docs(steward): add the product-side steward journey case - #4589
Conversation
The steward lane had a reproducible frontend journey but no product-side guidance: an owner had no case explaining what the workspace proves today and which beats it does not. Add `docs/product/use-cases/steward/` (English canonical plus a zh-CN mirror) and list it from the use-case index. The case walks the same seven beats as the browser scenario, states which are proven (first screen, asking the steward, reading the plan card, confirming) and which are recorded gaps, and gives six operating patterns derived from those facts instead of from optimism. Every claim points at the scenario and its report; the gaps table names the probing evidence and the owning surface (plan commit, readiness, alignment, recovery, return delivery). The case explicitly does not qualify a live steward turn, Lark audiences or cloud workers. Validation: `docs-governance-smoke.py` ok, `docs-asset-integrity-smoke.py` ok, private-marker scan clean. Signed-off-by: huangruiteng <14976749+huangruiteng@users.noreply.github.com>
huangruiteng
left a comment
huangruiteng
left a comment
There was a problem hiding this comment.
Approval conclusion (author-owned PR; GitHub blocks formal self-approval)
Exact head: a4faeb5
动机
owner 要求把「本机 steward 旅程」沉淀成产品侧的使用姿势与案例：英文正文加 zh-CN 镜像，并把尚未证明的环节如实作为 gap 记录，而不是写成已完成的能力。这个 PR 就是那份案例文档。
改动思路
案例只写在文档层，并明确绑定一个确定性的浏览器场景（examples/personal-workspace-browser/steward-journey.mjs，跑在合成数据上，由 fixture 代替真实 agent turn），所以文档里描述的每一条都是工作区真实渲染出来的事实，而不是对某个 live Goal 的断言。已证明的 beat 与未证明的 beat 分开列出，未证明的以 typed gap 加探针说明记录，并说明复现方式。
具体改动
三个纯文档文件：
- docs/product/use-cases/steward/README.md：旅程表（beat 1–7，逐条标注 Proven 或 Gap）、四条可复用姿势、以及未证明环节的复现方式；
- docs/product/use-cases/steward/README.zh-CN.md：同内容的 zh-CN 镜像；
- docs/product/use-cases/README.md：新增一行索引链接，指向该案例。
对主干的风险
纯文档，无 runtime、契约、权限、评分或调度改动，也不含私有状态与内部上下文。唯一可能被误解的点是索引页新增了一行链接，但它位于 docs/product/use-cases 列表，不改变任何产品首屏、hero 或主 CTA。
我的整体评价
Approve。文档只记录已验证事实并显式保留未证明项，符合「不把没有证据的推断写成能力」的要求；docs-only 变更使重型检查按 changes 判定跳过，merge-gate 通过。它与 #4588 一起构成 journey gap 1 收口的两个依赖。
English verdict: APPROVE
Signed-off-by: huangruiteng <14976749+huangruiteng@users.noreply.github.com>
01d49fa to
124f4d3
Compare
huangruiteng
left a comment
huangruiteng
left a comment
There was a problem hiding this comment.
Approval conclusion (author-owned PR; GitHub blocks formal self-approval)
Exact head: 124f4d3
动机
owner 要求把「本机 steward 旅程」沉淀成产品侧的使用姿势与案例：英文正文加 zh-CN 镜像，并把尚未证明的环节如实作为 gap 记录，而不是写成已完成的能力。这个 PR 就是那份案例文档。
改动思路
案例只落在文档层，并明确绑定一个确定性的浏览器场景（examples/personal-workspace-browser/steward-journey.mjs，跑在合成数据上，由 fixture 代替真实 agent turn），所以文档描述的每一条都是工作区真实渲染出来的事实，而不是对某个 live Goal 的断言。已证明的 beat 与未证明的 beat 分开列出，未证明的以 typed gap 加探针说明记录，并给出复现方式。
具体改动
三个纯文档文件：
- docs/product/use-cases/steward/README.md：旅程表（beat 1–7，逐条标注 Proven 或 Gap）、四条可复用姿势、未证明环节的复现方式；
- docs/product/use-cases/steward/README.zh-CN.md：同内容的 zh-CN 镜像；
- docs/product/use-cases/README.md：新增一行索引链接。
本 head 只多了一条把已更新的 main 合入的合并提交（带 sign-off），文档内容与上一 head 一致。
对主干的风险
纯文档：无 runtime、契约、权限、评分或调度改动，不含私有状态与内部上下文。索引页新增的一行位于 docs/product/use-cases 列表，不改变任何产品首屏、hero 或主 CTA。
我的整体评价
Approve。文档只记录已验证事实并显式保留未证明项，符合「没有证据的推断不写成能力」；docs-only 使重型检查按 changes 判定跳过，merge-gate 通过。它与 #4588 一起构成 journey gap 1 收口的两个依赖。
English verdict: APPROVE
1 participant
Goal And Delivered Outcome
Source: local steward lane (codex-local-steward), owner direction on 2026-09-17 for a frontend-first steward journey plus product-side usage patterns, with English canonical docs and a zh-CN mirror.
Before: the journey was reproducible (PR #4588) but an owner had no product-side case explaining what the workspace proves today and which beats are unproven.
After:
docs/product/use-cases/steward/holds the case in English and Chinese, listed from the use-case index.Scope And Continuation
Changed surfaces:
docs/product/use-cases/steward/README.md(new English canonical),README.zh-CN.md(semantic mirror), and one index line indocs/product/use-cases/README.md. Docs only: no runtime, protocol, permission or scoring change.Content: the same seven beats the browser scenario walks (first screen; asking the steward; reading the admitted plan card; confirming; readiness; correction; recovery and completion), which of them are proven, six operating patterns derived from those facts, reproduction commands, and a gaps table naming the probing evidence and the owning surface.
Deliberately not claimed: a live steward turn (the fixture substitutes the agent turn), Lark audiences, cloud workers, and any Goal whose plan was confirmed with real consequences.
Public And Private Boundary
No local paths, live Goal or agent names, credentials, private links or organizational context. Reproduction points at the public smoke and its gitignored report path. Private-marker scan over the new docs is clean.
Validation
python3 examples/docs-governance-smoke.py→ okpython3 examples/docs-asset-integrity-smoke.py→ ok (6 assets verified)Follows PR #4588, which supplies the scenario and report this case cites.