One db_url: mysql://homeassistant:password@core-mariadb line, or one Wi-Fi password in an ESPHome file, is enough to leak your passwords the day you push your config to GitHub. And trusted_proxies: 0.0.0.0/0 lets anyone on the internet choose the IP address Home Assistant sees, so IP bans and local-only webhooks stop protecting you. This skill finds those and nine other Home Assistant configuration mistakes before you share, commit or expose your setup, and tells you how to fix each one.

Your agent runs a scanner on your Home Assistant config folder: configuration.yaml, automations.yaml, scripts, packages, blueprints, files pulled in with !include and !include_dir_*, and ESPHome device files kept beside them. It understands !secret, !include, !env_var and !input as references, not values, knows both trigger styles (trigger: webhook and platform: webhook), reads included files that hold only an http: block, an auth provider list or a list of platforms, and skips YAML that is not Home Assistant's (CI workflows, Docker Compose, Kubernetes). It checks .gitignore rules itself, including the common "ignore everything, then allow *.yaml" style. It never runs Home Assistant, calls its API or connects to anything. It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach. Secrets are never printed.

What it checks

BLOCK (fix before shipping)

• H01: passwords, tokens and keys written inline instead of !secret: secret-named keys, passwords in db_url and rtsp:// camera URLs, Authorization headers, tokens in URLs, long-lived access tokens and ESPHome API encryption keys

• H02: secrets.yaml, .storage or .cloud in a git repository with no .gitignore rule covering it

• H03: use_x_forwarded_for with trusted_proxies open to the whole internet (0.0.0.0/0, ::/0)

• H04: a trusted_networks login provider that trusts the whole internet

REVIEW (check before release)

• H11: allow_bypass_login: true, or public addresses in trusted_networks

• H12: external_url on plain http:// for an internet host

• H13: ip_ban_enabled: false, login_attempts_threshold: -1, or no threshold behind a reverse proxy or TLS

• H14: templates inside shell_command and command_line commands

• H15: webhook triggers with local_only: false

• H16: allowlist_external_dirs that open /, /config, /etc, /root or /home

• H17: verify_ssl: false for https hosts on the internet

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 config folder (configuration.yaml, secrets.yaml, automations.yaml, a package, an ESPHome device and .gitignore), and the real report the scanner produces on the risky one

Limits

It reads each YAML file as written and does not merge !include files or resolve !secret values. Settings made in the UI live in .storage and are not read; from Home Assistant 2026.8 that includes the http: settings, which are imported once from YAML into Settings > System > Network, so check them there too. Users, MFA, port forwarding and add-ons are not checked. NOT FLAGGED is not a guarantee.

Contents

One ZIP file with 21 files: SKILL.md, the scanner in scripts/ (5 Python files), references/rules.md (every rule with its reason and safer approach), and examples/ (13 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.