agentleFS
SantiagoVqzSantiagoVqz/claude-skills/skills/4-ship/ship/SKILL.md

ship

Skill2 starsChanged 12 days agoMITView on GitHub

Open the PR for a built branch: rebase on the trunk, check the hard stops, run one cold read of the whole diff, write the body, open the PR, and wait for CI to pass.

  • Deletes or force-pushes
  • Commits and pushes

194 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ios-sync, open-gstack-browser and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Saqlain-811Skill

    ios-sync

    Regenerate the iOS debug bridge against the latest upstream gstack templates. (gstack)

    Also in 6 projects that use this

  • seek888Skill

    open-gstack-browser

    Launch GStack Browser — AI-controlled Chromium with the sidebar extension baked in. Opens a visible browser window where you can watch every action in real time. The sidebar shows a live activity feed and chat. Anti-bot stealth built in. Use when asked to "open gstack browser", "launch browser", "connect chrome", "open chrome", "real browser", "launch chrome", "side panel", or "control my browser". Voice triggers (speech-to-text aliases): "show me the browser".

    Also in 10 projects that use this

  • ayudaySkill

    setup-deploy

    Configure deployment settings for /land-and-deploy. Detects your deploy platform (Fly.io, Render, Vercel, Netlify, Heroku, GitHub Actions, custom), production URL, health check endpoints, and deploy status commands. Writes the configuration to CLAUDE.md so all future deploys are automatic. Use when: "setup deploy", "configure deployment", "set up land-and-deploy", "how do I deploy with gstack", "add deploy config".

    Also in 19 projects that use this

  • ShreyRaviSkill

    landing-report

    Read-only queue dashboard for workspace-aware ship. Shows which VERSION slots are currently claimed by open PRs, which sibling Conductor workspaces have WIP work likely to ship soon, and what slot /ship would pick next. No mutations — just a snapshot. Use when asked to "landing report", "what's in the queue", "show me open PRs", or "which version do I claim next". (gstack)

    Also in 9 projects that use this

  • zebbernSkill

    debug-fix

    Find and fix a bug or issue — from any source (GitHub issue, error message, user report, or observed behavior)

    Also in 4 projects that use this

  • ramybarsoumSkill

    document-release

    Post-ship documentation update. Reads all project docs, cross-references the diff, updates README/ARCHITECTURE/CONTRIBUTING/CLAUDE.md to match what shipped, polishes CHANGELOG voice, cleans up TODOS, and optionally bumps VERSION. Use when asked to "update the docs", "sync documentation", or "post-ship docs". Proactively suggest after a PR is merged or code is shipped.

    Also in 23 projects that use this

The file

We split this skill into the rules it gives an agent

8 rules. 0 are common, 2 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 181 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

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

    Read the ## Agent skills block in AGENTS.md, then docs/agents/tracker.md.

    line 16
  2. Very specific:

    Run git fetch origin, then git rebase origin/<trunk>.

    line 26
  3. Very specific:

    Run hardstop-check.sh origin/<trunk>, from the folder of this skill, in the root of the branch.

    line 32
  4. Distinctive:

    When it refuses, show its lines, and ask with the question tool: "Override", with a reason that goes in the PR body, or "Stop".

    line 34
  5. Very specific:

    Call the Skill tool with pr to write the body.

    line 56
  6. Very specific:

    Run git push --force-with-lease origin <branch>.

    line 67
  7. Very specific:

    Run gh pr checks <pr> --watch in the background, and wait for it to end.

    line 75
  8. Distinctive:

    Start with a short summary, three lines or fewer.

    line 84

“Common” is measured: kept by at least 6 in 10 of the 181 edited versions of this file we could read. “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

Ship

/ship turns a built branch into one PR that the team can read. It runs the one cold read of the whole diff and the one full check of the code. It works on any branch with commits, with no earlier skill needed. It ends with the PR open, CI finished, and a short summary for the user. The merge is always the user's, with the method that step 1 finds.

In a worktree, follow the shell rules of the build skill.

1. Load

Read the ## Agent skills block in AGENTS.md, then docs/agents/tracker.md. When the block is missing, read NO-SETUP.md in the folder of the setup-skills skill, and go on with its sources.

The target is the argument, or the number in the current branch name. A branch with no number has no issue. Go to the worktree that has the branch, when one does. When git status shows changes, stop and ask whether they belong in this PR.

Find the merge method. Search docs/adr/ for "merge commit", and read gh repo view --json rebaseMergeAllowed,mergeCommitAllowed. An ADR that needs merge commits wins. Otherwise use rebase and merge when the repo allows it, and a merge commit when it does not.

Completion: you have the branch, its issue when it has one, the trunk, as the worktree skill finds it, and the merge method.

2. Rebase

Run git fetch origin, then git rebase origin/<trunk>. On a conflict, call the Skill tool with resolve-conflicts.

Completion: the branch is on the latest trunk.

3. Hard stops

Run hardstop-check.sh origin/<trunk>, from the folder of this skill, in the root of the branch. It refuses a diff that holds a migration, a backfill, or auth or permission code when a commit says Hard stops: none.

When it refuses, show its lines, and ask with the question tool: "Override", with a reason that goes in the PR body, or "Stop".

Completion: the check passed, or the user gave the override reason.

4. Cold read

Start a subagent with a fresh context. Tell it to call the Skill tool with verify, with the issue, its criteria, and the diff origin/<trunk>...HEAD. Give it the shell rules, and no list of checks to skip. It needs only the conclusions, so keep the diff out of this context.

Fix each hard finding in this session, squash the fix into the commit, and run a new subagent. Run three cold reads or fewer. When a hard finding remains, stop and show it. The user fixes it, or overrides it with a reason that goes in the PR body. Judgment calls go in the report, and change nothing.

Completion: the last cold read has no hard finding, or the user gave each override reason.

5. Full checks

A workflow file is no proof that CI runs on this PR. Some repos have workflows that run only on main while the trunk is develop. Read the on: block of each file that git ls-files .github/workflows lists. CI covers the PR when a workflow has pull_request or pull_request_target, and its branches is missing or matches <trunk>, and its branches-ignore does not match <trunk>.

When CI covers the PR, skip the rest of this step. Otherwise start a subagent to run the full suite, the typecheck, the lint and the build, once, under the shell rules. Wait for its result, so that no check still runs when the user merges. When the repo has no script for a check, skip that check and name it in the report. A failure in the code of the branch stops /ship. Show its output.

Completion: CI covers the PR, or every check passed.

6. Write the body

Call the Skill tool with pr to write the body. Give it:

  • The title: the title of the issue, or the subject of the commit.
  • The evidence: the <criterion> · <test> lines of the commit message, the result of the cold read and each override reason. For a bug, the regression test.
  • The closing line: Closes #<n>, or Refs #<n> for a bug whose cause is not confirmed. A branch with no issue has no closing line.
  • The merge line: Merge with <merge method>.

Completion: you have the title and a body with each of these parts.

7. Open

Run git push --force-with-lease origin <branch>. Write the body to a file in the worktree root. Run gh pr create --base <trunk> --head <branch> --title "<title>" --body-file <file>. When the branch already has an open PR, run gh pr edit <pr> --body-file <file> in its place. Then delete the file.

Completion: the PR is open, ready for review, and git status is clean.

8. Watch CI

Skip this step when step 5 found that CI does not cover the PR.

Run gh pr checks <pr> --watch in the background, and wait for it to end. It can exit non-zero when every check passed, so read the result with gh pr checks <pr> each time it ends.

  • A check failed: show the job name and the end of gh run view <run> --log-failed. Fix nothing, because the user decides how to fix it.
  • No check was reported: CI did not start, for example because of a paths filter. Run the checks of step 5, and name the gap in the report.

Completion: each check passed, or you showed each failure, or the checks of step 5 ran because CI did not start.

Report

Start with a short summary, three lines or fewer. Do not show the PR body.

- <what changed> · because <why, from the issue's problem or the bug's cause>

Then: PR URL · hard-stop check · cold read result, with each judgment call · CI result, or no CI on PRs to <trunk> with where the checks ran · next step: merge with <merge method>, then /cleanup.

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.