指导开发者新增可持久化的业务记忆类型:按需选择 Markdown、SQLite 和 LanceDB,确定 Markdown 文件策略,再配置所需的架构、仓库和写入器。
- 新增可持久化的业务记忆类型
- 设计存储层组合
- 连接 Markdown 原文与检索索引
指导将 everos 新版本发布到 PyPI,包括更新版本号和变更日志、创建触发自动工作流的标签,并完成审批。工作流上传后,还需检查并发布 GitHub Release 草稿。
Cut a versioned release and publish everos to PyPI via the tag-triggered workflow
实测评分 未实测
查看安装方法指导将 everos 新版本发布到 PyPI。流程包括更新 pyproject.toml 中的版本号、整理 CHANGELOG.md,以及将改动合并到 main 后创建并推送标签。标签会触发工作流,执行构建、冒烟测试并上传到 PyPI。发布任务会等待审阅者批准。上传完成后,还需检查并发布 GitHub Release 草稿。
完成修复或功能开发后,准备发布正式版本时可以使用此技能;分发预发布版本时也适用。如果需要根据变更日志准备发布页面,并逐步了解发布前后的检查事项,也可以使用。
main 最新状态的 CI 必须成功。标签、pyproject.toml 中的版本号和稳定版 CHANGELOG.md 标题必须一致。发布到 PyPI 需要在 GitHub 环境中手动批准;GitHub Release 也要检查草稿后再发布。
本介绍由 AI(OpenAI Codex)根据公开的 SKILL.md 撰写。准确的规格请查看下方原文。
AI 根据 SKILL.md 的内容检查了需要注意的操作和使用前提。评分只依据实际运行的结果(实测对比)。
不计入评分。 由 AI(OpenAI Codex)于 2026年10月29日 评估。仅供参考,安装前请同时查看原文。
安装前可以查看交给 AI 代理的指令内容。
Publish a new version of everos to PyPI. Publishing is automated: pushing a
vX.Y.Z tag triggers .github/workflows/release.yml,
which builds, smoke-tests, and uploads via PyPI Trusted Publishing (OIDC —
no stored token) behind the release environment's manual-approval gate, then
drafts the GitHub Release page from the CHANGELOG.
A release is not finished when PyPI accepts the upload: the GitHub Release page is drafted, never auto-published, and someone has to write its lead summary and click Publish.
main, up to date, with green CI (the tag builds from main's tree).1. Bump the version → pyproject.toml [project] version = "X.Y.Z"
(single source; everos.__version__ reads installed package metadata)
2. Update CHANGELOG.md → move the Unreleased entries under a new
## [X.Y.Z] - <date> heading, and write the release page's prose here
(lead paragraph + `### Upgrade` group — see "The release page")
3. Commit → git commit -m "chore(release): vX.Y.Z"
4. Open a PR, merge to main after green CI
5. Tag main + push → git tag -a vX.Y.Z -m "vX.Y.Z" && git push origin vX.Y.Z
6. Approve → the release.yml run pauses on the `release`
environment; a reviewer approves in the Actions run
7. Verify → https://pypi.org/project/everos/X.Y.Z/
8. Publish the page → the run leaves a DRAFT GitHub Release, already
complete if step 2 was done properly. Read it once, click Publish.
The tag must equal the pyproject.toml version — the workflow refuses to
publish on a mismatch. A stable tag with no matching ## [X.Y.Z] CHANGELOG
section fails the release job for the same reason.
The whole page is written in CHANGELOG.md, during the release PR. Nothing is meant to be composed at publish time — by then the changes are weeks old and the text gets no review. Write the version section like this:
## [X.Y.Z] - 2026-09-01
**What this release is for.** One paragraph, prose, no bullets — it becomes the
lead of the release page. Say what changed for a user, not what was refactored.
### Added
### Changed
### Fixed
### Upgrade
What a reader must know before upgrading: what happens on first startup, which
command recovers a bad state, which pins moved. Omit the group entirely when a
plain `pip install --upgrade` is all there is — 1.1.4 and 1.2.0 have nothing
here. Do not write filler.
CI turns that into the page: everything above ### Upgrade is lifted verbatim
with the group headings demoted to ##, and the Upgrade prose is wrapped in the
boilerplate — pip line above it, compare link below — which is the shape every
release since 1.1.3 has. See
1.2.1.
So publishing is a read-through and a click. If the draft looks wrong, the fix
belongs in CHANGELOG.md on main, not only in the draft — otherwise the two
drift apart and the next release inherits the habit.
While it is a draft, GitHub serves the release at
releases/tag/untagged-<hash>, and that URL keeps serving a stale page after publication with no redirect. Never share it — a reader who opens it later concludes the release never went out. Linkreleases/tag/vX.Y.Zinstead; the job prints both URLs in its step summary.
Publish with the Publish release button in the web UI. Publishing through
the API by flipping draft alone drops the tag: GitHub rebinds the release to
the untagged-<hash> placeholder and creates a git tag by that name against the
default branch (observed on 1.2.2). Pass tag_name if you must do it from the
CLI:
gh api -X PATCH "repos/EverMind-AI/EverOS/releases/<id>" \
-F draft=false -f tag_name=vX.Y.Z -f make_latest=true
Re-running the release job replaces its own draft and leaves an already-published release untouched, so a re-run is always safe.
PEP 440 pre-release tags publish too (PyPI accepts them; pip install everos
ignores them unless --pre): vX.Y.ZrcN, vX.Y.ZaN, vX.Y.ZbN. Set the same
suffix in pyproject.toml version before tagging.
Their release page is drafted as a pre-release and never becomes
/releases/latest. A pre-release does not need its own CHANGELOG section — the
draft falls back to a one-line placeholder body when there is none.
everos → Settings →
Publishing → add: owner EverMind-AI, repo EverOS, workflow
release.yml, environment release.release
with required reviewers, so every publish needs a manual approval.No PyPI API token is ever stored; the workflow mints a short-lived OIDC token at publish time.
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
指导开发者新增可持久化的业务记忆类型:按需选择 Markdown、SQLite 和 LanceDB,确定 Markdown 文件策略,再配置所需的架构、仓库和写入器。
检查当前 Git 改动和近期提交记录,再按 Conventional Commits 格式创建聚焦的提交。可只暂存相关文件,并处理钩子检查失败的问题。
在更新后的 `main` 分支上创建用途明确的 Git 分支,并按改动类型选择前缀,如功能、修复、文档、CI、维护或重构,再添加简短的小写连字符名称。
使用 GitHub CLI `gh`,按仓库模板将当前分支的改动提交为面向 `main` 的拉取请求。创建前运行 `make ci`,通过后推送分支并填写改动摘要、所属领域、验证结果、检查清单和审阅备注。