docs(sharing-service): name the federated phantom-anchor case in the SHARING_NOT_ENABLED list (#8119)

[os-zhuang]

Docs-only follow-up to #8209 (Part of #8119). One line, one file.

Why

content/docs/kernel/runtime-services/sharing-service.mdx has a "Typical Errors"
section that enumerates the exact conditions producing each status. Its
SHARING_NOT_ENABLED (422) entry listed four: public sharing model, no owner_id
field, a bypass object, controlled_by_parent.

#8209 added a fifth — a federated (ADR-0015 external) object whose owner_id
is the platform's injected anchor rather than a real remote column. Leaving it out
is the enforced-but-undocumented inverse of a declared-but-unenforced gap: the
runtime now refuses more than the page says it does, and a reader hitting the new
422 would find no listed condition matching their object.

The docs-drift bot flagged this page on #8209; it was written there but could not be
pushed, because the PR had already been armed and added to the merge queue and a
queued branch cannot be updated. Dequeuing to land a doc line was not the right
trade, so it comes as its own PR — which is the better shape anyway.

The line

- `SHARING_NOT_ENABLED` (422) — `grant` on an object the sharing gates never consult
(public sharing model, no `owner_id` field, a bypass object, `controlled_by_parent`,
or a **federated** object whose `owner_id` is the platform's injected anchor rather
than a real remote column — the platform provisions no storage for a federated
object, so the gates read that column off a table that has not got it and can never
admit).

Re-read against current main before applying rather than assuming: #8209 landed the
code but not this line, so the surrounding list is byte-identical to what the wording
was written against and the original text still fits.

Notes

• No changeset: docs-only change to a hand-written page, no package behaviour. If the
  changeset gate disagrees I will add skip-changeset rather than a changeset.
• content/docs/releases/ is untouched — release-owned and read-only (AGENTS.md
  Documentation Guardrails); the drift bot listed implementation-status.mdx as
  advisory-only and it stays that way.
• The commit that carried this line on the fix(plugin-sharing): refuse a share row on a federated phantom owner anchor, with the single-record gate behaviour measured (#8119) #8209 branch also picked up a stray
  packages/cli/tsconfig.debt-remeasure.json — a check:type-check-debt scratch file
  swept in by a git add -A that ran while the gate was mid-measure. It never reached
  main (that commit was never pushed) and is not in this PR. Filed separately as
  the underlying trap: the gate's temp files are not gitignored.

---

Generated by Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
objectstack	Ignored		Aug 12, 2026 8:53pm

[github-actions]

⛔ merge queue 构建失败 — 先分诊，再决定要不要重排

队列构建 31641477795 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集），
所以失败的测试可能在本 PR 没碰过的包里 —— 那不是重排能修的。每次盲目重排都会让排在后面的所有 PR 重建一轮。

失败的 job（日志抽取，best effort）：

• Dogfood Regression Gate (3/3) — 失败步骤: Boot example apps and exercise real user flows

  �[41m�[1m FAIL �[22m�[49m �[30m�[46m isolated �[49m�[39m test/federated-phantom-share-grant.dogfood.test.ts�[2m > �[22m[#8119] federated phantom anchor: single-record gates + share posture�[2m > �[22m
  

历史信号：

• 本 PR 过去 24h 无队列失败记录（首次）。
• 过去 24h 队列共有 3 个失败构建（不含本次）。

分诊清单：

1. 失败测试在本 PR 改动的包里 → 真回归，修 PR。
2. 失败测试与本 PR 无关 → 在其他 PR 的同类评论里搜同名测试；出现过 ⇒ flaky 实锤，开 issue 修/隔离那条测试。修好前重排只会再烧一轮全队列。
3. 两者都不是 → 可能与同组 PR 语义冲突；等前面的 PR 落地或失败出队后再重排一次即可，不要连续重排。

---

Generated by Claude Code · merge-queue-triage workflow (#4859)