Skip to content

ddtcorex/maestro-skills / skills/govard-toolbox

govard-toolbox

Flagged

Use when starting or stopping a Govard environment, running container commands, database dump or import, syncing from staging, Xdebug, or a deploy and its rehearsal. Keywords govard up, govard sh, govard deploy.

FileKindSize
SKILL.mdskill md10.4 KB
references/AUDIT.mdmarkdown4.6 KB
references/COMMANDS.mdmarkdown7.3 KB
references/CONFIGURATION.mdmarkdown2.2 KB
references/DEPLOYMENT.mdmarkdown4.8 KB
references/FAQ.mdmarkdown3.5 KB
references/GUIDES.mdmarkdown2.8 KB
references/HOST-WITHOUT-DOCKER.mdmarkdown1.8 KB
references/SANDBOX.mdmarkdown5.0 KB
references/VERIFY.mdmarkdown23.8 KB

references/DEPLOYMENT.md

View on GitHub
# Deployment

The plan/apply split, a rehearsal against a container playing the target, the rollback and
the bootstrap into a live remote. The safety rule is stated in `SKILL.md`.

Ships a git revision to a remote from `.govard.yml`.

A deploy, a rollback, `--with-db` and a bootstrap into a live remote change a running
site. State the remote, the exact command and the way back, then wait for an explicit
yes before running one; never pass `--yes` to approve a production target on the
user's behalf, and never chain a plan and an apply on one line. Read-only first,
always:

```bash
govard deploy plan production          # every task with what it runs, or why this mode skips it (no ssh, no rsync, no Docker)
govard deploy plan production --json   # same plan machine-readable (carries schema_version, per-step skip_reason, run_on, needs_migration); timestamp-free, so two runs can be diffed in CI. `check` has no --json
govard deploy check production         # connectivity, permissions, layout, and the strategy that layout implies (ssh)
govard deploy releases production      # what the target has (a positional remote or --remote <name>)
govard deploy status production        # which revision the target serves (no remote: every configured remote)
govard deploy production               # ask the human first: this is a production change
govard deploy rollback production [--to <release>] [--with-db]   # default: the release before the live one; --with-db also restores its dump (destroys data)
govard deploy production --resume      # continue the newest unfinished release (or --from <task>)
govard deploy unlock production [--force]       # lock left by an interrupted run; a lock younger than lock_stale_after needs --force
```

Capabilities (confirm with `govard capabilities`): `plan`/`build` need nothing; `check`, `releases`, `status` and `unlock` need `ssh`; `deploy`/`rollback` also `rsync`; `sandbox *` needs Docker. A missing runtime is exit `3` `CAPABILITY_MISSING`, never a half-run. Recover a failed run with `--resume` (or `--from <task>`), never by unlocking and starting over. Resume adopts a recorded *migrate* verdict without re-probing; a recorded *skip* is discarded and re-probed; steps that already succeeded are not repeated.

**Conditional migrate (Magento only).** Before the downtime block the recipe
probes `cd {{release_path}} && {{php_bin}} bin/magento setup:db:status`: exit 0
skips the whole downtime block (maintenance mode, worker pause, config import,
db migrate, worker resume; reason `db up-to-date (probe exit 0)`); 1 or 2 runs it;
any other exit fails the deploy. Compile, cache flush and the DB backup run
regardless. Laravel, Symfony and WordPress have no probe, so their migrate steps
always run. `govard deploy plan` shows which steps are gated (`needs_migration`).

**Topology.** Project-wide defaults live in a `deploy:` block (`repository`, `branch`, `publish`, `deploy_path`); per-remote overrides live ONLY under `remotes.<name>.deploy:`. Flat per-remote deploy keys are rejected by the loader (exit 4) with a message naming the key to move. Unset `deploy_path` probes the target and is adopted only when exactly one layout candidate matches.

What runs on the target comes from the framework **recipe**; `--build=auto|server|artifact` decides who builds, and `govard deploy build --output <dir>` makes the artifact. `--db-backup` defaults off; when on, the dump lands in `shared/backups/deploy/<n>/` and `rollback --with-db` restores it. A framework whose recipe has no dump command refuses `--db-backup` up front (exit 4); see the framework skill.

```bash
# Rehearse the whole thing against a container playing the target
govard sandbox up --profile full --php 8.3 --docroot symlink   # profile, PHP series, target shape
govard deploy --remote sandbox --yes
govard sandbox down --purge          # also removes the image, key and mirror
```

Sandbox lists come from the recipe; `deploy.settings.sandbox_{packages,extensions,services,tools}` **replace** them. `sandbox reset` wipes the deploy directories (including `shared/`), so re-seed shared files. Details: [SANDBOX.md](references/SANDBOX.md).

> **Deploy preflight:** `govard deploy plan [remote]` is read-only and never connects (`--json` for machine-readable output); `govard deploy check [remote]` connects over ssh and is meant to leave nothing behind on the target (its layout probe uses a scratch directory it removes again). Both accept `--build`, `--artifact-dir`, `--branch`, `--revision` and `--tag`; `check` prints human text only. `plan` reflects the configuration, not `--publish`, `--keep` or `--verify`, which only take effect on a real run. Run them on the host.

Per-framework detail (recipes, the conditional-migrate probe, and each
framework's own step table): `govard-magento`, `govard-laravel`,
`govard-symfony`, `govard-wordpress`. Full reference:
<https://govard.ddtcorex.com/workflows/deployment>.
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 322kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall