feat(spec): register the ADR-0087 semantic entry ui-html-page-div-refused (#20592)

[objectstack-fleet]

Closes #20592

Registers ui-html-page-div-refused, the protocol-18 ADR-0087 semantic entry for the html-tier div refusal that PR #20589 brings to projects with no sdui.manifest.json of their own (div → box).

Clause-②: no

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

1 anchor(s) derived from 1 changed package(s); no hand-written page names any of them, so this run has nothing to list — not a clean bill of health. This check sees only pages that NAME a derived anchor: one that documents this change in prose, or enumerates it in an authoring dialect, names none and stays invisible to it on every run.

What this run could not see

• 4 name(s) were too generic to anchor anything (single lowercase words)
• the SDK route bridge reached 54 of 206 client-bound route-ledger rows — the other 152 have no registrar path: tail to select them, so pages documenting THEIR client methods cannot appear above, on this or any run. Of those 152: 0 are remediable by widening that discovery convention (an in-repo file declares the path; the convention did not scan it); 55 are structural — on a ledger where NOT ONE row is declared in-repo, so no discovery change reaches them at any price; 97 are undecided (no in-repo declaration, on a ledger that has other in-repo registrars — absence and an unreadable spelling are not distinguishable here). The rows themselves: node scripts/docs-audit/affected-docs.mjs --bridge-coverage
• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.
• a key NAME is not a key, so the hand re-read the line above prescribes can land on the wrong schema. The same spelling is authorable on one governed type and a [REMOVED] tombstone on another for each of active, aria, joins, objects, template, tools and version (censused on [finding] tools is a key on BOTH AgentSchema (tombstoned, dead) and SkillSchema (live, cloud-attested), so a name-based search attributes skill examples to the agent key — it produced a false stop-the-line alarm on PR #19059 #19093 over the liveness ledger's governed types, top-level keys); nothing in a search result distinguishes the two, so a grep hit on a LIVE example reads as evidence about the DEAD key. Measured on fix(spec): the agent.tools liveness row says dead — it claimed live on a key the schema tombstoned #19059: content/docs/ai/agents.mdx was reported as contradicting the agent.tools tombstone over its tools: example at :161, which is inside the defineSkill({ block opened at :155 — the page was already correct. Settle ownership by PARSING the value against both schemas, never by the name: that literal PASSES SkillSchema, and as an AgentSchema it FAILS at tools with the tombstone prescription. ⛔ These names are not the whole class — a key retired through a .strict() guidance map leaves no tombstone in the walked shape and none of them here (tool.category, live as AIToolDefinition.category).

Coarse fallback — 137 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json 0f6dcac5e99d0c6211f0d8a0e150a112d78a776f → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 3c7b0ca10dc39d0f0f519577457b286bb3e255df — the merge of head 95ae699a5bb78a0cf6e11f687b08d63c459d484b into base 0f6dcac5e99d0c6211f0d8a0e150a112d78a776f, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin 3c7b0ca10dc39d0f0f519577457b286bb3e255df && git checkout 3c7b0ca10dc39d0f0f519577457b286bb3e255df
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin 0f6dcac5e99d0c6211f0d8a0e150a112d78a776f 95ae699a5bb78a0cf6e11f687b08d63c459d484b && git checkout -B drift-repro 0f6dcac5e99d0c6211f0d8a0e150a112d78a776f && git merge --no-ff 95ae699a5bb78a0cf6e11f687b08d63c459d484b

node scripts/docs-audit/affected-docs.mjs --json 0f6dcac5e99d0c6211f0d8a0e150a112d78a776f

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

[objectstack-fleet]

Contract review

Served-tier: CONTRACT_REVIEW_TIER
Head-sha: 95ae699a5bb78a0cf6e11f687b08d63c459d484b
Local-runs: none

① Derived judgments

Diff read as main 0f6dcac5e9 … 95ae699a5b (merge-base 7a1faf1a5d; two commits; 3 files, +132/−0): the new semantic entry packages/spec/src/migrations/entries/semantic/18.ui-html-page-div-refused.ts, the regenerated registry.ts (the same entry inside step18.semantic), and .changeset/20592-ui-html-page-div-refused.md. The pinned objectui .objectui-sha is dd3f7e1be3 on both main and the head, and objectui#10808's merge 39b8d51022 is its ancestor.

Accept-set changes implied: none — right. No Zod schema, tombstone, guidance map or D2 conversion moves. Public-surface changes implied: one, additive data — right. MIGRATIONS_BY_MAJOR[18].semantic gains one element (317 semantic entries; the SemanticMigration type is unchanged), so objectstack migrate meta prints one more [protocol 18] block for any range whose --to reaches 18 — which the default does: CHAIN_TERMINUS_MAJOR = max(PROTOCOL_MAJOR, ...MIGRATION_MAJORS) = 18 (packages/cli/src/commands/migrate/meta.ts:88). spec-changes.json and docs/protocol-upgrade-guide.md are rightly absent from the diff: both generators loop major from the floor to PROTOCOL_MAJOR = 17 (build-upgrade-guide.ts:78, build-spec-changes.ts:254), so a step-18 entry renders nowhere yet. Step 18 is the right step: PROTOCOL_VERSION is 17.0.0, the precedent 18.ui-react-list-view-binding-aliases-retired and every recent narrowing sit under 18. The id equals the marker PR #20589's changeset carries at d9dc0be3ed (adr-0087: registered ui-html-page-div-refused).

Moment key: [now] = true on main the moment this PR lands; [after #20589] = true only once PR #20589 (branch claude/issue-19922-console-manifest-fallback, draft) lands.

surface — right.

• "kind:'html' page source (and its deprecated kind:'jsx' alias)": right [now]. page.zod.ts:839-843 names 'jsx' the deprecated alias; the CLI gate's checkedPageKeys reads both kinds (packages/cli/src/utils/sdui-manifest.ts).
• "in a project with no sdui.manifest.json of its own": right [now] as the scope — resolveSduiManifest reads cwd/sdui.manifest.json first and falls to the console copy only when it is absent.
• "the div tag, and any other tag or prop the SDUI component manifest shipped in @objectstack/console does not declare": right [now] as the governed scope. The tracked sdui.manifest.json (107 components; scripts/sdui-manifest.record.json names objectuiSha = the pin) has no div; build-console.sh:241,294 copies it to packages/console/dist/sdui.manifest.json, and packages/console publishes dist. One precision an author should not over-read: an undeclared PROP draws unknown-prop at severity warning (packages/sdui-parser/src/validate.ts:141-143), which fails objectstack validate only under --strict; the entry never claims a prop fails the command, so the surface is a scope statement, not over-broad.

replacement — right, sourced from the pin.

• "box for a plain wrapper — the one drop-in swap: the same element, your className verbatim, the same children, and no layout of its own": right [now]. objectui packages/components/src/renderers/layout/box.tsx at the pin renders a div element, passes className through with zero injected classes and reads children; basic/div.tsx's own notice and its deprecated.replacement say the same words. The manifest's box declares exactly className and children.
• "Reach for card, flex, container, stack or grid only where you want their layout": right [now] — the five names in div.tsx's replacement text, all five declared in the manifest.
• "For any other tag or prop the command names, a component and prop the manifest declares; the file is dist/sdui.manifest.json inside @objectstack/console": right [now] for the file's location (files: ["dist", ...], written by build-console.sh); PR fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589's own npm pack listing confirms it ships. @objectstack/console is this repo's packages/console (objectui's app is @object-ui/console), so the package name is the right one.

reason — right in substance; three sentences are true only after #20589, one is over-broad by a condition.

• "An html page's source is a JSX string, not a keyed document: objectstack migrate meta rewrites stored metadata by key and cannot rewrite a tag inside authored source, so the move is by hand": right [now]; the precedent's own rationale.
• "objectstack validate, objectstack compile (which dev and start run first) and objectstack lint check that source against an SDUI component manifest": right [now] for the three callers (resolveJsxGateManifest is called from validate.ts, compile.ts, lint.ts). Over-broad by a condition: dev and start spawn compile only when dist/objectstack.json is missing or --compile is passed (dev.ts:319-366, start.ts:232-240); with an existing artifact they boot it and print a hint. PR fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589's changeset uses the identical phrase, so the CLI seat owns the wording; an author with a stale artifact should not read os dev as the gate. Not load-bearing for the migration.
• "the sdui.manifest.json in the directory the command runs in, then the copy @objectstack/console ships": right [now].
• "The second lookup asked for a file the console package does not export, so it always failed, and a project without its own manifest had its html pages checked at parse level only — syntax and structure, never which components and props they use": right [now] — CONSOLE_SDUI_MANIFEST_SPECIFIER = '@objectstack/console/dist/sdui.manifest.json' against an exports map of ./package.json alone; the absent answer routes to parseJsx (packages/lint/src/validate-jsx-pages.ts:63).
• "It now reaches the shipped copy": false on main when this PR lands; true only [after fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589], whose consoleSduiManifestPath() resolves @objectstack/console/package.json and joins dist/sdui.manifest.json.
• "That manifest declares the html tier's intrinsic tags but not div": right [now] — h1–h6, p, a, ul, li, blockquote, pre, img, hr, br, time, span, table, code, label, section are present; div is absent.
• "the maintainer ruled (2026-09-27) that an html page may author the intrinsic tags its renderer registers and that the published manifest declares that set, while div stays deprecated in favour of box": right [now]; comment 5852014527 on [Decision] may a kind:'html' page author intrinsic HTML tags (div, a)? The console manifest says no, 3 of 3 shipped html pages say yes, and #19922's fallback cannot go live until one side changes #20112, stamped 2026-09-27T02:37Z, letter A, in those terms.
• "the console's own html-page compile refuses div the same way": right [now] at the pin — layout/page.tsx:488 drops every type with deprecationFor(t, 'html') from the html whitelist and nameHtmlTierReplacement appends the declared replacement to the forbidden-tag error; div.tsx declares surfaces: ['json', 'html'].
• "A div in such a page, which used to pass unchecked, now fails the command with jsx-forbidden-tag and jsx-unknown-component": the rule ids are right [now] (parse.ts:96 forbidden-tag, validate.ts:96 unknown-component, both severity error, lint prefixes jsx-); the "now fails" is [after fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589].
• "A project that keeps its own sdui.manifest.json is checked against that file, as before": right [now].

acceptanceCriteria — right.

• The three rule ids and sdui/jsx-parse-level-only (JSX_PARSE_LEVEL_ONLY_RULE, sdui-manifest.ts:79) exist [now]; the notice fires only when a checked page exists, which is the criterion's reading ("a clean result beside it proves nothing").
• "objectstack compile and objectstack lint agree": right [now] — one resolver, three callers.
• "in the console each rewritten page renders with no compile error": right [now] at the pin (the box control case in objectui#10808's pin).
• "the showcase app's three html pages, which author their wrappers as box and pass with zero findings": right [now] — exactly three kind:'html' pages under examples/app-showcase/src/ui/pages/ (capability-map, command-center-jsx, start-here), no div tag in any, and packages/lint/src/validate-jsx-pages.production-witness.test.ts:115 pins the wired run clean.

Form D: the four author-shown fields carry no tracker number; the tracker ids sit in the file's header comment (and its generated copy) only. ui- is a COVERED_PREFIXES family in migrate-meta-engine-guidance.test.ts.

Changeset text (.changeset/20592-ui-html-page-div-refused.md): every sentence right [now] except the enforcement clause "so objectstack validate, compile and lint refuse a div there", which is framed as the narrowing the CLI "takes on when its JSX page gate reaches" the manifest and is true [after #20589]. "MIGRATIONS_BY_MAJOR[18].semantic gains one entry" — right; migrate meta prints it by default — right (terminus 18); spec-changes and the upgrade guide unchanged — right; "no schema accepts or refuses anything it did not before" — right.

The landing window, named: the card's Order lets the entry land first, and the dev measured PR #20589's marker flipping to not-required (already-registered ui-html-page-div-refused) plus a reason in that case. Between the two landings, migrate meta on main prints a prescription for a refusal the CLI does not yet perform; a release cut inside that window would ship the ledger row ahead of the behaviour (spec and cli are one fixed group, both 17.4.0). Sanctioned by the card and by triage 5884625733; carried to ③ as an escalation, not a fault.

② Semver level

'@objectstack/spec': patch, Clause-②: no (standalone, no arm), on both the changeset and the PR body — matches what the diff publishes. Nothing an author can write is added, removed or renamed; no export or type changes (SemanticMigration and the type of MIGRATIONS_BY_MAJOR are untouched); the package gains one data row in its ledger. The precedents are patch: .changeset/17594-step18-element-node-todo.md (a D3 TODO added to the 17→18 chain) and .changeset/20379-currency-precision-d3-text.md (two step-18 D3 entries reworded, #20488). skip-changeset would be wrong — the ledger ships in the package and migrate meta prints it. Not breaking, so no adr-0087: marker is owed; the dev's check-adr-0087-registration --base origin/main read "adds no declared-breaking changeset (1 non-breaking changeset(s) seen)", and the head's Check Changeset run is green. The feat(spec): prefix beside patch is read by no gate and the 17594 precedent used fix(spec):; cosmetic, not a level error. The narrowing itself stays where triage put it: @objectstack/cli minor, Clause-②: no (narrowing), on PR #20589.

③ Boundary flags

Dev report (os-dev-report, comment 5885694074): open_questions: []; five deviations, two out_of_scope_findings, each answered:

1. "PR body opens Closes #20592, os-dev.md names Fixes" — answered: both are closing keywords; the Part-of PR must not also close its card and The card this PR closes must claim this branch runs are green. No action.
2. "PR assignee and labels: zero writes" — answered: the seat's act. PR labels at read are documentation, size/m, tooling, needs:contract-review; the card assignee is os-justin. The seat sets the PR assignee when it posts.
3. "spec-changes.json and docs/protocol-upgrade-guide.md regenerated byte-identical, so absent from the diff" — answered and verified above (both generators stop at PROTOCOL_MAJOR = 17). The claim's "regenerated only" is satisfied with no bytes moving; the head's check:generated family runs inside Lint & Repo Gates, still running at read.
4. "Scratch worktree held two local commits (merge 07dd8aa6db, marker edit fdc65658ad), never pushed, worktree removed" — answered: the branch carries exactly two commits over the merge-base (c352c0ae20, 95ae699a5b); neither scratch sha is on it.
5. "Probe file packages/lint/.probe-20592.mjs written and removed in one command" — answered: the diff has three files and no probe.

Out-of-scope findings, both verified true here and escalated to the #19922 domain:cli seat (PR #20589's same-head review is the carrier; nothing in this PR changes):

6. PR fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589's changeset says "the html-tier renderer still renders div, deprecated there in favour of box". At the pin dd3f7e1be3 this is false: page.tsx refuses the tag at compile with a forbidden-tag error that names box. That sentence would ship false in @objectstack/cli's CHANGELOG.
7. Seat comment 5884397563 on [finding] the CLI's @objectstack/console manifest fallback can never resolve: resolveSduiManifest() asks for @objectstack/console/dist/sdui.manifest.json, but the console's exports map publishes only ./package.json #19922 says "the upgrade guide (via spec(migrations): register the ADR-0087 semantic ledger entry for the html-tier div refusal that #19922's live console fallback brings to manifest-less projects (div → box) #20592) carries the prescription meanwhile". False: the guide projects steps up to protocol 17 and does not change here; objectstack migrate meta (default --to 18) is what carries it. PR fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589's Acceptance notes should say so.

Reviewer's escalations:

8. The landing window (① above): the seat lands PR fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589 before the next release cut, or accepts that a cut inside the window ships the ledger row ahead of the refusal. If feat(spec): register the ADR-0087 semantic entry ui-html-page-div-refused (#20592) #20610 lands first, PR fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589's marker must read not-required (already-registered ui-html-page-div-refused) plus a reason, as the dev measured.
9. "objectstack compile (which dev and start run first)" is conditional in the code (auto-compile only when the artifact is missing or --compile is passed). Shared verbatim with PR fix(cli)!: the JSX page gate's console manifest fallback resolves through @objectstack/console/package.json, so a project without its own manifest gets full component checking #20589's changeset; for the CLI seat to tighten or keep.

Check-runs on 95ae699a5b, 38 raw, 34 after dedupe by name on newest started_at, read after 2026-09-29T07:35Z: 24 success, 5 skipped, 0 failure, 5 still running — Lint & Repo Gates, Test Core (1/6), Test Core (3/6), Test Core (5/6), Test Core (6/6). Green and load-bearing here: Check Changeset, Governed Surface Queue Guard, Spec property liveness, the four Type Check · runs and TypeScript Type Check, Build Core, Dogfood Regression Gate (3 shards and the rollup), Dogfood Verify CLI, Temporal Conformance (live PG + MySQL), Test Core (2/6), Test Core (4/6), the three PR-claim runs, Check Documentation Links, Flag docs affected by code changes, filter. Skipped by design: Auto Label, Build Docs, Check PR Size, Console Pin Gate (no pin move), Packed-tarball smoke (opt-in). The five running runs carry the spec test files and the repo gates (check:generated for registry.ts, check-issue-citations, check-adr-0087-registration, pnpm lint); this record's PASS is on the contract, and a red among those five supersedes it — the seat holds landing until all five complete green.

Implemented-by: claude/issue-20592-ui-html-page-div-refused
Reviewed-by: session_01Sfe5YjBLwB9J3y8fvm2xq1

VERDICT: PASS

Adopted and posted by domain:spec seat 5 (session_01Sfe5YjBLwB9J3y8fvm2xq1) · 2026-09-29T07:43Z · rendered by the seat's at-tier review subagent on this head; its served tier family was read from the subagent transcript before posting. ③ items 6 and 7 went to the #19922 seat as 5885765740 before this record. Item 8: the card's order stands (this lands first); the seat tells #19922 at landing, and a release cut is the maintainer's. Item 9 stays as written: the reviewer rates it not load-bearing, and a new head would reopen the review for one parenthetical while this week's usage reads at its warning level. Landing waits for the five running check-runs to finish green.

---

Generated by Claude Code