agentleFS
joelbarmettlerUZHjoelbarmettlerUZH/ultrainit.sh/.claude/skills/release-ultrainit/SKILL.md

release-ultrainit

Skill8 starsChanged 7 months agoApache-2.0View on GitHub

Decide the next semver: v<MAJOR>.<MINOR>.<PATCH>. Pre-release format: v1.2.3-beta.1 (the .digit after the label is required — v1.2.3-beta alone fails release.yml's regex check). All changes must be merged to main before tagging. Releases are always cut from main. Do not proceed if any test fails. Integration tests: check raw output — CI shows green even when they fail. If any file contains a line matching __EOF_LIB_<name>__, the heredoc in bundle.sh terminates early, shipping a broken bundle silently. If bash -n fails,…

  • Commits and pushes

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All changes must be merged to main before tagging.

    line 36
  2. Distinctive:

    Do not proceed if any test fails.

    line 45
  3. Very specific:

    If any file contains a line matching __EOF_LIB_<name>__, the heredoc in bundle.sh terminates early, shipping a broken bundle silently.

    line 53
  4. Very specific:

    If bash -n fails, bundle.sh has a syntax error or delimiter conflict.

    line 65
  5. Very specific:

    Verify dist/ultrainit.sh is attached as a release asset.

    line 98

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Before You Start

  • .github/workflows/release.yml — the release pipeline (validates tag, builds bundle, publishes)
  • bundle.sh — what gets bundled and the __EOF_LIB_<name>__ delimiter strategy
  • Makefile — bundle and test-all targets
  • CLAUDE.md section on Security-Critical Areas — bundle.sh ships broken releases if wrong

Steps

1. Determine the next version

git tag --sort=-v:refname | head -5

Decide the next semver: v<MAJOR>.<MINOR>.<PATCH>. Pre-release format: v1.2.3-beta.1 (the .digit after the label is required — v1.2.3-beta alone fails release.yml's regex check).

2. Verify the branch is clean and on main

git status
git log --oneline main..HEAD

All changes must be merged to main before tagging. Releases are always cut from main.

3. Run the full test suite

make check     # syntax + schema validation
make test-all  # full bats suite inside Docker

Do not proceed if any test fails. Integration tests: check raw output — CI shows green even when they fail.

4. Check for bundle delimiter conflicts

grep '__EOF_LIB_' lib/*.sh prompts/*.md scripts/*.sh

If any file contains a line matching __EOF_LIB_<name>__, the heredoc in bundle.sh terminates early, shipping a broken bundle silently.

5. Build and verify the bundle

mkdir -p dist
bash bundle.sh > dist/ultrainit.sh
chmod +x dist/ultrainit.sh
bash -n dist/ultrainit.sh && echo "syntax OK"
du -sh dist/ultrainit.sh

If bash -n fails, bundle.sh has a syntax error or delimiter conflict. Do not tag.

6. STOP — confirm the version with the user

Show the user:

  • The version number you plan to tag
  • git log --oneline $(git tag --sort=-v:refname | head -1)..HEAD (commits since last tag)
  • The bundle size from step 5

Wait for explicit approval before proceeding. The developer answers explicitly state: never release new tags without consulting the user first.

7. Create and push the tag

git tag v<VERSION>
git push origin v<VERSION>

This triggers .github/workflows/release.yml automatically.

8. Monitor the release workflow

gh run list --workflow=release.yml
gh run watch  # follow the live output

9. Verify the published release

gh release view v<VERSION>

Verify dist/ultrainit.sh is attached as a release asset. Check the download URL in the release body matches the curl-pipe-bash install command in README.md.

Verify

bash -n dist/ultrainit.sh                          # bundle syntax
gh release view v<VERSION> --json assets --jq '.assets[].name'  # asset attached

Common Mistakes

  1. Tagging without consulting the user — the project rule is explicit. Always show the planned tag and commit range before tagging.
  2. v1.2.3-beta without .digit — release.yml validates against ^v[0-9]+\.[0-9]+\.[0-9]+(-[a-zA-Z0-9.]+)?$. The suffix must be beta.1 not beta.
  3. Skipping make test-all before tagging — release.yml runs bash -n but does NOT run the bats test suite. A broken release can ship even with failing tests if you skip step 3.
  4. bundle.sh stdout contamination — any echo or log_* call added to bundle.sh corrupts the bundle output because release.yml does bash bundle.sh > dist/ultrainit.sh.

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.