agentleFS
TakiTakeTakiTake/pall8t/.claude/skills/release/SKILL.md

release

Skill4 starsChanged 49 days agoMITView on GitHub

Cut and publish a new pall8t release end to end — version bump, changelog, tag, GitHub Release verification, and Homebrew tap update. Use when the user asks to cut, ship, or publish a release, or wants the next version released. Takes the new version (e.g. 0.2.0) as an argument.

  • Reads credentials
  • Installs packages
  • Commits and pushes

193 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use create-agent-adapter, changelog-style 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.
  • MicrosoftSkill

    create-agent-adapter

    Technical guide for creating a new Paperclip agent adapter. Use when building a new adapter package, adding support for a new AI coding tool (e.g. a new CLI agent, API-based agent, or custom process), or when modifying the adapter system. Covers the required interfaces, module structure, registration points, and conventions derived from the existing claude-local and codex-local adapters.

    Also in 5 projects that use this

  • GrinvSkill

    changelog-style

    Turn a commit range into CHANGELOG.md entries following common-changelog style. Use when finishing a PR or cutting a release, to write/fix the [Unreleased] section's entries — moving [Unreleased] under a new dated version heading is a separate, automated step (see the release skill / scripts/sync-version.ts), not this skill's job.

    Also in 4 projects that use this

  • Microsoft AzureSkill

    component-release

    Prepares a verifiable Agent Landing Zone UI release. Use for release branches, VERSION, changelog entries, tags, release notes, and compatibility evidence; never publish without explicit human approval.

    Also in 4 projects that use this

  • GrinvSkill

    docs-consistency-check

    Check README/manifest.json/server.json/CHANGELOG.md/AGENTS.md and docs/*.md for drift against the actual registered tools and source. Use after adding, renaming, or removing a tool, or as part of a live-audit pass.

    Also in 4 projects that use this

  • Microsoft AzureSkill

    documentation-consistency

    Keeps AI Landing Zone contributor, user, and operator documentation aligned with shipped Bicep behavior. Use for parameters, defaults, outputs, topology, deployment flow, operations, or releases.

    Also in 8 projects that use this

  • parcadeiSkill

    agent-context-isolation

    Agent Context Isolation

    Also in 3 projects that use this

  • +3 more from Grinv

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 1 is distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 183 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Very specific:

    Run /bump $VERSION — it owns this step: version resolution (including what a pre-1.0 breaking change means for the number), preflight, the …

    line 17
  2. Very specific:

    After /bump opens the PR, run /code-review on it and fix what it finds.

    line 30
  3. Very specific:

    Do not merge the PR: If it turns out to be merged already when this step runs — /release invoked after the human merged the bump — say what …

    line 42
  4. Distinctive:

    Only proceed here once step 3 has confirmed a real Release with both assets verified — don't run this step against a release that failed or …

    line 91
  5. Very specific:

    Update Formula/pall8t.rb: url to the new release tarball, sha256 to the verified checksum from step 3.

    line 96
  6. Very specific:

    Commit locally — git add and git commit in one call, with no push in the same call.

    line 98

“Common” is measured: kept by at least 6 in 10 of the 183 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

Cutting a pall8t release

Full narrative in docs/release.md; this is the operational, step-by-step version with who does what. Steps alternate between the agent and the user — several publish-only actions are gated to the user by the permission system (see the repo's root CLAUDE.md). Don't retry a denial or work around it — hand the exact command to the user instead.

1. (agent) Preflight and release-prep PR

Run /bump $VERSION — it owns this step: version resolution (including what a pre-1.0 breaking change means for the number), preflight, the Cargo.toml/Cargo.lock/CHANGELOG.md edits, a check that release.yml's extractor will actually find the notes, the quality gates, and the PR.

Kept there rather than repeated here so the two can't drift — the changelog heading format is parsed verbatim by release.yml, and one description of it is enough.

Two things this skill adds on top:

  • Decide $VERSION from this skill's argument if given, otherwise ask before invoking /bump.
  • After /bump opens the PR, run /code-review on it and fix what it finds. /bump runs local-review on the diff it creates; every revision you push on top of that is a diff nobody has reviewed, so run that again before each push, per CLAUDE.md. Findings from a reviewer on the PR itself go through the review-loop skill like any other.
  • The bots do not review this PR unprompted. CodeRabbit's automatic review is off (.coderabbit.yaml), so it reviews only when someone comments @coderabbitai review — the human's call, per CLAUDE.md — and the Codex workflow, while it does run and report green, skips every step until an OPENAI_API_KEY secret exists. Waiting for either review to arrive on its own is waiting forever.

Do not merge the PR. If it turns out to be merged already when this step runs — /release invoked after the human merged the bump — say what review the diff did and didn't get, and carry on from step 2 rather than reopening it.

2. (user) Merge and tag

  • The user merges the PR themselves — PR merges are gated for agents.
  • (agent) Before handing over the tag command, pull latest main and confirm Cargo.toml's version and the CHANGELOG.md heading both read exactly $VERSION — release.yml only checks the tag against Cargo.toml after the tag is already public, so catching a mismatch here avoids pushing a tag that's guaranteed to fail downstream (see "If the release workflow fails" below).
  • Then hand over the tag command:
    ! git checkout main && git pull
    ! git tag v$VERSION && git push origin v$VERSION
    Tag pushes are gated too (they trigger the public release workflow) — hand this exact command over rather than attempting it.

3. (agent) Watch the release workflow and verify the artifact

  • gh run watch <run-id> (or poll gh run list --workflow=release.yml --limit 1) until .github/workflows/release.yml finishes.
  • gh release view v$VERSION — confirm both assets exist: pall8t-v$VERSION-aarch64-apple-darwin.tar.gz and its .sha256.
  • Download both into the scratchpad and verify:
    • sha256sum -c the tarball against its .sha256 file — this is the authoritative check. gh release view --json assets may also report a digest field to cross-check against, but treat that as a bonus, not a required step: it's absent/null on some gh versions.
    • tar -tzf shows a single top-level dir pall8t-v$VERSION-aarch64-apple-darwin/ containing pall8t + LICENSE
      • README.md — that single-dir layout is what lets Homebrew cd into it.
    • The binary's magic bytes are Mach-O ARM64: od -A x -t x1z -v pall8t | head -1 should start cf fa ed fe with cputype 0c 00 00 01 (0x0100000c = ARM64).

If the release workflow fails (step 3 never finds a Release, or gh run watch reports a failed run): stop — do not proceed to step 4. The tag is already public at this point; deleting and reusing it requires git push --delete origin v$VERSION (gated, hand it to the user), fixing whatever failed on main in a new PR, then retagging. Report the failure and the fix needed instead of guessing.

4. (agent, then possibly user) Update the Homebrew tap

Only proceed here once step 3 has confirmed a real Release with both assets verified — don't run this step against a release that failed or doesn't exist yet.

  • Clone https://github.com/TakiTake/homebrew-tap.git into the scratchpad.
  • Update Formula/pall8t.rb: url to the new release tarball, sha256 to the verified checksum from step 3.
  • Commit locally — git add and git commit in one call, with no push in the same call. A denied compound command rolls back the whole thing, including the local-only parts, so keep the push isolated.
  • Attempt git push origin main. Pushes to homebrew-tap are gated — don't retry on denial, report the exact push command for the user to run from the scratchpad clone instead.
  • Once pushed, verify the live formula: fetch https://raw.githubusercontent.com/TakiTake/homebrew-tap/main/Formula/pall8t.rb and confirm it's byte-identical to what was committed, with the right url/sha256.

5. (user, optional) Smoke test

! brew upgrade TakiTake/tap/pall8t || brew install TakiTake/tap/pall8t
! pall8t --version

This can't be run from a Linux dev container — it's the one step only the user can actually execute end to end.

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.