More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+22 more

    Shipgate

    1

    Ship from laptop to production in the right order: full gate, green CI, database before code, a human on every irreversible step, then verified live.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Shipgate

    Shipgate

    Example session with this skill installed

    I've finished the new auth flow locally. Ship this to production.

    • Read your context and instructions
    • Compiled the shipgate

    Read SHIP.md first: this repo ships through a PR, never a push to main, and production steps are run by you. The auth flow is on its own worktree and branch, auth-flow.

    StepStatusNotes
    Ship policyReadPR to main, production steps by the owner
    Local gatePasstsc --noEmit, lint, tests, pnpm build with a frozen lockfile
    Staged diffCleanfile list reviewed, secret scan found nothing
    PR #42CI greenwaiting for your "merge"
    MigrationReady0042_user_metadata.sql applied to local and test, suite green
    ProductionYour turncommands below, database first

    After you say "merge": run these yourself, in order

    
# from the repo root, account: my-team
supabase db dump -f backup-$(date +%F).sql     # backup first
supabase db push --dry-run                      # lists 0042_user_metadata only
supabase db push
vercel deploy --prod --scope my-team

    Secrets stay in your terminal or host settings, never in this chat. Reply "done" when the deploy finishes.

    What I check after "done"

    • vercel inspect: the production URL points at the NEW deployment id
    • /api/health returns 200, /admin returns 401, not 500
    • the served JS chunk contains a string from this change
    • a read-only Playwright smoke test of sign-in on a phone viewport

    Then I record the new deployment id, the previous one as the rollback target, and the migration applied, in SHIP.md.

    Next steps

    1. Review PR #42 and reply "merge".
    2. Run the four commands above.
    3. Reply "done" for the live checks and the record.

    shot-19500

    Connects securely to your tools. The creator never sees your data.

    About this skill

    "Deploy to prod" is where agents do the most damage: a push to main, an app deployed before its migration, a secret pasted into chat, or a green dashboard that is still serving the old build. Ask your agent to "ship this" and Shipgate moves the change through every gate in order, hands the irreversible production steps to you as copy-ready commands, then proves the new version is live and records how to roll it back.

    Why it's different

    • A human on every irreversible step: production database push, production deploy, secrets and production SQL come to you as exact commands with the working directory and account scope. Secrets go into your own terminal, never into chat.
    • Database before code: migrations are committed, dry-run and applied before the deploy, with a backup first.
    • Verified, not assumed: confirms the URL serves the new deployment, checks status codes, greps the served bundle for this change and runs a read-only smoke test.
    • Built from real projects: 23 failures that actually happened, each with its guard.

    What it does

    • Ship policy first: reads the repo's own rules (local-only, never push main, auto-push) and defaults to local-only when there are none.
    • Local gate: typecheck, lint, tests and a production build with a frozen lockfile.
    • PR with green CI: conventional commit, PR, CI watched until green, merge only on your explicit "merge".
    • Go-live and lockdown: route audit for missing auth, row-level security, deployment protection, firewall and rate limits on first launch.
    • Record and learn: deployment id and rollback target written down, and a blameless post-mortem only when something broke and you say yes.

    How it works

    1. Read the ship policy and work in one worktree per task.
    2. Run the local gate and stage on purpose, with a secret scan of the diff.
    3. Open the PR and wait for green CI.
    4. Migrate, then deploy: gated steps handed to you as copy-ready commands or one go-live script.
    5. Verify live, lock down, record the deployment id and rollback target.

    Frameworks & tools

    Git, GitHub CLI and Actions, any Node package manager, Vercel or Fly.io, Supabase or another Postgres with migrations, curl and Playwright for the live checks.

    Use cases

    • Shipping a finished change from local to production without skipping steps.
    • Getting a schema migration to production safely.
    • First go-live of an app on a new host.
    • Working out why a deploy failed or the site went down right after a release.

    Known limitations

    It does not run irreversible production commands for you; you paste them. Not for repos under a deploy freeze or client repos you may only pull from.

    How to use

    Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scanned

    Verified clean today

    • Passed all security checks, Safe to install

    Requires

    Git

    Listed3 days ago
    Updatedtoday

    What's inside

    Frequently Asked Questions