Back to Skills

create-multistage-dockerfile

pjt222
Updated 2 days ago
8 views
17
2
17
View on GitHub
Metaaidesign

About

This Claude skill generates optimized multi-stage Dockerfiles that separate build and runtime environments to create minimal production images. It helps when your images are too large, contain unnecessary build tools, or need deployment to constrained environments like edge computing. The skill covers builder/runtime stage separation, artifact copying, and targets like scratch, distroless, and Alpine bases.

Quick Install

Claude Code

Recommended
Primary
npx skills add pjt222/agent-almanac -a claude-code
Plugin CommandAlternative
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternative
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/create-multistage-dockerfile

Copy and paste this command in Claude Code to install this skill

Documentation

Create Multi-Stage Dockerfile

Build multi-stage Dockerfiles that produce minimal production images by separating build tooling from runtime.

When to Use

  • Production images are too large (>500MB for compiled languages)
  • Build tools (compilers, dev headers) are included in the final image
  • Need separate images for development and production from one Dockerfile
  • Deploying to constrained environments (edge, serverless)

Inputs

  • Required: Existing Dockerfile or project to containerize
  • Required: Language and build system (npm, pip, go build, cargo, maven)
  • Optional: Target runtime base (slim, alpine, distroless, scratch)
  • Optional: Size budget for final image

Procedure

Step 1: Identify Build vs Runtime Dependencies

CategoryBuild StageRuntime Stage
Compilersgcc, g++, rustcNot needed
Package managersnpm, pip, cargoSometimes (interpreted langs)
Dev headers-dev packagesNot needed
Source codeFull source treeOnly compiled output
Test frameworksjest, pytestNot needed

Step 2: Structure the Multi-Stage Build

The core pattern: build in a fat image, copy artifacts to a slim image.

# ---- Build Stage ----
FROM <build-image> AS builder
WORKDIR /src
COPY <dependency-manifest> .
RUN <install-dependencies>
COPY . .
RUN <build-command>

# ---- Runtime Stage ----
FROM <runtime-image>
COPY --from=builder /src/<artifact> /<dest>
EXPOSE <port>
CMD [<entrypoint>]

Step 3: Apply Language-Specific Patterns

Node.js (pruned node_modules)

FROM node:22-bookworm AS builder
WORKDIR /src
COPY package.json package-lock.json ./
RUN npm ci
COPY . .
RUN npm run build && npm prune --omit=dev

FROM node:22-bookworm-slim
RUN groupadd -r app && useradd -r -g app app
WORKDIR /app
COPY --from=builder /src/dist ./dist
COPY --from=builder /src/node_modules ./node_modules
COPY --from=builder /src/package.json .
USER app
EXPOSE 3000
CMD ["node", "dist/index.js"]

Python (virtualenv copy)

FROM python:3.12-bookworm AS builder
WORKDIR /src
RUN python -m venv /opt/venv
ENV PATH="/opt/venv/bin:$PATH"
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .

FROM python:3.12-slim-bookworm
COPY --from=builder /opt/venv /opt/venv
ENV PATH="/opt/venv/bin:$PATH"
WORKDIR /app
COPY --from=builder /src .
RUN groupadd -r app && useradd -r -g app app
USER app
EXPOSE 8000
CMD ["python", "app.py"]

Go (static binary to scratch)

FROM golang:1.23-bookworm AS builder
WORKDIR /src
COPY go.mod go.sum ./
RUN go mod download
COPY . .
RUN CGO_ENABLED=0 GOOS=linux go build -ldflags="-s -w" -o /server ./cmd/server

FROM scratch
COPY --from=builder /etc/ssl/certs/ca-certificates.crt /etc/ssl/certs/
COPY --from=builder /server /server
EXPOSE 8080
ENTRYPOINT ["/server"]

Rust (static musl binary)

FROM rust:1.82-bookworm AS builder
RUN apt-get update && apt-get install -y musl-tools && rm -rf /var/lib/apt/lists/*
RUN rustup target add x86_64-unknown-linux-musl
WORKDIR /src
COPY Cargo.toml Cargo.lock ./
RUN mkdir src && echo "fn main() {}" > src/main.rs \
    && cargo build --release --target x86_64-unknown-linux-musl \
    && rm -rf src
COPY . .
RUN touch src/main.rs && cargo build --release --target x86_64-unknown-linux-musl

FROM scratch
COPY --from=builder /src/target/x86_64-unknown-linux-musl/release/myapp /myapp
EXPOSE 8080
ENTRYPOINT ["/myapp"]

Got: Final image contains only the runtime and compiled artifacts.

If fail: Check COPY --from=builder paths. Use docker build --target builder to debug the build stage.

Step 4: Choose Runtime Base

BaseSizeShellUse Case
scratch0 MBNoStatic Go/Rust binaries
gcr.io/distroless/static~2 MBNoStatic binaries + CA certs
gcr.io/distroless/base~20 MBNoDynamic binaries (libc)
*-slim50-150 MBYesInterpreted languages
alpine~7 MBYesWhen shell access needed

Note: Alpine uses musl libc. Some Python wheels and Node native modules may not work. Prefer -slim (glibc) for interpreted languages.

Step 5: Build Args Across Stages

ARG APP_VERSION=0.0.0

FROM golang:1.23 AS builder
ARG APP_VERSION
RUN go build -ldflags="-X main.version=${APP_VERSION}" -o /server .

FROM gcr.io/distroless/static
COPY --from=builder /server /server
ENTRYPOINT ["/server"]

Build with: docker build --build-arg APP_VERSION=1.2.3 .

Note: ARG before FROM is global. Each stage must re-declare ARG to use it.

Step 6: Compare Image Sizes

# Build both variants
docker build -t myapp:fat --target builder .
docker build -t myapp:slim .

# Compare sizes
docker images --format "table {{.Repository}}\t{{.Tag}}\t{{.Size}}" | grep myapp

Got: Production image is 50-90% smaller than the build stage.

Validation

  • docker build completes for all stages
  • Final image does not contain build tools (compilers, dev headers)
  • docker run works correctly from the slim image
  • Image size is significantly reduced vs single-stage
  • COPY --from=builder paths are correct
  • No source code leaks into the production image

Pitfalls

  • Missing runtime libraries: Compiled code may need shared libraries (libc, libssl). Test the slim image thoroughly.
  • Broken COPY --from paths: The artifact path must match exactly. Use docker build --target builder then docker run --rm builder ls /path to debug.
  • Alpine musl issues: Native Node.js addons and some Python packages fail on Alpine. Use -slim instead.
  • Global ARG scope: An ARG declared before FROM is available to FROM lines only. Re-declare inside each stage that needs it.
  • Forgetting CA certificates: scratch has no certificates. Copy /etc/ssl/certs/ca-certificates.crt from the builder or use distroless.

Related Skills

  • create-dockerfile - single-stage general Dockerfiles
  • create-r-dockerfile - R-specific Dockerfiles with rocker images
  • optimize-docker-build-cache - layer caching and BuildKit features
  • setup-compose-stack - compose configurations using multi-stage images

GitHub Repository

pjt222/agent-almanac
Path: i18n/caveman-lite/skills/create-multistage-dockerfile
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Related Skills

content-collections

Meta

This skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.

View skill

polymarket

Meta

This skill enables developers to build applications with the Polymarket prediction markets platform, including API integration for trading and market data. It also provides real-time data streaming via WebSocket to monitor live trades and market activity. Use it for implementing trading strategies or creating tools that process live market updates.

View skill

creating-opencode-plugins

Meta

This skill helps developers create OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It provides the plugin structure, event API specifications, and implementation patterns for JavaScript/TypeScript modules. Use it when you need to intercept, monitor, or extend the OpenCode AI assistant's lifecycle with custom event-driven logic.

View skill

sglang

Meta

SGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.

View skill