vs ADD, ENV vs ARG (build-time only), EXPOSE (documentation only), USER for non-root, ENTRYPOINT vs CMD in exec vs shell form, HEALTHCHECK, multi-stage with COPY --from, .dockerignore. · BuildKit: the default builder; docker buildx for multi-platform (linux/amd64, linux/arm64), RUN --mount=type=cache, RUN --mount=type=secret so credentials never land in a layer, cache export to a registry or GitHub Actions cache. · Compose (docker compose v2): services, build, ports, volumes, env_file, depends_on with condition: service_healthy, healthcheck, networks, profiles, restart policies, compose.override.yaml, compose watch. · Image hygiene: slim or alpine bases, dependencies before source, caches cleaned in the same RUN, no latest in production, scanning with docker scout or Trivy, docker history to read layers. · Runtime: PID 1 and signals (exec form or tini), bind mount ownership, --memory and --cpus limits, read-only root filesystem, dropped capabilities, logging drivers. · Networking: containers reach each other by service name on a user-defined network; -p host:container publishes ports; localhost inside a container is not the host. · Gotchas: cache busted by copying the project before installing dependencies, secrets baked into layers via ENV or COPY, containers exiting because CMD ends, permission denied on mounted volumes, disk full from dangling images and unused volumes. · Docker Desktop on macOS or Windows, or Engine on Linux? Docker and Compose versions? · Language and framework, plus build and start commands. This decides the Dockerfile shape. · Target: local dev, CI or production, and which architectures? · Paste the existing Dockerfile, compose file and .dockerignore rather than starting from scratch.
Trademarks of their respective owners. KissMySkills is not affiliated with or endorsed by them.
// ask your assistant
Or let your AI set up this skill for you.
Go to kissmyskills.com/pages/connector and set up the "AI Skill for Docker" skill (handle: ai-skill-for-docker) for me
Paste it into Claude, ChatGPT, Claude Code, Cursor or Codex. Your AI walks you through connecting KissMySkills (about 3 minutes, 7-day free trial) and loads the skill. What your AI will read →
What you're actually buying
This Claude/ChatGPT skill is a Docker AI assistant for developers: it writes multi-stage Dockerfiles with pinned bases, cached dependency layers and a non-root user, a matching .dockerignore, compose.yaml stacks with healthchecks, depends_on conditions and named volumes, buildx commands and GitHub Actions jobs for amd64 and arm64 images, and a prioritised review of the Dockerfile you already have.
It knows BuildKit cache and secret mounts, ENTRYPOINT vs CMD and PID 1 signal handling, ENV vs ARG, user-defined networks, bind mount permissions, Docker Desktop vs Engine differences and the reasons a container exits or cannot reach the host. It does not build, run or push anything; it prepares files and commands you run in your terminal, and marks assumptions about your app.
Works the same way in Claude, ChatGPT and Gemini. One purchase, download the skill file, install in under 2 minutes, then paste your stack and ask. Pairs with Rami, DevOps Engineer, the persona skill; this one is the helper inside Docker.
// how to install
Under 2 minutes
Four steps. Any AI chat.
01
Download the file
After checkout, the download link lands in your inbox. Save the .txt file anywhere on your device.
02
Open your AI chat
Any AI chat you already use: Claude, ChatGPT, Gemini, Grok, Copilot or another.
03
Paste the prompt
Open the .txt file, copy the prompt section and paste it into Claude or ChatGPT - or add it to a Project so it is always there.
04
Start working
Your AI is now configured as a specialist. Ask it anything inside its domain.
No technical knowledge required.
With Library there is nothing to download. Connect your AI once, then ask for any skill by name.
It turns Claude or ChatGPT into a Docker AI assistant that writes and reviews your Dockerfiles, compose files and build pipelines. Name your stack and target, or paste what you have, and it returns pasteable files with pinned bases, correct layer order, non-root users and healthchecks, plus the exact commands to build and verify.
Does it operate Docker for me?+
No. The skill never builds an image, starts a container or pushes to a registry. It prepares the Dockerfile, .dockerignore, compose.yaml, buildx command or CI job, explains each non-obvious line, and you run the commands in your own terminal or pipeline.
Can it make my Docker image smaller or my build faster?+
Yes. Paste your Dockerfile with the current image size or build time and it rewrites it: dependency layer before source copy, BuildKit cache mounts, a slim base, build tools left out of the final stage, and it tells you how to confirm the gain with docker history and docker image ls.
Does it help debug a Docker container that exits or cannot connect?+
Yes. Paste the docker logs output or the error and it walks the checks in order: whether CMD exits immediately, docker inspect for env and mounts, exec into the container, published ports, user-defined network DNS and the localhost vs host.docker.internal mistake, then names the likely cause and the fix.
Ready to give your AI a specialist?
This one and 2,500+ others, inside Claude, ChatGPT, Claude Code, Codex or Cursor. Cancel anytime.