release_skill

Overview

This skill orchestrates the full release workflow for a new version, covering testing, documentation, and Git tagging. It centralizes verification, changelog updates, and tag creation so releases are consistent and repeatable. Use it to prepare a release candidate before pushing tags and publishing artifacts.

How this skill works

The skill first determines the release version from an argument or the project manifest (e.g., package.json). It runs the test suite through a QA specialist agent and halts on failures. If tests pass, it invokes a technical writer agent to ensure CHANGELOG.md contains an entry for the new version, adding a concise summary of recent commits when needed. Finally, it stages and commits release changes and creates a Git tag for the new version, then notifies you that the release is ready to push.

When to use it

• Preparing a new semantic version release (patch, minor, major).
• Before pushing tags and publishing packages to registries.
• When you need an automated, repeatable release checklist.
• When you want tests and documentation verified before tagging.

Best practices

• Always run the skill from a clean working tree to avoid accidental commits.
• Provide the explicit version argument when bumping to avoid ambiguity.
• Fix failing tests before proceeding; the process stops on test failures.
• Review the changelog entry added by the writer agent before committing.
• Tag releases using the exact format v<VERSION> to keep tags predictable.

Example use cases

• Release v1.2.0 after completing sprint work: runs tests, updates CHANGELOG.md, commits and tags v1.2.0.
• Patch release for a hotfix: supply version argument, validate tests, and create a v1.2.1 tag.
• Preparing a major release: ensure documentation summaries capture breaking changes before tagging.
• Automated CI step that invokes the skill after merge to main, gating the tag creation on passing tests.