agentleFS
skforecastskforecast/skforecast-ai/.claude/skills/release-bump/SKILL.md

release-bump

Skill52 starsChanged 47 days agoApache-2.0View on GitHub

Bumps the skforecast-ai version in every place that carries it and prepares the release notes section for the new version.

6 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

Bump the version

New version: $ARGUMENTS (ask for it if empty). The current one is __version__ in skforecast_ai/__init__.py.

1. Branch

A new minor version starts a new release branch X.Y.x from the previous one (ask before creating or pushing it). A patch version stays on its release branch. The bump itself goes in a chore/<slug> branch, as any other change.

2. Update every version reference

Replace the old version with the new one in:

  • pyproject.toml (version = ...)
  • skforecast_ai/__init__.py (__version__)
  • plugin/.claude-plugin/plugin.json (version)
  • plugin/.mcp.json (the skforecast-ai-mcp==X.Y.Z pin)
  • .claude-plugin/marketplace.json (version, twice)
  • packages/skforecast-ai-mcp/pyproject.toml (version and the skforecast-ai[mcp]==X.Y.Z pin)
  • server.json (version, twice)
  • tests/test_cli_config.py (the version in the output)
  • docs/quick-start/how-to-install.md (the example of a pinned install)

Then grep the old version across the repository (excluding docs/releases/, dev/, site/, build/, dist/ and the reports under tools/mcp/agent_reports/ and tools/perf/results/) and review each remaining hit. Most are historical and must stay: "added in X.Y.Z" in comments and docstrings, the run tables of tools/mcp/README.md and tools/ai/ask_context_reports/README.md. Others (a test that mentions the version, a TODO such as "review in X.Y.Z") need a decision. List them for the user instead of changing them blindly.

CITATION.cff has no version on purpose.

3. Release notes

The section of the version being closed must not stay as In development: check that it went through the consolidation pass of the release-note skill (/release-note consolidate) and that its heading has the release date (<small>Oct 8, 2026</small>). If not, tell the user before going on.

Then, in docs/releases/releases.md, add a section above the previous one following the existing format:

## X.Y.Z <small>In development</small> { id="X.Y.Z" }

The main changes in this release are:


**Added**


**Changed**


**Fixed**

4. Check

With the Python command of /verify:

python -m pytest tests/test_plugin_distribution.py tests/test_mcp_registry_distribution.py tests/test_cli_config.py -q -p no:cacheprovider
python tools/docs/check_release_notes.py

Remind the user of the steps of the release that follow the bump, all in packages/README.md ("A release, step by step"): the pull request of the release branch to main, opened and green but not merged; the local check of the release on the head of that pull request (tools/release/publish_pypi.py release), before creating it; the GitHub release vX.Y.Z on that commit, which starts .github/workflows/publish-pypi.yml and publishes skforecast-ai and skforecast-ai-mcp to PyPI after the two approvals; the entry of the MCP registry; the merge into main; and selecting Publish for the new version in the plugin directory of Anthropic (auto-publish is off there on purpose).

Report the files changed and the remaining hits of the old version that need a decision.

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.