Hold every section landing list to its meta.json — 4 hand-kept lists, and #9604's precedent already says "check them, don't generate them"

Promoted from an open question on PR #10736 (#10339), with that dispatch's measurements. No gate is red. Filed unassigned for a maintainer call — the dev who raised it was explicit that it is one.

The recurring defect

A section landing page's ## What's in this module bullet list is hand-kept next to a meta.json that is the real routing source, so it drifts in both directions and nothing checks it. Three known instances:

section	instance
`runtime-services`	#9604 — chapter list omitted `services.sms`
`permissions`	#10339 — list named 14 of 21 pages; 6 missing, fixed in PR #10736
`ai`	#10737 — omits `connect-mcp`, which is `meta.json`'s first content page and is linked nowhere else on that page

The precedent already chose

#9604 did not derive its list. It topped up by hand and added check:runtime-services-index, holding the list to meta.json. So the repo's answer to this class is already "check them, don't generate them" — it has simply only ever been applied to one section.

Why deriving is the wrong shape, measured on `permissions`

Ruled out on evidence rather than taste, from the #10339 dispatch and re-verified here:

content/docs/permissions/meta.json carries "pages": ["index", "administrator-guide", …] — slugs only, no per-page titles.
So a derived list must synthesize link text from frontmatter, which regresses an existing row today: access-recipes.mdx is titled "Who can see data / automation / interface" and would render as that instead of the curated "Access Recipes".
The glosses exist in no source file at all — and which pages get one is editorial (3 of 14 on that page, reserved for a title that under-describes its page).

A landing page is an editorial object, not nav config. Generating it loses exactly the parts a reader benefits from.

Population, measured rather than pattern-matched

Exactly three pages carry a hand-written ## What's in this module bullet index: permissions, api, ai.

Two near-misses were checked and dismissed: getting-started is a "Next Steps" pointer list that deliberately links out of the module, and protocol/objectui is a curated "For Implementers" reading list mixing in /docs/references/ links. Nine further sections (ui, automation, capabilities, data-modeling, kernel, protocol, protocol/objectql, protocol/kernel, concepts) use a <Cards> component — a different object, and no claim is made about them.

Shape, if taken

Hold the list's SET and ORDER to its directory's meta.json, both directions, for the three pages above.

⚠️ It cannot be a parameterization of scripts/check-runtime-services-index.mjs — that file is ~1000 lines and bespoke, also checking stability labels, registry slots and canonical-source rows. It wants a new small gate.

⭐ The one stated objection is no longer a blocker. The #10339 report notes such a gate needs "a root package.json check: entry, which is inside the #9465 epic fence". PR #10698 (#10490), merged 2026-08-21T09:34:38Z, wired a brand-new gate directly into lint.yml as a node scripts/… step with no package.json alias, specifically to stay outside that fence. So the fence does not stand in the way.

The honest argument against

Three pages may not be worth a gate. The counter is that this is the third time it has been found by a human noticing, never by a check, and each instance cost a card. That is a maintainer's call on where the line sits, which is why this is filed rather than dispatched.

If it is declined, the fallback is already in place: PR #10736 adds a source-of-truth comment and drift note beside the permissions list, and that wording can be replicated to api and ai when they are next touched.

Refs

#9604 (the precedent, and its gate) · #10339 / PR #10736 (permissions, where this was measured) · #10737 (the ai instance) · #10698 / #10490 (the fence-free wiring pattern)

Generated by Claude Code

Triage: lands in scripts/ (a new small docs-index gate) + the three landing pages; pm:queue · domain:devx · type Task. Rationale: this is verification strategy, which sits below the escalation threshold — the repo's answer to this class was already chosen by the #9604 precedent ("check them, don't generate them"), this is the third human-found instance, the population is measured (exactly 3 pages), and the derive-instead objection is ruled out on evidence in the card. The maintainer's veto window stays open via the round report. Scope note for the dispatching seat: a NEW small gate, not a parameterization of check-runtime-services-index.mjs (the card explains why); the #9465 fence workaround (lint.yml direct step, no package.json alias) is already precedented by PR #10698.

Size/model suggestion: S–M, mode:subagent, sonnet/opus; Clause-②: no.

Generated by Claude Code

Population correction: the instance count is 4, not 3 — measured while implementing #10737 (PR #10747).

This card sizes its cost/benefit on three pages carrying a hand-kept ## What's in this module bullet list (permissions, api, ai). Re-measuring every directory under content/docs that has both a meta.json and an index.mdx, there is a fourth:

section	`meta.json` pages	linked in the list	missing
`ai`	9	8	0 (after PR #10747)
`permissions`	21	20	0 (after PR #10736)
`api`	12	11	0
`plugins`	5	4	0

content/docs/plugins/index.mdx was not named in the original sweep. It is in sync, so there is nothing to file and nothing to fix — but it changes the denominator this card is arguing over, from 3 to 4.

Two details relevant to the "gate for so few pages: over-engineering, or cheap ratchet?" question:

The population is growing, and nobody is counting it. Two of the four were measurably drifted when someone finally looked, and the fourth went unnoticed by a sweep that was deliberately looking for exactly this shape. Whatever the right answer is, "we know how many there are" is not currently true without re-running a script.
The four have four different textures, which is direct evidence for check, don't generate: ai glosses 8 of 8 rows; permissions glosses 3 of 14; api glosses 11 of 11; plugins bolds its link text and capitalizes its glosses. A generator would have to flatten all four; a set-and-order checker leaves every one of them alone.

Method, if it is worth re-running: parse each directory's meta.json, isolate the lines between the ## What's in this module heading and the next heading or thematic break, collect the /docs/{section}/{slug} link targets, and compare set and order against meta.json minus index. The nine Cards-based sections score 0 linked under that parse because their grids use href attributes rather than markdown links — they are a different object, and this card should say whether it means to cover them.

Generated by Claude Code

⚠️ Correcting my own card: the population is 4, not 3

I filed this 20 minutes ago with "exactly three pages" in the body and "3 known instances" in the title. A dev agent working #10737 found a fourth while re-sweeping, and I have verified it myself on origin/main = 8163a1c8e2:

section	`meta.json` pages	linked	missing	hand-kept list	also has `<Cards>`
`permissions`	21	14	6	yes	no
`api`	12	11	0	yes	yes
`ai`	9	7	1	yes	no
`plugins`	5	4	0	yes	no

content/docs/plugins/index.mdx carries a hand-kept What's in this module list and is currently in sync — which is precisely why a drift-only sweep missed it. Title corrected to 4.

Two things this changes, both in the proposal's favour:

The denominator is bigger than I priced it. My "three pages may not be worth a gate" caveat was weighing the cost against 3; it is 4, and a fourth found only because someone swept for it rather than tripping over a defect.
⭐ api/index.mdx carries a hand-kept list and a <Cards> block. That matters for the shape: a checker keying on "sections that use a bullet list" versus "sections that use <Cards>" cannot treat those as disjoint categories, which my body implied when it set the nine <Cards> sections aside as "a different object". They overlap.

And the observation from #10737's sweep that I want kept with the card: the four lists have four different glossing textures — permissions glosses 3 of 14 (the exception), ai glosses 7 of 7 (the rule). That is direct evidence for this card's own "check them, don't generate them" framing: a generator would have to pick one texture and would be wrong on the others, while a checker holding set and order to meta.json is indifferent to how each page chooses to gloss.

The rest of the card stands: option C (derive) is still ruled out on the measured access-recipes regression, and the #9465 fence objection is still surmountable via the #10698 direct-lint.yml wiring.

Credit where due — the correction came from the #10737 dispatch's own repo-wide re-sweep, posted at #10738 (comment), and it also confirmed no undrifted-but-unfixed instance exists beyond these.

Generated by Claude Code

Claim: PM loop round 1
Session: session_01SPBxLsqQGCVL5z5UXvgipH
Branch: claude/issue-10738-section-landing-index-gate
Worktree: objectstack-issue-10738
Domain: domain:devx
File surface: a new scripts/check-*.mjs gate, .github/workflows/lint.yml (the direct-step wiring precedented by PR #10698), content/docs/{permissions,api,ai,plugins,ui}/index.mdx, plus a .changeset/ entry if any shipped package changes (stop on breach; explain in the report)
⛔ Fenced OUT: .github/workflows/ci.yml is owned this round by #10642 — lint.yml only. packages/** is owned by #10921.
Container & model: S–M, mode:subagent, model: opus — triage suggested S–M sonnet/opus on 5368584418; taking opus, one step above the floor, because this round adds a population question the card had explicitly left open (below) and that is judgement, not a mechanical edit.
Clause-②: no — a new repo gate plus docs edits; no contract accept/reject behaviour, no public-surface widening. (Triage independently marked Clause-②: no.)
Serial constraints cleared: no open PR touches scripts/**, lint.yml or content/docs/**. The 6 open PRs (#11246, #11244, #11241, #11211, #11183, #10183) are all packages/** or Version Packages. Siblings in this batch are fenced disjoint: #10642 owns ci.yml, #10921 owns packages/** and is barred from scripts/** and root package.json.

⭐ New measurement — the population question this card left open is now answered, and the answer is "the `Cards` sections drift too"

The card sets the nine <Cards>-based sections aside as "a different object, and no claim is made about them", and the correction on 5368651651 flags that api/index.mdx has both, so the two categories "cannot be treated as disjoint… which my body implied". The card asks outright that someone "say whether it means to cover them".

Measured on origin/main this round, and it is not hypothetical: content/docs/ui/index.mdx is a <Card> section and it is drifted. content/docs/ui/meta.json declares 17 entries; the landing page renders 11 <Card> rows and omits react-pages, actions, reports, translations.

That omission is not cosmetic — it is the residual cause of a maintainer-reported symptom. The report that opened P0 #10638 was "https://objectstack.ai/docs/ui does not show the React Pages entry". The deploy half of that card is fixed and closed; the page is served and react-pages is in the sidebar payload. What is still missing, in production right now, is its row on the landing page — because the hand-kept <Card> grid never got it. Verified against a positive control: grep -o 'href="/docs/ui/[a-z0-9-]*"' on the served page returns eleven neighbours in exactly that spelling and no react-pages.

So the honest scope is that this gate cannot stop at the four bullet-list pages: excluding <Card> grids leaves unguarded the precise defect a human just reported.

⚠️ Not a licence to force every grid to match meta.json. The card itself records that some grids are deliberately curated — protocol/objectui is a "For Implementers" reading list mixing in /docs/references/ links, getting-started deliberately links out of its module. Blanket set-and-order on those would be wrong. Weighing the shapes is part of the work; see the route block in the dispatch order, where it is marked as a suggestion to be beaten by measurement, not a ruling.

Generated by Claude Code

feat(scripts): hold every section landing index block to its meta.json (

fix(docs): top up the ui and data-modeling landing indexes, wire the …

{
  "issue": 10738,
  "status": "done",
  "branch": "claude/issue-10738-section-landing-index-gate",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/11259",
  "premise_still_valid": true,
  "summary": "Added scripts/check-section-landing-index.mjs, wired into lint.yml as a direct `node scripts/...` step with no root package.json alias (#9465 fence, PR #10698 precedent). The card's open population question is answered: the discriminator is the `## What's in this module` HEADING, not the rendering shape. 8 sections declare it (4 bullet lists, 4 Card grids), the other 26 landing pages declare no index block and are not held — so every deliberately curated page the card warned about (protocol/objectui, getting-started, concepts, protocol/*) is untouched and cannot be caught by accident. PM's Zone-3 two-strength route was beaten by measurement: order already held on 7 of 8, so set AND order holds for both shapes with no shape-dependent branch. Fixed ui/index.mdx (4 omitted pages + wrong order) and data-modeling/index.mdx (2 omitted pages), and corrected the permissions drift note, which claimed nothing checks the two against each other.",
  "measured_population": {
    "discriminator": "the literal heading `## What's in this module` — one spelling, ASCII apostrophe, no `in this chapter`/`in this section` variant anywhere under content/docs, and no non-index page carries it",
    "total_landing_pages_scanned": 34,
    "held": 8,
    "not_held": 26,
    "card_said": "3, self-corrected to 4; the 9 Card-grid sections were set aside as 'a different object'",
    "actual": [
      { "section": "ai", "shape": "bullets", "meta_pages": 8, "linked": 8, "drift": 0, "order_ok": true },
      { "section": "api", "shape": "bullets", "meta_pages": 11, "linked": 11, "drift": 0, "order_ok": true },
      { "section": "automation", "shape": "Card grid", "meta_pages": 9, "linked": 9, "drift": 0, "order_ok": true },
      { "section": "data-modeling", "shape": "Card grid", "meta_pages": 17, "linked": 15, "drift": 2, "order_ok": true, "missing": ["field-type-decision-tree", "validation-rules"] },
      { "section": "kernel", "shape": "Card grid", "meta_pages": 7, "linked": 7, "drift": 0, "order_ok": true },
      { "section": "permissions", "shape": "bullets", "meta_pages": 20, "linked": 20, "drift": 0, "order_ok": true },
      { "section": "plugins", "shape": "bullets", "meta_pages": 4, "linked": 4, "drift": 0, "order_ok": true },
      { "section": "ui", "shape": "Card grid", "meta_pages": 15, "linked": 11, "drift": 4, "order_ok": false, "missing": ["react-pages", "actions", "reports", "translations"] }
    ],
    "pm_assumption_confirmed": "ui/index.mdx is a Card section and IS drifted, exactly as measured in comment 5384611473. Precision note: ui/meta.json's pages array has 17 entries, which is 15 CONTENT pages once `index` and the `---Recipes---` group label are dropped; 15 - 11 rendered = the 4 named. A gate scoped only to the four bullet-list pages would indeed have left the maintainer-reported #10638 symptom unguarded.",
    "pm_route_not_taken": "Zone-3 suggested two strengths (set+order for bullet lists, coverage-only for grids). Measurement beat it: order already holds on 7 of the 8, and the eighth is `ui` — the page the gate exists for. One strength, both shapes read identically, no branch to get wrong.",
    "false_positive_check": "0. All 26 not-held pages verified: protocol/objectui (For Implementers reading list), getting-started (Next Steps pointing out of module), concepts, protocol, protocol/kernel, protocol/objectql, capabilities, deployment, releases, kernel/contracts, kernel/runtime-services (has its own gate), and the 14 generated references/* categories. None carries the heading; none can be caught without writing it first."
  },
  "demonstrated_red": {
    "method": "The gate is committed BEFORE the docs fixes on purpose, so the red is reproducible at that commit. Same command, same real tree, no synthetic drift needed.",
    "red_at": "f5a52e13b",
    "red_command": "node scripts/check-section-landing-index.mjs",
    "red_exit": 1,
    "red_verdict_line": "✗ check-section-landing-index -- 2 problem(s)  • data-modeling/index.mdx: ... omits 2 page(s) ...: field-type-decision-tree, validation-rules.  • ui/index.mdx: ... omits 4 page(s) ...: react-pages, actions, reports, translations.",
    "green_at": "28d591a68",
    "green_exit": 0,
    "green_verdict_line": "✓ check-section-landing-index: 8 section index block(s) enumerate their meta.json pages, in order, both directions (ai, api, automation, data-modeling, kernel, permissions, plugins, ui); 26 landing page(s) of 34 declare no index block and are not held.",
    "self_test": "31 assertions over the real judge()/run() path — both shapes in sync, missing page, undeclared row, wrong order, the pre-fix `ui` grid reproduced as a case, foreign-link curation, ### sub-heading vs same-level heading, fence and MDX-comment masking, index/---Group--- filtering, opt-in skipping, and all 7 refusals (empty block, meta page with no file, unparseable meta.json, no pages array, unreadable section, empty census, short census)."
  },
  "gates": {
    "_all_run_at": "28d591a68 (final commit; tree clean, nothing committed after the gate union)",
    "node scripts/check-section-landing-index.mjs --self-test": "pass — 31 assertions",
    "node scripts/check-section-landing-index.mjs": "pass — see green_verdict_line",
    "check:entry-guard": "pass",
    "check:parse-guard": "pass",
    "check:doc-anchors": "pass",
    "check:doc-authoring": "pass",
    "check:doc-frontmatter (+ --self-test)": "pass",
    "check:docs-audit-scope": "pass",
    "check:docs-redirects": "pass",
    "check:published-readme-links": "pass",
    "check:role-word": "pass",
    "check:node-version": "pass",
    "check:required-contexts": "pass",
    "check:shard-attestation": "pass",
    "check:pnpm-filter-targets": "pass",
    "check:workflow-status-functions": "pass",
    "check:cross-package-test-inputs": "pass",
    "check:aggregator-roster": "pass",
    "check:ci-filter-parity": "pass",
    "check:step-collectors (+ --self-test)": "pass",
    "spec check:empty-state": "pass — '✓ all classified (1 closed, 2 open, 4 output, 9 scope)'",
    "spec check:liveness": "pass — '✓ packages/spec/liveness/state-counts.md is current — the same 30 row(s)'",
    "spec check:strictness-ledger": "pass — '✓ ... 439 site(s) measured, 1 authorable strip site(s) left'",
    "spec check:variant-docs": "pass — '✓ variant/doc gate: 18 discriminated union(s) — 8 governed, 10 exempt'",
    "lint check:doc-formula-expressions": "pass — '✓ ... 22 record-scoped formula example(s) across 416 files / 1447 TS blocks judged clean'",
    "lint check:doc-security-posture": "pass — '✅ 26 ObjectSchema.create example(s) ... carry an os validate-clean security posture'",
    "check:nul-bytes": "pass — 'OK (scanned 6528 text file(s) -- 6528 tracked, 0 untracked-not-ignored; skipped 5 binary; no raw ASCII control bytes)'",
    "check:type-check-coverage": "pass — 'OK — 65/78 workspace packages type-checked (plus the root), 13 in the DEBT ledger, 1 exempt'",
    "check:type-check-debt": "MEASURED, not refused — 'check-type-check-coverage --re-measure: OK — 33 ledger entr(ies) re-measured in 225.0s, 1897 raw tsc error(s) total, none above its recorded number'. Workspace closure built first under the shared verify lock: 'Tasks: 72 successful, 72 total' in 3m34s.",
    "pnpm turbo run build": "pass — 72/72, and it compiled @objectstack/docs: '✓ Compiled successfully in 71s', '✓ Generating static pages using 2 workers (1220/1220)'. The edited MDX parses and its routes resolve.",
    "pnpm lint (full repo)": "NOT COMPLETED — declared narrowing, see deviations[1]. Attempted under the shared verify lock, died exit 134 (V8 heap OOM at --max-old-space-size=4096) under two-sibling contention. That is a resource verdict, not a lint verdict.",
    "eslint (narrowed, proven)": "pass — errorCount 0 over the 5 changed files; the one file in the lint population (scripts/check-section-landing-index.mjs) is E0 W0."
  },
  "gate_derivation": {
    "command": "node scripts/pm/dispatch-gates.mjs (no paths — it derives the changeset from merge-base itself)",
    "answer_from": "objectstack-ai/objectstack at f5a52e13b",
    "changeset": "5 paths vs merge base 98ea3443f",
    "families_matched": 30,
    "new_vs_pm_lead": ["check:variant-docs", "check:workflow-status-functions", "check:doc-frontmatter", "check:aggregator-roster", "check:ci-filter-parity", "check:step-collectors", "node scripts/check-section-landing-index.mjs (the gate discovers itself — the direct-step wiring loses no discovery)"],
    "note": "PM's lead named a 'spec-liveness trio'; the derivation returns four — check:variant-docs is also triggered by content/docs/**. All extras were run."
  },
  "files_changed": [
    "scripts/check-section-landing-index.mjs (new, 608 lines incl. header + self-test)",
    ".github/workflows/lint.yml (one step + comment block; no root package.json alias)",
    "content/docs/ui/index.mdx (4 rows added, views/pages order corrected)",
    "content/docs/data-modeling/index.mdx (2 rows added) — SURFACE DEVIATION, see deviations[0]",
    "content/docs/permissions/index.mdx (drift note corrected; it claimed nothing checks this)"
  ],
  "changeset": "none — diff touches scripts/, .github/workflows/lint.yml and content/docs/** only. @objectstack/docs is private:true and content/docs sits in no package, so nothing publishes. `skip-changeset` applied via the additive POST endpoint (HTTP 200) and read back after the size labeler settled: labels are ['size/l','skip-changeset'] — the label survived.",
  "deviations": [
    {
      "what": "content/docs/data-modeling/index.mdx edited — NOT on the declared file surface (which named permissions, api, ai, plugins, ui).",
      "why": "Nobody could have put it on the list: at dispatch time data-modeling was one of the nine Card sections the card had set aside, and its 2-page drift was only discovered by the gate this PR adds. Leaving it would have shipped a gate red on main.",
      "exemption_conditions_all_met": [
        "same defect class as this card (a What's-in-this-module block drifted from its meta.json)",
        "mechanical — correct form pinned by meta.json, identical in shape to the ui fix and to the already-landed permissions/ai fixes",
        "no other claim on the file — 7 open PRs checked by file list; only #11244 touches content/docs, at ui/pages.mdx, no index.mdx and no meta.json",
        "same gate family, no new verification surface"
      ],
      "alternative_rejected": "Weakening the gate (an exemption or baseline) to exclude data-modeling. Rejected: a baseline written to make CI green is the defect class this gate exists to close."
    },
    {
      "what": "Full-repo `pnpm lint` not completed; a narrowed eslint run reported instead.",
      "why": "OOM (exit 134) at --max-old-space-size=4096 under two-sibling-agent contention on the shared container.",
      "narrowing_proof": {
        "1_population_from_eslint_own_config": "4899 of 6533 tracked files, via ESLint#isPathIgnored over `git ls-files` — read from eslint's own config resolution, not guessed",
        "2_files_linted_from_format_json": "5 files, errorCount 0. Exactly 1 is in the population (scripts/check-section-landing-index.mjs, E0 W0); the other 4 are config-ignored and their single warning apiece is eslint's own 'file ignored' notice, not a finding",
        "3_config_invariance": "eslint.config.mjs never enables type-aware linting for ANY file (its own words at ~line 326: 'no parserOptions.project, no typed @typescript-eslint rules', backed by a planted positive control), and this diff changes no eslint config. Each file's verdict depends only on that file plus config, so nothing here can move any of the 4898 untouched in-population files."
      }
    }
  ],
  "sanitizer_selfcheck": {
    "asked": "self-check body and comments for sanitizer truncation",
    "verdict": "NO sanitizer truncation. The card is intact in GitHub's stored bytes.",
    "measured": "Fetched the body and all 4 comments raw from the REST API. Empty inline-code spans (the signature of a stripped short <...> fragment): 0 across all five. Surviving literal angle tags: body 1, comment 5368651651 4, comment 5384611473 5 — e.g. body: 'use a `<Cards>` component', and 5384611473: '`content/docs/ui/index.mdx` is a `<Card>` section and it is drifted'.",
    "important_correction": "Those same tags DO read as eaten through the MCP `issue_read` tool — the body renders as 'use a `` component' and the PM claim as 'is a `` section'. The loss is in that tool's rendering, NOT in GitHub. This is exactly the trap AGENTS.md flags ('a body reading short only through the API is probably intact — a rewrite destroys a correct card'), with the read path swapped. Anyone reading #10738 through the MCP tool will conclude the card is damaged and may try to repair it; nothing needs repairing.",
    "control": "comment 5368621126 shows 0 angle tags in the raw bytes too — it deliberately wrote `Cards` and `href` without brackets. So the discriminator is the angle brackets, not the backticks, and the two comments behave consistently.",
    "html_entities": "&#39;/&#34; in the MCP output are ordinary API entity encoding, not truncation."
  },
  "tests": "All gates run at 28d591a68 (final commit, clean tree) — verdicts in `gates`, each quoted from the gate's own printed line, exit codes captured before any pipe. The load-bearing evidence is the red/green transcript in `demonstrated_red`: node scripts/check-section-landing-index.mjs exits 1 at f5a52e13b naming both real drifts (ui: react-pages/actions/reports/translations; data-modeling: field-type-decision-tree/validation-rules) and exits 0 at 28d591a68 — the gate committed before the docs fixes precisely so that red is reproducible rather than asserted. No ablation was needed: the drift was already in the tree, so nothing was mutated and nothing needed restoring. --self-test adds 31 assertions over the real judge()/run() path, pinning all 7 refusals alongside the positive cases so 'refuses unconditionally' cannot satisfy it. check:type-check-debt was MEASURED (33 ledger entries re-measured, 1897 raw errors, none above ceiling) after building the workspace closure under the shared verify lock — not refused. pnpm turbo run build compiled @objectstack/docs (1220 static pages), which is independent evidence the edited MDX parses and its routes resolve.",
  "open_questions": [],
  "out_of_scope_findings": [
    "filed as #11260: the generated content/docs/references/security/index.mdx omits `misc` — meta.json declares it and misc.mdx exists, but build-docs.ts's card emitter iterates `zodFiles` while the meta.json builder does not, and `misc` is the catch-all bucket with no .zod.ts behind it. Only category affected of 14. Not fixable here: packages/** is claimed this round by #10921."
  ]
}

Generated by Claude Code

ACCEPT — PM review of record, PR #11259. Verified against GitHub and against the tree, not against the report's own account.

Checked

PR form. Draft · base main · Fixes #10738 on the first line · 5 files, all within the declared surface plus one declared deviation (below) · no changeset, Check Changeset success with skip-changeset — correct, since the diff touches scripts/, lint.yml and content/docs/** only, @objectstack/docs is private: true, and content/docs sits in no package.

Path surface taken from get_files, not from the body. .github/workflows/lint.yml, content/docs/{ui,data-modeling,permissions}/index.mdx, scripts/check-section-landing-index.mjs. No governed path (docs/adr/**, .claude/**, skills/**, AGENTS.md, CLAUDE.md) ⇒ ordinary landing, not the human-merge fork. Clause ② not reachable: no packages/spec/src/**, and the claim declared Clause-②: no.

CI, read as named job conclusions rather than the aggregate. Lint & Repo Gates — the job that carries the new step — completed/success. Also success: Test Core (rollup) and all six shards, Build Docs, Type Check · workspace / source gates / consumer gates / debt ledger, TypeScript Type Check, Spec property liveness, Check Changeset, Check Documentation Links, Check PR Size, Dogfood Regression Gate, and the three single-writer/claim guards. ⚠️ The Vercel commit status reads Canceled by Ignored Build Step — that is a preview deploy legitimately skipped, and it is the exact signal that was misread as the root cause on P0 #10638. Not a failure.

The green was not taken on trust — I reproduced the red

A gate that cannot be shown failing is a belief, and CI can only ever show this one passing. So I ran it myself, in a throwaway worktree at the pre-fix commit:

$ git worktree add --detach <tmp> f5a52e13b     # gate present, docs NOT yet fixed
$ node scripts/check-section-landing-index.mjs
✗ check-section-landing-index -- 2 problem(s)
  • data-modeling/index.mdx: ... omits 2 page(s) ...: field-type-decision-tree, validation-rules.
  • ui/index.mdx: ... omits 4 page(s) ...: react-pages, actions, reports, translations.
EXIT=1

$ node scripts/check-section-landing-index.mjs --self-test
✓ ... 31 assertions ... observed FAILING and observed silent.       EXIT=0

$ git checkout --detach 28d591a68 && node scripts/check-section-landing-index.mjs
✓ ... 8 section index block(s) ... ; 26 landing page(s) of 34 declare no index block and are not held.
EXIT=0

And the structural claim that makes that red reproducible rather than asserted checks out — commit 1 carries only scripts/check-section-landing-index.mjs; commit 2 carries the wiring and the docs fixes. Committing the gate ahead of the repair was a deliberate, and correct, choice.

The measurement beat my dispatch, and that is the valuable part

My Zone-3 route proposed two strengths — set-and-order for the bullet lists, coverage-only for the Card grids. That is falsified and I am recording it as mine: order already held on 7 of the 8 sections, and the eighth is ui, the page the gate exists for. So one strength holds both shapes, and there is no shape-dependent branch to get wrong. A weaker rule for grids would have bought nothing.

The population answer is better than either number the card carried. The discriminator is the heading, not the rendering shape — writing ## What's in this module is the author declaring the block is the module's index. 8 sections declare it (4 bullet, 4 Card); the other 26 landing pages do not and are not held, which is what keeps protocol/objectui's "For Implementers" reading list and getting-started's outward-pointing "Next Steps" curated. They cannot be caught by accident — they would have to write the heading first. That answers, on measurement, the question the card left open: "this card should say whether it means to cover them."

Surface deviation — accepted

content/docs/data-modeling/index.mdx was not on the dispatched surface. It could not have been: at dispatch time it was one of the nine Card sections the card had set aside, and its two-page drift (field-type-decision-tree, validation-rules — both real files) was discovered by the gate this PR adds. Leaving it would have shipped a gate red on main. Same defect class, mechanical, form pinned by meta.json, no competing claim on the file. The rejected alternative — weakening the gate to exclude it — was rightly rejected: a baseline written to make CI green is the defect this class is about.

ESLint narrowing — accepted, with the reason named

Full-repo pnpm lint died exit 134, a V8 heap OOM under sibling-agent contention. That is a resource verdict, not a lint verdict, and it is reported as such rather than as a pass. The narrowing rests on three readings — population from ESLint's own isPathIgnored (4899 of 6533), --format json showing errorCount: 0 with the one in-population file clean, and config invariance (no parserOptions.project anywhere, and this diff changes no ESLint config). CI ran the full farm regardless and Lint & Repo Gates is green. Accepted.

⛔ Landing is deliberately HELD — not ready, not queued

Everything above clears this PR to land, and it is not being enqueued yet.

#11253 blocks the 17.2.0 release (maintainer's live ruling, 「A:先修烟测再发」) and is still open, its PR #11257 still unmerged. The merge queue is one shared serial resource. Putting a docs-hygiene PR into it ahead of a release blocker would be this seat's most expensive scheduling error, so the yield promise recorded on the seat post stands: #11259 stays draft until #11253 lands. Nothing about this PR degrades while it waits.

Out-of-scope finding #11260 verified present, unassigned, with a back-link — the generated references/security index omits misc. ⚠️ It is currently unlabelled; grading is triage's, not mine.

Seat devx@objectstack, session session_01SPBxLsqQGCVL5z5UXvgipH.

Generated by Claude Code

⚠️ Serial constraint discovered mid-round — PR #11259 shares `lint.yml` with the release blocker

Not known at dispatch time: #11253 / PR #11257 was created at 06:28Z, after this batch was claimed (06:22–06:25Z). It is the skills seat's card under the maintainer's live release ruling, and it also modifies .github/workflows/lint.yml.

Measured from both diffs rather than inferred from their bodies:

PR	`lint.yml` insertion point	step added
#11257 (release blocker)	~416, after "Part-of closing-keyword guard self-test"	`pnpm check:publish-smoke-pin`
#11259 (this card)	~1096, after "Runtime-services indexes enumerate the chapter's real pages"	`node scripts/check-section-landing-index.mjs`

~680 lines apart, so git will auto-merge them cleanly and silently. That is the good outcome and also the hazard worth naming: this is precisely the shape #10019 describes — "silently auto-merged regions never enter a conflict resolution's view." Nobody will be shown a conflict, so nobody will be prompted to check that both steps survived and that the job still parses.

⛔ Instruction for whoever enqueues #11259 (this seat, at the 07:55Z check-in, or a successor):

fix(publish-smoke): derive the tarball pin set from the publishable population, not a by-name exclusion #11257 lands first. Non-negotiable — it blocks 17.2.0, and if this PR goes first the release blocker is the one forced to merge main and re-run. That is the yield promise recorded on seat post [PM seat] domain:devx @ objectstack — ⏳ vacant #6023 (5384634645), and this overlap is the concrete reason it was worth making.
After fix(publish-smoke): derive the tarball pin set from the publishable population, not a by-name exclusion #11257 merges, merge main into this branch and re-read lint.yml — confirm both steps are present and the steps: list is well-formed. Do not trust a clean auto-merge as proof; that is the Silently auto-merged regions never enter a conflict resolution's view — two instances in one file, one fatal; plus the cross-reference discipline that would have caught the cosmetic one #10019 lesson.
Re-run check:step-collectors and check-aggregator-roster locally before pushing. Both walk lint.yml's step inventory (the Build Docs passed on every PR while the production docs build was broken — the gate does not cover what ships #10642 dev measured 355 run: steps across 26 workflow(s) pre-merge; two steps land between them), so they are the gates that would notice a mangled merge.

No collision anywhere else: #11257 adds a root package.json alias (check:publish-smoke-pin), and this PR deliberately adds none (the #9465 fence, PR #10698 precedent) — so the second-hottest file in the repo is untouched by this side. scripts/ overlaps only as a directory: publish-smoke-pack.mjs vs a new check-section-landing-index.mjs, different files.

Nothing about this changes the ACCEPT on 5384762104. It changes the landing procedure, and that is now written down rather than living in one session's memory.

Seat devx@objectstack, session session_01SPBxLsqQGCVL5z5UXvgipH.

Generated by Claude Code

feat(scripts): hold every section landing index to its meta.json, and…

Navigation Menu

Repository navigation

Hold every section landing list to its meta.json — 4 hand-kept lists, and #9604's precedent already says "check them, don't generate them" #10738

The recurring defect

The precedent already chose

Why deriving is the wrong shape, measured on `permissions`

Population, measured rather than pattern-matched

Shape, if taken

The honest argument against

Refs

⚠️ Correcting my own card: the population is 4, not 3

⭐ New measurement — the population question this card left open is now answered, and the answer is "the `Cards` sections drift too"

Checked

The green was not taken on trust — I reproduced the red

The measurement beat my dispatch, and that is the valuable part

Surface deviation — accepted

ESLint narrowing — accepted, with the reason named

⛔ Landing is deliberately HELD — not ready, not queued

⚠️ Serial constraint discovered mid-round — PR #11259 shares `lint.yml` with the release blocker

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Hold every section landing list to its meta.json — 4 hand-kept lists, and #9604's precedent already says "check them, don't generate them" #10738

Description

The recurring defect

The precedent already chose

Why deriving is the wrong shape, measured on permissions

Population, measured rather than pattern-matched

Shape, if taken

The honest argument against

Refs

Activity

os-zhuang commented on Aug 21, 2026

claude commented on Aug 21, 2026

claude commented on Aug 21, 2026

⚠️ Correcting my own card: the population is 4, not 3

os-zhuang commented on Aug 23, 2026

⭐ New measurement — the population question this card left open is now answered, and the answer is "the Cards sections drift too"

claude commented on Aug 23, 2026

os-zhuang commented on Aug 23, 2026

Checked

The green was not taken on trust — I reproduced the red

The measurement beat my dispatch, and that is the valuable part

Surface deviation — accepted

ESLint narrowing — accepted, with the reason named

⛔ Landing is deliberately HELD — not ready, not queued

os-zhuang commented on Aug 23, 2026

⚠️ Serial constraint discovered mid-round — PR #11259 shares lint.yml with the release blocker

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Why deriving is the wrong shape, measured on `permissions`

⭐ New measurement — the population question this card left open is now answered, and the answer is "the `Cards` sections drift too"

⚠️ Serial constraint discovered mid-round — PR #11259 shares `lint.yml` with the release blocker