agentleFS
leonardourcileonardourci/pnyx/skills/pnyx-set-up/SKILL.md

pnyx-set-up

Skill4 starsChanged 2 days agoApache-2.0View on GitHub

Use when you are Hydra and your digest shows "Setup needed", or the owner asks what Pnyx still needs to be set up.

  • Installs packages

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

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

    After a fix, run setup check again and report only what is still missing.

    line 15

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Pnyx setup

Pnyx checks its setup each time a Hydra session opens, in its own code, and lists only what is missing under "Setup needed" in your digest. A complete setup prints nothing. pnyx-axi setup check --hydra <hydra> prints the same list again (exit 1 while anything is missing).

Each row names who fixes it. Do the you rows yourself, without asking. For owner rows that name a pnyx-axi command, run that command with --hydra <hydra> as usual, except setup, which serves every Hydra and refuses --hydra. Pnyx turns owner actions into a question card in your chat, and only the owner's Confirm runs them; when your session is set to Bypass permissions or Codex's never, they run at once instead, so report the result rather than a pending confirmation. For owner rows without a command, ask the owner in one short line. Greet a first setup with one line ("A few things to set up") before the cards; that line is the turn's reply, so never add NO_REPLY after it. When the owner dismisses one, note it in owner memory (preferences) and do not raise it again.

After a fix, run setup check again and report only what is still missing.

Setup of hooks, skills and the login service changes user-wide configuration. If a command returns CUSTOM_HOME_SETUP, keep the selected PNYX_HOME and ask whether the owner wants those changes for that home. Only then add --allow-custom-home; do not unset PNYX_HOME to get around the refusal. Use --preview --allow-custom-home after that choice to inspect the targets. A custom home's service also needs an unused --port that no installed Pnyx service reserves.

Rows

  • postgres: Pnyx's Postgres is not ready, and setup check lists only this row until it is. When Postgres cannot be reached, no Hydra loads. When it runs without wal_level=logical, or is older than 15, Hydras load but zero-cache cannot serve the page. Ask the owner to run the row's fix in a terminal. On macOS, pnyx-axi setup postgres starts Homebrew's postgresql@17 service with wal_level=logical and creates the home's database; it fails with brew install postgresql@17 as its next step when Postgres is not installed, and for a Postgres older than 15 the fix says to stop it and install postgresql@17. Elsewhere setup postgres only checks, and the fix names what to start or set. A refused role or password needs a working PGUSER or postgresUrl, not a start. When the postgresUrl setting names the Postgres, Pnyx does not manage it: the fix asks the owner to start that Postgres with wal_level=logical, or to correct or clear the URL with pnyx-axi setup postgres --url '<url>' (--url '' clears it) and restart Pnyx, since the server cannot start while that URL is wrong.
  • config: no local dispatch rules or account registry. Run pnyx-axi rule init --hydra <hydra> --reason "first-run setup" yourself; it saves the templates only when nothing is saved yet and keeps a file that exists. It registers starter accounts using CLAUDE_CONFIG_DIR and CODEX_HOME from the command's environment when set, or the template's folders otherwise; any whose login folder is missing show up as account:<name> rows next time.
  • account:<name>: a registered account's login folder does not exist. Ask the owner whether they use that account. If they do, they sign in with that folder; if not, run the pnyx-axi account remove command the row names so the owner confirms its removal.
  • login:claude, login:codex: the rules name a harness with no login. Ask the owner to sign in to it; Pnyx only stores login folders. When accounts are registered, the row also names an account set command: run it with the folder the owner signed in to, so the owner confirms it. If they do not use that harness, propose rules without it (rule propose set).
  • hooks: the reporting hooks are not installed for this Pnyx home. Run pnyx-axi setup hooks --hydra <hydra>; the owner confirms it. Codex then asks the owner to trust the new hooks once with /hooks.
  • skills: the global skill links are missing. Run pnyx-axi setup skills --hydra <hydra>; the owner confirms it. A conflict row means another skill sits at that path: raise no card, since setup skills refuses to replace it; ask the owner to move it aside, then raise the card.
  • quota-axi: not on PATH, so dispatch cannot rank accounts by quota. Ask the owner to install quota-axi; quota stays unknown without it, and a start with several unknown candidates can require an explicit harness or rule.
  • service: no login service, so Hydras stop when this server stops. Run pnyx-axi setup without --hydra; the owner confirms it, and the service takes over once a manually started server stops. This row only applies on macOS and Linux; a foreground server works without the service.
  • profile:<project>: an attached project has no profile. Write it yourself as your contract's Project memory section describes.

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.