A Postgres port published as "5432:5432" is open on every network interface, and Docker's port rules skip most host firewalls. This skill finds that and ten other risky Compose settings, and tells you how to fix each one.

Your agent runs a scanner on your docker-compose.yml and compose.yaml files (overrides included). It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach for every finding. Secret values are never printed: the report shows the variable name and length only.

What it checks

BLOCK (fix before shipping)

• C01: a password, token or key written as a literal value in environment, or inside a database URL

• C02: privileged: true

• C03: the Docker socket mounted into a container

REVIEW (check before release)

• C11: an image with no tag or :latest

• C12: a database or cache port published on every interface

• C13: network_mode, pid or ipc set to host

• C14: dangerous capabilities such as SYS_ADMIN or ALL

• C15: seccomp, AppArmor or SELinux turned off

• C16: host system folders such as / or /etc mounted

• C17: a database that accepts logins with no password

• C18: a well-known default password such as postgres or changeme

What's included

• The scanner (Python 3.9+, standard library only)

• A rules guide with the reason and safer approach for each rule

• A risky and a safe example Compose file, and the real report the scanner produces on the risky one

Limits

It reads each Compose file on its own. It does not merge override files or resolve extends, include or ${VAR} values (YAML anchors and merge keys are followed), and it cannot see host firewall rules. NOT FLAGGED is not a guarantee.

Contents

One ZIP file with 8 files: SKILL.md, the scanner in scripts/ (3 Python files), references/rules.md (every rule with its reason and safer approach), and examples/ (2 example files plus report.md, the real scanner output on the risky example).

How to use: This is an agent skill: a folder with a SKILL.md file and Python scripts that an AI coding agent such as Claude Code, Cursor, or Codex CLI can run. Unzip it into your agent's skills folder (for Claude Code: ~/.claude/skills/), then ask your agent to check your project. The scripts also run on their own with Python 3.9 or newer, with nothing to install. They read your files only: no network access, and nothing is sent anywhere.