agentleFS
hugobatistahugobatista/agent-skills/skills/python/docker-python/SKILL.md

docker-python

Skill1 starsChanged 53 days agoMITView on GitHub

Dockerfile best practices for Python projects. Covers non-root user (ARG UID/GID/APP_USER), COPY --chown, runtime-writable files, and entrypoint diagnostics.

  • Reads credentials
  • Installs packages

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 5 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read assets/Dockerfile.tmpl, replace appname, and write it to the project root.

    line 9
  2. Very specific:

    NPM dependencies: use npm ci --ignore-scripts in Dockerfile stages to block postinstall attacks.

    line 51
  3. Distinctive:

    When setting up Docker for a project, offer to create a GHCR publish workflow.

    line 63
  4. Distinctive:

    Ask the user: *"Will this project be hosted on GitHub?"* If no, skip this section.

    line 67
  5. Very specific:

    Read config.toml from this directory.

    line 68
  6. Very specific:

    Ask the user for the GitHub owner (default: <github_owner from config>)

    line 69
  7. Distinctive:

    Ask the user for the URL base for badge links:

    line 70
  8. Very specific:

    Ask the user for the repo name (read from pyproject.toml [project].name if it's a Python project.

    line 74
  9. Very specific:

    Read assets/ghcr.yml.tmpl, then write it to .github/workflows/ghcr.yml (no substitutions needed).

    line 76
  10. Distinctive:

    Add the following badge to README.md:

    line 78
  11. Distinctive:

    Create the directory before writing:

    line 84

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Dockerfile

Read assets/Dockerfile.tmpl, replace appname, and write it to the project root.

Key patterns:

  • uv image pinned to tag@sha256:digest — guarantees the exact uv version; update the digest when a new stable release has been available for > 30 days
  • ARG UID=1000 GID=1000 APP_USER=appuser — configurable at build time via --build-arg UID=568
  • Privileged RUN (package installs, user creation, app directory) happens before USER
  • COPY --chown=${UID}:${GID} sets ownership at build time — no runtime chown needed
  • /app is pre-created and chowned before any copy so the working directory is writable
  • Runtime-writable config files (e.g. config.js) should be pre-created with chmod 666 so docker-compose user: overrides still work
  • USER ${APP_USER} is the last step before ENTRYPOINT

Bind mounts

For host bind mounts in docker-compose, users must chown the host directory to the container UID:

sudo chown -R 1000:1000 /host/path

To use a custom UID, add user: "<UID>:<GID>" in docker-compose and ensure runtime-writable files have chmod 666 set at build time.

.dockerignore

.git/
__pycache__/
*.pyc
.venv/
.env
dist/
*.egg-info/
.mypy_cache/
.pytest_cache/
.coverage

Dependency pinning

  • Docker images: pin all FROM lines to tag@sha256:digest. Tags are mutable; digests guarantee the exact image.
    FROM python:3.13.12-slim@sha256:f1927c75e81efd1e091dbd64b6c0ecaa5630b38635a3d1c04034ac636e1f94c8 AS builder
  • NPM dependencies: use npm ci --ignore-scripts in Dockerfile stages to block postinstall attacks.
    RUN npm ci --prefer-offline --ignore-scripts
  • Python dependencies: export from lockfile with uv export and install with pip install --require-hashes to verify every package hash.
    RUN uv export --no-dev --no-emit-project -o requirements.txt
    RUN pip install --no-cache --require-hashes -r ./requirements.txt

GitHub Actions workflow

When setting up Docker for a project, offer to create a GHCR publish workflow. Ask the user: "Do you want to set up a GitHub Actions workflow to build and push this image to GHCR?"

If yes:

  1. Ask the user: "Will this project be hosted on GitHub?" If no, skip this section.
  2. Read config.toml from this directory. If it exists, extract github_owner and github_url_base for defaults. If not, ask the user for these values and write them to config.toml (one-time setup).
  3. Ask the user for the GitHub owner (default: <github_owner from config>)
  4. Ask the user for the URL base for badge links:
    • go.<owner>.com/gh (default, e.g. https://<github_url_base from config>)
    • github.com/<owner> (e.g. https://github.com/<github_owner from config>)
    • Custom — user provides their own
  5. Ask the user for the repo name (read from pyproject.toml [project].name if it's a Python project; otherwise ask directly)

Read assets/ghcr.yml.tmpl, then write it to .github/workflows/ghcr.yml (no substitutions needed).

Add the following badge to README.md:

[![GHCR Tag](https://img.shields.io/github/v/tag/<owner>/<repo>?logo=docker&logoColor=white&label=GHCR)](<base>/<repo>/packages)

Create the directory before writing:

mkdir -p .github/workflows

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.