agentleFS
oborchersoborchers/proqi/.agents/skills/release/SKILL.md

release

Skill45 starsChanged 52 days agoMITView on GitHub

Prepare, verify, publish, or diagnose a Proqi release through the repository's protected tag workflow. Use only when explicitly invoked for this repository.

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

  • 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

    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

  • 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

  • rivet-devSkill

    launch-graphics

    Generate agentOS marketing/launch graphics (launch & social heroes, code-snippet cards) from the committed generators in docs-internal/graphics. Use when asked to make, update, or render an agentOS launch image, social image, "agentOS Apps" graphic, or technical code-snippet image.

    Also in 3 projects that use this

  • +3 more from Grinv

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 9 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 183 edited versions we read keep these.

  • 9

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

    Never advertise, install, or publish this skill as an end-user Proqi skill.

    line 9
  2. Very specific:

    Read the repository instructions and docs/RELEASING.md completely.

    line 13
  3. Very specific:

    Verify the branch, HEAD, remotes, tags, and complete worktree state.

    line 15
  4. Distinctive:

    Read references/release-notes.md before drafting or reviewing release notes.

    line 20
  5. Distinctive:

    Do not infer a shipped capability from a commit title alone.

    line 27
  6. Very specific:

    Update the workspace version in Cargo.toml, refresh Cargo.lock through normal Cargo tooling, create .github/release-notes/vX.Y.Z.md, and …

    line 32
  7. Very specific:

    Run focused checks appropriate to the changed code.

    line 44
  8. Distinctive:

    Never draft highlights from commit titles alone.

    line 54
  9. Distinctive:

    Before asking for publication confirmation:

    line 61
  10. Distinctive:

    Always stop immediately before creating or pushing the release tag.

    line 73
  11. Distinctive:

    Create and push one annotated stable tag.

    line 92
  12. Distinctive:

    If promotion fails, inspect the exact failing step.

    line 98
  13. Very specific:

    Verify the public GitHub Release, exact crates.io version, Homebrew formula, and installed proqi --version.

    line 102
  14. Distinctive:

    If crates.io or any immutable public artifact exists, recover with a new semantic version.

    line 109
  15. Distinctive:

    Do not merge dependency pull requests, alter protection rules, change repository settings, or expand distribution channels as a side effect …

    line 114

“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

Proqi Release

This is a repository-local maintainer skill. Act only after explicit invocation. Never advertise, install, or publish this skill as an end-user Proqi skill.

Begin safely

  1. Read the repository instructions and docs/RELEASING.md completely. The runbook and checked-in workflows are authoritative over this skill.
  2. Verify the branch, HEAD, remotes, tags, and complete worktree state. Never discard, hide, or publish unexpected changes.
  3. Determine whether the user asked to prepare a release, inspect release status, diagnose a failure, or proceed toward publication. Preparing files, committing them, or pushing main does not authorize a tag or release.
  4. Read references/release-notes.md before drafting or reviewing release notes.

Prepare

  1. Derive the change set from the previous stable tag through HEAD. Inspect commits, user-visible diffs, tests, documentation, and known limitations. Do not infer a shipped capability from a commit title alone.
  2. Recommend the next semantic version with reasons. Before 1.0, use a minor increase for a new capability or intentional compatibility break, and a patch increase for compatible fixes and operational hardening. The user owns the final version choice.
  3. Update the workspace version in Cargo.toml, refresh Cargo.lock through normal Cargo tooling, create .github/release-notes/vX.Y.Z.md, and add or update one exact matching entry with three to six concise user-facing items in release-highlights.json. GitHub notes remain the only changelog. The manifest is the bounded in-product projection of those reviewed notes. Set the version in herdr-plugin.toml and the proqi plugin version in .claude-plugin/marketplace.json to the same version so Claude Code announces the updated skills; release-plan rejects a Herdr plugin version that differs from Cargo. Replace every Next release and next-release scope marker in README.md and docs/ with the new version, and rewrite version-free wording such as the Herdr plugin's "first Proqi release that includes proqi herdr toggle" to name it.
  4. Run focused checks appropriate to the changed code. Routine release preparation then runs the cheap cargo xtask release-plan vX.Y.Z contract and Git diff hygiene. Run workflow lint only when workflows changed. Full check, PTY, coverage, audit, package, crate dry run, full MSRV, rehearsal, and Linux parity remain milestone or diagnostic gates as documented in docs/RELEASING.md; do not repeat them merely because metadata is prepared.
  5. Review generated or changed release artifacts explicitly. Never weaken a gate, accept snapshots automatically, or claim an unavailable platform check passed.
  6. Review the exact GitHub notes and matching packaged highlights together. Never draft highlights from commit titles alone. Show both texts, the version diff, gate results, and any limitations before committing or pushing release preparation unless the current instruction already authorizes those exact local and remote changes.

Pre-publication checks

Before asking for publication confirmation:

  1. Require the release commit on main, its required CI result and immutable candidate workflow to be green, and the exact vX.Y.Z tag to match the Cargo version, notes filename, and reviewed highlights.
  2. Confirm that the exact tag is absent locally and remotely and that no conflicting GitHub Release or crates.io version exists.
  3. Record the full release commit SHA and verify that the worktree contains no unreviewed release input.

Mandatory publication confirmation

Always stop immediately before creating or pushing the release tag. Ask the user to confirm one exact action in this form:

Create and push the annotated tag vX.Y.Z at <full-commit-sha>, triggering the
public GitHub Release, crates.io publication, and Homebrew tap update?

This confirmation is mandatory even if the user previously asked to prepare, finish, cut, or publish a release, approved the release notes, or approved the release preparation commit. Do not infer it from earlier authority. Do not create a local tag, push a tag, call a release workflow, or create a GitHub Release before the user answers affirmatively to the exact tag and commit.

Publish after confirmation

  1. Revalidate that the commit, CI result, version, notes, and tag state have not changed since confirmation. If any changed, stop and ask again with the new exact tag and commit. Never move, recreate, or force a public release tag.
  2. Create and push one annotated stable tag. Do not create or publish a GitHub Release manually. The protected tag is the workflow's sole publication trigger.
  3. Monitor the tag-triggered Release workflow. It selects the exact prior main candidate without rebuilding native binaries, publishes crates.io through OIDC, publishes and verifies GitHub assets, then notifies the Homebrew tap.
  4. If promotion fails, inspect the exact failing step. Rerun the same promotion when safe. If the candidate is missing or expired, manually run the non-publishing candidate workflow for the exact main SHA or protected tag, then retry promotion. Never publish around a failed verification.
  5. Verify the public GitHub Release, exact crates.io version, Homebrew formula, and installed proqi --version. Report each channel independently.

Recovery boundaries

  • The manual Release candidate workflow is diagnostic or recovery tooling. It never publishes and is not part of an ordinary release.
  • If crates.io or any immutable public artifact exists, recover with a new semantic version. Never overwrite, delete, yank, or retag merely to make the workflow green.
  • A failed or ambiguous external operation is not success. Preserve the command output and current public state before proposing a retry.
  • Do not merge dependency pull requests, alter protection rules, change repository settings, or expand distribution channels as a side effect of a release.

Finish with the released version, commit and tag, workflow URL and result, GitHub Release URL, crates.io result, Homebrew result, installed version, and any skipped or failed verification.

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.