agentleFS
emiliodominguezemiliodominguez/flow-skills/skills/flow-release/SKILL.md

flow-release

Skill1 starsChanged 47 days agoMITView on GitHub

Versioned software releases and publication recovery. Use when preparing a version, publishing artifacts or resuming a partial release. Hands off to `flow-ship` or `flow-triage`.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the versioning policy, release mechanism, supported targets, pending changes, existing tags and published versions.

    line 15
  2. Distinctive:

    Verify that an existing tag resolves to the intended source commit, accounting for tag objects separately from their target commits.

    line 39
  3. Distinctive:

    Read back the tag, release metadata and artifact identities from the actual destinations.

    line 49
  4. Distinctive:

    Stop at the requested release stage.

    line 55
  5. Very specific:

    Use flow-ship for requested Git delivery or flow-triage for an observed release incident.

    line 71

≈ “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

Release

Make the requested version available with evidence tying its source, metadata and artifacts together. Git delivery belongs to flow-ship; package or release publication is a separate action.

Phase 1: Establish the release contract

  • Read the versioning policy, release mechanism, supported targets, pending changes, existing tags and published versions. Determine the authorized stage: prepare, tag or publish. Reuse current authorization; publication never implies deployment or announcements.
  • Select the version from actual compatibility changes and repository policy, including prerelease rules and coordinated packages. Write user-facing release notes from the real diff, identifying breaking behavior, migration steps and meaningful limitations.
  • Record source revision, version, expected artifacts, destinations, build inputs and required gates in a release record. A version string alone does not identify the source or artifact.

Phase 2: Verify the candidate

  • Build from the intended clean revision with the recorded toolchain and locked inputs. Preserve unrelated work. Run repository gates and inspect the packaged contents, metadata, checksums and any required signatures or provenance. Verify the contents users will receive.
  • Install or consume the candidate in a disposable environment. Check its reported version and at least one representative behavior; a version command alone is not a functional test.
  • Reuse an existing verified artifact when resuming. If rebuilding is necessary, establish equivalence using the project's reproducibility/provenance policy. Never assume that the same source produces identical bytes or that different bytes are automatically acceptable.

Phase 3: Reconcile remote state before publication

  • Inspect the authoritative state of every target. Classify each expected artifact as absent, present and matching, present and conflicting, or unknown. A timeout means unknown, not absent.
  • Verify that an existing tag resolves to the intended source commit, accounting for tag objects separately from their target commits. Never move or delete it to include newer work.
  • Retry only missing steps whose preconditions still hold. Query unknown outcomes before repeating a mutation. Matching published artifacts stay untouched; a conflict stops that target and needs an explicit repair decision under the registry's immutability policy.
  • Complete authorized steps in dependency order and persist each observed result. Preparation can finish without publication; a partially published version must remain explicitly partial.

Phase 4: Verify the consumer result

  • Read back the tag, release metadata and artifact identities from the actual destinations. Fetch and consume the published version in a clean environment, checking checksums or the applicable provenance and the same representative behavior tested on the candidate.
  • Report per-target evidence, the source revision, artifact identities and any pending or conflicting state. Keep a recovery path that respects immutable versions: a forward fix, withdrawal or correction may need a separately authorized action.
  • Stop at the requested release stage. Record later changes as remaining work; do not assign them a new version, open another release, deploy or announce unless that was requested.

Anti-patterns

  • Rerunning a whole release pipeline after an upload timeout without checking the destination.
  • Publishing from the current checkout when the release tag identifies another revision.
  • Calling a release complete because a tag exists or a publish command exited successfully.
  • Replacing conflicting immutable artifacts, inventing release notes, or starting a follow-up release.

Done when

  • The authorized stage is reached for every required target, or its exact partial state is recorded.
  • Version, source, notes and artifact identities agree, with consumer-side verification evidence.
  • Unrun checks and recovery decisions are explicit; no extra delivery stage is implied.

Use flow-ship for requested Git delivery or flow-triage for an observed release incident; when unavailable, perform the equivalent delivery checks or record the incident evidence.

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.