agentleFS
redhat-developerredhat-developer/rhdh-skills/skills/ci/rhdh-yarn-bump/SKILL.md

rhdh-yarn-bump

Skill20 starsChanged 21 days agoApache-2.0View on GitHub

Bumps Yarn Berry across the RHDH repos — rhdh-plugins, rhdh midstream, rhdh-plugin-export-overlays, rhdh-cli, and GitLab CEE rhidp/rhdh and rhdh-plugin-catalog — with `yarn set version` plus install, and rewrites the pins Yarn cannot see: `packageManager`, `yarnPath`, `ENV YARN=`, and Containerfile lines. Use for "bump yarn to 4.17.1", "upgrade Yarn Berry across the repos", "weekly yarn bump", "which Yarn version is each repo pinned to", or scanning yarn pins.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 0 are distinctive, and 6 are 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.

  • 6

    Very specific

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

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

    When package.json defines prettier:check or prettier:fix, yarn-bump.sh runs yarn prettier --ignore-unknown --write on the changed files …

    line 41
  2. Very specific:

    Do not re-enable Renovate Yarn bumps.

    line 85
  3. Very specific:

    Do not yarn set version stable — repos must share one exact 4.x.

    line 86
  4. Very specific:

    Do not download a separate Yarn binary per repo.

    line 87
  5. Very specific:

    Do not special-case rhdh-1-rhel-9.

    line 88
  6. Very specific:

    Do not invent a model in --attribution.

    line 89

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

RHDH multi-repo Yarn bump

Goal

Propagate a Yarn Berry 4.x bump across:

RepoNotes
redhat-developer/rhdh-pluginsroot workspace (+ Fullsend if hardcoded)
redhat-developer/rhdhroot + nested workspaces + Containerfile
redhat-developer/rhdh-plugin-export-overlaysmany packageManager pins
redhat-developer/rhdh-cliroot packageManager / yarnPath
gitlab.cee.redhat.com/rhidp/rhdhdistgit binary + ENV YARN= (fetched CLI, same bytes as GitHub)
gitlab.cee.redhat.com/rhidp/rhdh-plugin-catalogper-workspace pins + Containerfiles

Renovate Yarn packageManager updates are disabled (RHIDP-17563). This skill is the bump path.

Branch: live line is main on GitHub and GitLab CEE.

What actually changes

yarn set version <to>                 # packageManager + yarnPath + .yarn/releases
chmod +x .yarn/releases/yarn-<to>.cjs
yarn install --mode=update-lockfile

When package.json defines prettier:check or prettier:fix, yarn-bump.sh runs yarn prettier --ignore-unknown --write on the changed files before commit. Yarn writes "**" in .yarnrc.yml; Prettier (rhdh-cli) requires '**'.

Plus pins Yarn cannot see: ENV YARN=, Containerfile / Dockerfile yarn_version= and literal yarn set version (including versions not in --from; denylist pins stay). yarn set version $yarn_version is left as a variable.

yarn-bump.sh runs curl -fsSL once against https://repo.yarnpkg.com/<to>/packages/yarnpkg-cli/bin/yarn.js (the same URL yarn set version uses) and installs that file into every bumped .yarn/releases directory. GitHub PRs and GitLab MRs commit yarn-<to>.cjs (mode 100755) plus the removed older binary. If the download fails, GitLab falls back to copying a binary produced by a GitHub bump.

Default path: yarn-bump.sh

SKILL=<this skill's directory>
JIRA_ISSUE=RHIDP-1234 "$SKILL/scripts/yarn-bump.sh"   # resolve latest 4.x, clone, bump, PR/MR
"$SKILL/scripts/yarn-bump.sh" --to 4.18.1 --issue RHIDP-1234 --dry-run

--issue / JIRA_ISSUE is required: the opener (/rhdh-pr-mr) refuses to publish without a key, so the scheduled bump carries the standing automation ticket. The key goes in the Related Jira: line; it does not mean a human is working the bump.

--attribution / RHDH_ATTRIBUTION (alias --assisted-by, RHDH_ASSISTED_BY) becomes the commit trailer and is forwarded to the opener. Default rhdh-yarn-bump — the script names itself and claims no model. An agent-driven run should pass what it knows, e.g. --attribution 'paseo (pi + DeepSeek V4.1 Flash (EnMaaS), thinking: off)'. A value that already starts with Assisted-By: or Generated-by: is used verbatim.

It resolves Yarn 4.x (@yarnpkg/cli; never stable / 5.x unless --to), downloads that CLI once, clones the six repos, skips open chore/automated-yarn-bump* PRs/MRs and denylist pins (4.8.1 / 4.9.2 / 4.15.0), bumps each tree with --bin (so the release binary is in the diff), commits as rhdh-bot with the Assisted-By: trailer, and opens PRs/MRs via sibling /rhdh-pr-mr, forwarding --issue and --attribution.

GitLab weekly-maintenance clones this skill pack and runs yarn-bump.sh after digest/bootc and before Quay cleanup.

Manual mutator (existing checkouts)

node "$SKILL/scripts/bump-yarn.js" --scan --root /path/to/repo
node "$SKILL/scripts/bump-yarn.js" --to 4.18.1 --from-all --root /path/to/rhdh-plugins
curl -fsSL -o /tmp/yarn-4.18.1.cjs \
  https://repo.yarnpkg.com/4.18.1/packages/yarnpkg-cli/bin/yarn.js
node "$SKILL/scripts/bump-yarn.js" --to 4.18.1 --bin /tmp/yarn-4.18.1.cjs \
  --from-all --root /path/to/rhdh-downstream

--from 4.12.0,4.14.1 is the default when --from-all is omitted. Lock refresh can exceed 45 minutes.

Anti-patterns

  • Do not re-enable Renovate Yarn bumps; they miss Containerfile / ENV YARN=.
  • Do not yarn set version stable — repos must share one exact 4.x.
  • Do not download a separate Yarn binary per repo. One fetch from repo.yarnpkg.com, then --bin copies those bytes into every PR/MR. --copy-bin is only the fallback when that fetch fails.
  • Do not special-case rhdh-1-rhel-9; both CEE repos default to main.
  • Do not invent a model in --attribution. Unknown parts are dropped: the tool alone is honest, a guessed model is not.

Tests

node --test "$SKILL/tests/bump-yarn.test.mjs"
bash "$SKILL/tests/smoke-caller.sh"   # commit trailer + forwarded opener metadata

Completion

Done when matching workspaces sit on --to (packageManager, yarnPath, yarn-<to>.cjs mode 100755), extras rewritten, locks refreshed unless --no-refresh-locks, and PRs/MRs opened (or skipped for open bot PR / no diff).

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.