fix(spec): remove 8 dead page names from build-docs integration category

[os-zhuang]

Fixes#8166

What

packages/spec/scripts/build-docs.ts's integration category (SECTION_GROUPS) listed 9 page names across 3 sections. Re-enumerated against a fresh gen:docs run rather than trusting the issue's count: packages/spec/src/integration/ contains exactly one .zod.ts module (connector.zod.ts), and content/docs/references/integration/ emits exactly one page (connector.mdx). So 8 of the 9 listed names are dead — mapping, translation, http, message-queue, object-storage, offline, tenant, misc — not the 9 the card's title/body claimed (that number was the size of the whole list, not the dead subset; the dispatch brief carried this correction and asked for re-derivation, which this PR's numbers reflect).

The 8 names trace to #4480, which removed the per-provider connector template files (connector/saas.zod.ts, connector/database.zod.ts, file-storage, message-queue, github, vercel) as the losing side of ADR-0023's rejected per-system modelling — ADR-0097's answer is that provider shapes come from the provider itself, not from the spec (packages/spec/src/integration/index.ts already records this). The SECTION_GROUPS entry was never updated to match.

This integration-category message-queue is that connector-template entry — not the system category's own message-queue.zod.ts (#8075's surface, at build-docs.ts:519, untouched here) nor data's external-lookup (:497, also #8075's, also untouched). #8075 had not landed on main as of this branch's base; the data/system lists here are unmodified.

Removed the 8 dead names and added a deliberate-removal comment, matching the style the automation category already established in the same file (naming what was removed, why, and the issue number).

Verification

• packages/spec/scripts/build-docs.ts's buildCategoryPages filters SECTION_GROUPS entries by what a run actually emits, so this change is a no-op on generated output by construction. Confirmed with a fresh pnpm gen:schema && pnpm gen:docs: 230 files generated both before and after the edit (not the 228 the card cited — confirmed the real number), and git status/git diff show zero bytes changed under content/docs/references/ — only build-docs.ts itself differs.
• pnpm --filter @objectstack/spec check:docs (the --check / CI mode): green, "230 generated files in sync with packages/spec".
• pnpm --filter @objectstack/spec typecheck: green (tsc --noEmit, check:scripts-typecheck, check:test-typecheck).
• pnpm --filter @objectstack/spec exec vitest run scripts --maxWorkers=2: 35 test files / 720 tests, all green.
• pnpm --filter @objectstack/spec exec eslint scripts/build-docs.ts: clean.
• Named gate families (all green): check:adr-anchors, check:authz-resolver, check:changeset-gate-self-tests, check:cross-package-test-inputs (surfaced by a live re-derivation of node scripts/pm/dispatch-gates.mjs against the actual changed path — not in the original named list), pnpm --filter @objectstack/lint run check:doc-formula-expressions, check:docs-audit-scope, check:i18n, check:merge-driver, check:nul-bytes, check:quick-reference-counts, check:release-body, check:role-word, check:spec-parsed-alias.

Out of scope

• data/system page lists in the same file — [security] Two more spec schemas permit inline credentials: ExternalDataSourceSchema.authentication (clientSecret/apiKey) and MessageQueueConfigSchema.sasl.password — census toward #7990's parked boundary-guard reopen trigger #8075, in flight separately.
• content/docs/releases/** — not touched, per standing rule.
• No changeset: this is a generator-input-only change verified to emit zero different bytes (@objectstack/spec is a published package, checked rather than assumed private).

---

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 8:56pm

[github-actions]

📓 Docs Drift Check

No hand-written docs reference the 0 changed package(s). ✅

[github-actions]

⛔ merge queue 构建失败 — 先分诊，再决定要不要重排

队列构建 31641653166 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 4 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)