Skip to main content
Agent skill

Maxxing Infra

Put a site online, create a database or an S3 bucket on this server the standard way, add another server, or answer "where is X / what runs where / is it up" - sites (Caddy), PostgreSQL/SQLite, Garage buckets, domains, other servers over SSH, Cloudflare (Workers, Pages, R2, D1) and Hostinger. Use when the user wants something hosted, published, stored, or asks about their sites, databases, buckets, domains or servers.

All install options
Quick install · Personal (~/.claude/skills)Project / Cursor / Desktop →
mkdir -p ~/.claude/skills/maxxing-infra && curl -fsSL 'https://raw.githubusercontent.com/makushingu/harness-maxxing/main/skills/maxxing-infra/SKILL.md' -o ~/.claude/skills/maxxing-infra/SKILL.md

Get the complete skill folder, including its scripts and reference files.

When to use this skill

Use Maxxing Infra when an AI agent needs a reusable SKILL.md workflow for this job: Put a site online, create a database or an S3 bucket on this server the standard way, add another server, or answer "where is X / what runs where / is it up" - sites (Caddy), PostgreSQL/SQLite, Garage buckets, domains, other servers over SSH, Cloudflare (Workers, Pages, R2, D1) and Hostinger. Use when the user wants something hosted, published, stored, or asks about their sites, databases, buckets, domains or servers.

When not to use it

Skip Maxxing Infra when the task is outside the coding category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.

How to install

Where will you use it?

Install scope
One-line install (Claude Code)
mkdir -p ~/.claude/skills/maxxing-infra && curl -fsSL 'https://raw.githubusercontent.com/makushingu/harness-maxxing/main/skills/maxxing-infra/SKILL.md' -o ~/.claude/skills/maxxing-infra/SKILL.md
  1. Claude Code personal install: place the skill and its bundled files at ~/.claude/skills/maxxing-infra/SKILL.md. For a shared project, commit the folder at .claude/skills/maxxing-infra/ instead.
  2. Current Claude Code detects SKILL.md changes in watched directories during the session. If you just created a top-level skills directory, run /reload-skills. Check that the skill is listed before testing its trigger.
  3. For Claude chat or Cowork in the Desktop app, manage uploaded skills in Customize > Skills; copying files into a local Claude Code folder is not a Desktop install. Other agents may use different install locations.

Full install guide for Claude, Cursor, and Codex →

Try a starter prompt

After installing the skill, paste this starter prompt into your agent to verify it triggers on a small task.

Starter prompt
Use the Maxxing Infra skill for this task: Use Maxxing Infra when this documented workflow matches the task. Confirm any required inputs or files before making changes.

What supports this recommendation?

71/100 provisional content score. This measures documentation, practical detail and source signals. It is not a measured task success rate or a security certification.

Documentation
77/100
Practical detail
68/100
Source evidence
63/100
Source signals
72/100
Model compatibility evidence
  • claudeinferred

    The skill text mentions Claude or a closely associated term. This is keyword-derived compatibility; no hands-on model test is recorded.

  • chatgptuntested

    No model-specific signal or recorded compatibility test was found.

  • geminiuntested

    No model-specific signal or recorded compatibility test was found.

  • copilotuntested

    No model-specific signal or recorded compatibility test was found.

  • llamauntested

    No model-specific signal or recorded compatibility test was found.

  • perplexityuntested

    No model-specific signal or recorded compatibility test was found.

  • mistraluntested

    No model-specific signal or recorded compatibility test was found.

  • grokuntested

    No model-specific signal or recorded compatibility test was found.

Inferred compatibility is based on content signals. Check the app, its available tools and setup requirements before running a skill.

What this skill does

# Infra: see everything, create on this server the standard way

The **Infra page** (Settings -> Features -> Infra, off by default) shows every site, database, bucket, domain and server with its status and links. After the first site, database or bucket you create, if the page is off (`"infra_page": false` in the state `config.json`), tell the user in one line that Settings -> Features -> Infra shows it live - never switch it on yourself. The command `maxxing infra` is the same thing in the terminal **and the only way to create things here**: never hand-write a Caddy block, a database or a bucket - use the recipes below, they validate, roll back on error, store credentials in the vault, register the resource and back it up.

## Look first ```bash maxxing infra status # sources (servers, Cloudflare, Hostinger), counts, everything down or to check maxxing infra list [--kind site|database|bucket|domain|server|worker|pages] [--json] maxxing infra refresh # gather now (10-60 s); the server also does it every 15 min when the page is on maxxing infra mail [--json] # e-mail (view only): mailboxes, Cloudflare / Hostinger forwards, DNS health of mail domains ``` `list --json` gives each resource's `id`, `urls`, `status`/`status_detail`, `details` (target port, service unit, repository, DNS records, certificate, sizes...) and `links` (domain <-> sites, site <-> server...). When you talk about resources in a reply, add an ` ```infra ` block so the UI shows them live: ```infra site: shop.example.com database: shop ``` (one line per resource: `site`, `database`, `bucket`, `domain`, `server`, `worker`, `pages`, `id`; or filters `project: acme`, `kind: database`, `location: <server name>`).

## Put a site online ```bash maxxing infra site add NAME --dir /path/to/folder [--project P] # static files maxxing infra site add NAME --port 3100 [--project P] [--keep-prefix] # an app listening on 127.0.0.1:3100 maxxing infra site add NAME (--dir D | --port N) --host shop.example.com [--dns] # its own address maxxing infra site list | site remove NAME | reserved ``` - **Default = path site**: `https://<address of Harness>/<NAME>/`, nothing to configure, works at once. The page is **sandboxed** (Content-Security-Policy sandbox): it cannot touch Harness, its session or its API, and has no localStorage/cookies of its own. Perfect for pages, demos, landing pages, reports, static builds. - Built sites must use **relative paths or the base `/<NAME>/`**: Vite `base: '/<NAME>/'`, Next `basePath`, Astro `base`. A build with absolute `/assets/...` loads Harness's files instead. - An app (`--port`) gets the path stripped (it sees `/`); `--keep-prefix` keeps `/<NAME>/` for apps configured with a base path. - **Use `--host` (advice, explain it) for**: an app with a login, a cart, cookies or localStorage, an API called from its own pages, anything in production. `--dns` creates the A record on Cloudflare when the token may edit DNS; otherwise tell the user which record to add (A `<host>` -> the server's IP, DNS only / grey cloud if the zone is proxied two levels deep). - Names: a-z 0-9 -, not a route of Harness (`maxxing infra reserved`). - An app needs a process: write a systemd unit (`/etc/systemd/system/<name>.service`, `Restart=always`, `WorkingDirectory=`, the port on 127.0.0.1 only), `systemctl enable --now`, then `site add --port`. Pick a free port (`ss -ltn`), 3100-3999 by convention.

## Databases ```bash maxxing infra db install # PostgreSQL, localhost only, + the nightly job "Infra: database backups" maxxing infra db create NAME [--project P] # role + database NAME; URL in the vault: DB_<NAME>_URL (not exposed) maxxing infra db sqlite /path/app.db [--project P] # declare a SQLite file: shown and backed up maxxing infra db backup # dump everything now into /var/backups/maxxing-db (7 kept per database) ``` - Ask before `db install` (it installs a package). The app reads its URL with `maxxing keys get DB_<NAME>_URL` (put it in the app's own env file, mode 600) - never print the password in the chat. - Remote access = SSH tunnel only (`ssh -L 5432:127.0.0.1:5432 root@server`). Never open 5432. - Databases in Docker are shown but not backed up by the job: say so if it matters.

## Buckets (S3 on this server) ```bash maxxing infra bucket install # Garage, single node, every port on 127.0.0.1 maxxing infra bucket create NAME [--public] [--project P] ``` - Keys in the vault: `S3_<NAME>_ACCESS_KEY_ID`, `S3_<NAME>_SECRET_ACCESS_KEY`, `S3_<NAME>_ENDPOINT` (`http://127.0.0.1:3900`), `S3_<NAME>_BUCKET`, `S3_<NAME>_REGION` (`garage`). Any S3 client works (path-style addressing). - `--public` serves its files at `https://<address of Harness>/<NAME>/`. - A bucket here lives on this disk, without a copy elsewhere: for big public files or production, advise Cloudflare R2 (free egress) - check the free disk first (`maxxing infra list --kind server`).

## Other servers ```bash maxxing infra server connect HOST [--user root] [--port 22] [--label "Shop server"] # finds a key that works maxxing infra server connect <Hostinger VPS id> # id = the N of server:hostinger:N in `list --kind server --json` maxxing infra server add NAME HOST [--key /root/.ssh/id_x] [--label L] # a given key maxxing infra server rename NAME "Label" | server list | server remove NAME ``` Prefer `connect`: for a Hostinger VPS it uses the key the VPS already accepts (read through the API) or adds the Harness key (`<state>/ssh/infra_ed25519`) to its root account through the Hostinger API (tell the user: it is a change on their VPS). For another server, when no key works it prints the line to run there. A running Hostinger VPS that already accepts one of our keys is connected by the refresh itself. The name is the short host name (`srv100200`, part of the ids); the label ("VPS Hostinger") is what the page shows. The server must accept SSH from this one without a password (key) and have `python3`. The probe only reads (Caddy/nginx sites, ports and services, Docker, PostgreSQL/MySQL including in containers, SQLite, Garage). To act on that server, use `ssh` yourself - the Infra page never changes anything remotely.

## Domains from any registrar `maxxing infra domain add example.com` (or Domains -> Add a domain): no key needed, its registrar, expiry, DNS provider and host come from RDAP and the public DNS. Use it for every domain the providers below do not list.

## Providers (keys in Settings -> Keys) - **Cloudflare**: `CLOUDFLARE_API_TOKEN` (+ `CLOUDFLARE_ACCOUNT_ID`). Read rights: Zone, DNS, Workers Scripts, Workers Routes, Pages, R2, D1, Registrar (Account settings: Read). `--dns` needs DNS Edit. A missing right only hides that part (`maxxing infra status` says which). - **Hostinger**: `HOSTINGER_API_TOKEN` (hPanel -> API): domains with their expiry, VPS, DNS zones. - **Vercel** `VERCEL_TOKEN`, **Netlify** `NETLIFY_AUTH_TOKEN`, **Supabase** `SUPABASE_ACCESS_TOKEN` (beta): sites and deployments, domains, Supabase databases/buckets/functions. **GitHub Pages**: the GitHub tokens of the vault. - Other registrars and hosts: no key, `maxxing infra domain add` (RDAP + public DNS: registrar, expiry, DNS provider, who hosts the address). A provider line in error: read its message in `maxxing infra status`.

## Organise `maxxing infra annotate ID --project P --note "..." [--hide|--show]` - attach a resource to a Harness project, leave a note, hide noise. The page's resource panel does the same, plus "Ask Claude".

## Alerts While the page is on: a site/server going down (and back up), a certificate expiring within 14 days, a domain expiring within 30 then 7 days (not auto-renewed), a failed backup -> the notifications feed (and Telegram if the feed thread is on).

Intended uses

  • Use Maxxing Infra when this documented workflow matches the task.

Implementation notes

  • Review the source instructions and adapt inputs before running the workflow.

Try it, then tell us how it went

A copied prompt does not tell us whether the task worked. Report your result after checking the output. We collect your tool and outcome, never your files or prompt.