fix(spec): refuse an absent value on a value-taking view filter operator at authoring time

[objectstack-fleet]

Fixes #19751

Clause-②: no (narrowing)

What changes

checkViewFilterRuleValueShape (the value-shape refinement of ViewFilterRuleSchema, packages/spec/src/ui/view.zod.ts) now refuses a rule with NO value on every operator that takes one. Its scalar arm returned early on value === undefined for every operator, so { field: 'name', operator: 'icontains' } parsed green, while the key's published description says every operator outside in / not_in / between and the four unary operators takes a scalar, and the query path refuses the lowered rule with 400 INVALID_FILTER.

• The four unary operators (is_empty, is_not_empty, is_null, is_not_null) are answered first and stay valueless, with or without a value.
• in / not_in / between keep their own arms, which already refused an absent value.
• The key stays .optional() on the shape; the coupling lives in the refinement, like the other arms.
• The value .describe() is unchanged (it already declares this contract), so no generated reference page moves.
• The code comment above the scalar arm, which named an absent value as a carve-out "the query path itself makes", now says the opposite and why. The docblock's "mirrors the query path" list and its runtime-wording section name the new arm.

Refusal text, one issue at the rule's value path:

Filter comparand for operator "icontains" on field "name" is undefined. The rule carries no value, and "icontains" compares the field against one — write the value to compare against, or, if the rule means the field has no value, use an operator that takes none ("is_empty" / "is_not_empty" / "is_null" / "is_not_null"), which reads its direction from its name. This is refused at authoring time because the query path refuses it too (400 INVALID_FILTER).

The leading sentence is the runtime's undefined-comparand sentence ("Filter comparand at PATH is undefined.") with the location named in the view vocabulary: operator and field, the same substitution the list and range arms already make. A view rule has no where path, and the $ spelling in that path is not one a view author can write. The unary operator names in the tail come from the schema's own VIEW_FILTER_VALUELESS_OPERATORS.

Producer reading (step 1): objectui at the pinned .objectui-sha 87af769e9a3ee28ace099fdd653d3ebd79fe82e2

Read with git show SHA:PATH from a local clone at that sha, not from a working tree.

Does the console ever save a value-taking rule with no value? No.

writer	file at the pinned sha	what happens to a half-filled row
foldFilterGroupToSpecRules, the one fold every view-filter writer shares	packages/app-shell/src/views/viewFilterFold.ts	a row whose operator takes a value is dropped when isFilterValueComplete(operator, value) is false (if (takesValue && isMissingValue(...)) continue)
isFilterValueComplete	packages/components/src/custom/filter-builder.tsx	false for value == null (also '', [], a half-filled pair), so an absent value is always incomplete
FilterBuilderField / FilterBuilderWidget: the filter-builder widget that view.form.ts names for filter and page.form.ts for filterBy, plus the per-tab filter editor	packages/app-shell/src/views/metadata-admin/widgets.tsx	calls the fold on every change; the runtime ViewConfigPanel hosts the same inspector (ViewConfigPanel.tsx, ViewVariantInspector)
list toolbar	packages/app-shell/src/views/ObjectView.tsx	no automatic write at all (its docblock: "There is deliberately NO persistViewFilter"); explicit saves go through the fold
drill-down "Save as view", foldUrlFilterTriplesToSpecRules	packages/app-shell/src/views/ObjectDataPage.tsx	ViewFilterRuleSchema.safeParse per rule, refused rules dropped; the URL triples (drillUrlFilters.ts, parseUrlFilterTriples) skip an empty param and always carry a value

One edge, stated rather than hidden: handleViewConfigSave (ObjectView.tsx) persists the config draft whole. A view whose STORED body already carries such a rule (hand-authored, or written by another tool) and is re-saved through the panel without its filter being touched now gets the refusal at save. That view already fails every query today (next table).

Does anything drop a valueless row between storage and the query? No.

layer	file	reading
console lowering: viewFilterRuleToNode, behind toFilterNode / mergeFilterNodes (plugin-list buildEffectiveFilter, plugin-view ObjectView, ObjectGrid, RelatedList, LineItemsPanel)	objectui packages/core/src/utils/filter-converter.ts	a rule without value lowers to the 2-tuple [field, operator] and nothing skips it; its own comment records the runtime throwing INVALID_FILTER / 400 for ['name','icontains']
REST lookup-picker route: lowerViewFilterRule	this repo, packages/rest/src/view-filter-rule-lowering.ts	the same 2-tuple; the module forwards and never drops
query normalizer	this repo, packages/metadata-protocol/src/protocol.ts	isFilterAST, then parseFilterAST, which throws

Measured on this tree's spec source (4112752): isFilterAST(['and', ['name','equals'], ['status','equals','open']]) is true, and parseFilterAST of it throws INVALID_FILTER / 400, "Filter comparand at where.$and[0].name is undefined". One valueless rule fails the WHOLE view's query, its good rules included.

So no working flow saves or executes this shape, and refusing it at save breaks nothing that works today.

Today's behaviour for the whole class (step 2)

Measured at origin/main 4112752 by script. The operator list is VIEW_FILTER_OPERATORS read at runtime; the unary set was derived by behaviour from the schema's own scalar arm (an array is refused on every non-list, non-range operator outside the private valueless set).

operators	ViewFilterRuleSchema, value omitted, before this change	parseFilterAST([field, op])
equals, not_equals, contains, not_contains, icontains, starts_with, ends_with, greater_than, less_than, greater_than_or_equal, less_than_or_equal, before, after (13)	ACCEPT	throws INVALID_FILTER / 400, "Filter comparand at where.name (or where.name.$op) is undefined"
in, not_in	refused by the list arm	throws, "requires an ARRAY of values"
between	refused by the range arm	throws, "requires a [min, max] value array"
is_empty, is_not_empty, is_null, is_not_null	accept	{ "$null": true } / { "$null": false }

After this change the 13 are refused. The other rows are unchanged.

ADR-0087 reading (step 4)

• This narrows a published accept set. The repo's rule for that during the launch window is in the header of scripts/check-changeset-no-major.mjs: the level does not carry breaking-ness, and "the mandatory information carriers for breaking-ness in the meantime are the BREAKING banner the author writes in the changeset body and the ADR-0087 migration-ledger disposition". scripts/check-adr-0087-registration.mjs then requires a disposition on the declared-breaking changeset.
• The disposition is registered, not not-required. The author has a hand prescription (write the value, switch to a unary operator, or delete an unfinished row), and no-migration-prescription is refused for a body that carries one. None of the other categories fits: the package publishes, no existing entry covers absence, and the surface is a schema, not a runtime interface or a type surface.
• Precedents: view-filter-rule-scalar-operator-array-refused (the sibling arm of this same check) and filter-preset-ordering-comparand-refused (a shape that never executed usefully) both registered a semantic entry under protocol major 18.
• Added: packages/spec/src/migrations/entries/semantic/18.view-filter-rule-absent-value-refused.ts. packages/spec/src/migrations/registry.ts was regenerated by pnpm --filter @objectstack/spec gen:migration-registry and not hand-edited; check:migration-registry is green. No D2 conversion: there is no value to infer.
• check-adr-0087-registration --base origin/main reads the changeset as [BREAKING+clause-②-narrowing] registered view-filter-rule-absent-value-refused (new here).
• spec-changes.json and docs/protocol-upgrade-guide.md did not move. The protocol-18 step stays inert until the protocol major reaches 18, and check:spec-changes / check:upgrade-guide are green without regeneration.

Changeset (step 7)

.changeset/19751-view-filter-rule-absent-value-refused.md, minor on @objectstack/spec. files[] ships dist and src/**/*.zod.ts, and both carry the refinement. Its summary is the user-visible change: a stored view filter rule with no value on a value-taking operator is now refused at save instead of failing every query. It carries the BREAKING banner, a FROM/TO block, Clause-②: no (narrowing) and the registered disposition marker. One sentence names that it reverses the carve-out the still-pending #19514 changeset records (an omitted value "still parses", an absent comparand "is left unjudged"), so the two entries do not contradict each other in the compiled CHANGELOG. The #19514 file itself is not edited.

Level: minor. An accept-set narrowing declared (narrowing) is BREAKING (AGENTS.md, Post-Task Checklist step 3), and during the launch window a breaking change ships as minor: the header of scripts/check-changeset-no-major.mjs says "During the launch window we ship breaking changes as minor", and that the BREAKING banner and the ADR-0087 disposition carry the break, not the level. Both precedents above shipped minor with the same banner. The first round graded this patch; the at-tier contract review (record 5808364674) failed that, and the patch-round commit ab0104d changes the frontmatter to minor and rewrites the banner sentence to state the convention ("Shipped as minor under the repo's launch-window convention for accept-set narrowings"). That commit moves no package file.

The PR's Clause-②: no (narrowing) line is the claim's, copied verbatim, and matches the changeset's line. With the arm present, the level axis of check-changeset-no-major.mjs judges the level instead of standing down. Measured offline with --event on this body, it refuses the first round's patch head 22a14a1 (exit 1) and passes minor at ab0104d (exit 0).

Fixtures, examples and pins (step 5)

• An AST scan of every tracked .ts / .tsx / .mts / .js / .mjs / .json outside content/docs/references/ (1,739 files mention operator) found 311 object literals with field plus a string-literal value-taking operator (aliases folded). 20 of them have no value key, and none is a view filter rule in a shipped example or seed:
  • 7 are QA assertions (expectedValue, a different schema) in examples/app-showcase/qa/platform-smoke.test.json;
  • 5 are QA assertions in packages/core/src/qa/runner.test.ts and packages/spec/src/qa/testing.test.ts;
  • 1 is a skill trigger condition in packages/spec/src/ai/skill-trigger-condition-value-shape.test.ts;
  • 4 are a structural walk with no schema in packages/metadata-protocol/src/protocol.graft-normalized-operators.test.ts;
  • 3 are in view-filter-rule-value-shape.test.ts.
    Markdown (.md / .mdx) has no match. No fixture was an authoring mistake, so no fixture was edited.
• Pins that pinned the removed carve-out and moved with it:
  • packages/spec/src/ui/view-filter-rule-value-shape.test.ts: equals + omitted and greater_than + omitted, from accepted to refused.
  • packages/spec/src/data/filter-icontains-parse-door.test.ts: "ABSENCE is left unjudged" now asserts that absence is refused once, in the absent-value arm's words and never in the conformance table's. ⚠️ This file is outside the claim's declared file surface. It is a view-filter-rule test that lives in src/data/, not beside view.zod.ts, and it had to move with the carve-out it pinned.
• Carriers named in the docblock (ListView.filter, a tab filter, Page.filterBy, a related-list filter, a lookup picker filter, plus ObjectGridProps.defaultFilters) are all z.array(ViewFilterRuleSchema). The full spec suite is green, and a new pin drives the refusal through ListView.filter at filter.1.value.

Tests

• New pins, with operator lists derived at runtime: value-taking is VIEW_FILTER_OPERATORS minus the four valueless operators. The valueless set is module-private in view.zod.ts and deliberately not exported, so the test reuses the file's existing transcription, and a new two-way sweep holds that transcription equal to the private set by behaviour: over every operator, an absent value is accepted exactly when the operator is valueless.
• vitest run --project local on view-filter-rule-value-shape.test.ts and filter-icontains-parse-door.test.ts: 104 passed.
• Firing control at 325052f, through scripts/ablation-replace.mjs: the anchor if (value === undefined) { was replaced by if (value === undefined) return; followed by if (false) {, which is the base behaviour (an absent value returns before any issue). The anchor went 1 to 0 and the blob b6b2f445 to 8c2839db. Result: 11 new pins red, 57 green. After the restore, the blob equals HEAD and git diff HEAD is empty. A first attempt was refused by the tool before anything ran, because its replacement contained the anchor; nothing was measured on that attempt.
• Full spec local project at 22a14a1: 522 files passed, 15,420 tests passed. One file skipped by its own stale-dist condition (scripts/root-entry-type-nameability.pin.test.ts); after a rebuild at the same head it ran with OS_EXPECT_ROOT_NAMEABILITY=1: 2 passed.
• Spec repo project at 22a14a1: 34 files, 587 tests passed.
• pnpm --filter @objectstack/spec typecheck at 22a14a1: exit 0 (tsc, scripts typecheck, check:test-typecheck OK).

Gates

node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --commands at 22a14a1 derived 86 commands. All were run and reconciled with --ran: 84 exit 0, 2 NOT MEASURED, 0 unrun. The two NOT MEASURED are check:dual-build-cjs-loads and check:type-check-debt, both exit 3 PREREQUISITE NOT MET because they need the whole-workspace build. They are left to CI. check:generated is 15/15 up to date after a fresh build at that head.

Generated files that moved: packages/spec/src/migrations/registry.ts only, via gen:migration-registry.

Patch round at ab0104d: node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --commands derived 86 commands from a tree 53 commits behind origin/main 2c1011b. origin/main's selector over the same six paths adds one family, check:migration-registry, which was run too (exit 0). The --ran reconciliation reads 84 run, all exit 0, 2 NOT MEASURED (check:dual-build-cjs-loads, check:type-check-debt: exit 3 PREREQUISITE NOT MET, they need the whole-workspace build, left to CI), and 0 unrun. check-changeset-no-major --base origin/main prints "This diff introduces no major bump." check-adr-0087-registration --base origin/main reads [BREAKING+clause-②-narrowing] registered view-filter-rule-absent-value-refused. CI at ab0104d: 35 check runs, 32 success, 3 skipped (Console Pin Gate, Build Docs, Packed-tarball smoke), 0 failure.

Scope held

• In view.zod.ts, only checkViewFilterRuleValueShape and its docblock changed. The ViewFilterRuleSchema block is untouched: its JSDoc and .describe() are still true. None of PR fix(spec): one row bound per view — retire the unpublished per-kind view limit #19809's regions is touched.
• FILTER_TEXT_CASES gains no row. There are no objectui or runtime (parseFilterAST) edits.
• origin/main has moved 8 commits past the branch point: spec(ui): declare the columns x hiddenFields x fieldOrder composition on the list view, with pins and a re-cited liveness row (ruling B) #19598 touches the ListView shape in view.zod.ts, and feat(spec)!: retire connector.connectionTimeoutMs — carried everywhere, applied nowhere #19657 touches registry.ts. A driver-less merge-tree of HEAD onto origin/main 8cbc3c0 is clean. Main is not merged in.

Acceptance notes

• The sibling entry view-filter-rule-scalar-operator-array-refused says, in its replacement prose, "An omitted value is still an omitted value". That was true of its own arm; after this change an omitted value on a scalar operator is refused. Both entries sit in the uncut protocol-18 step. The new entry's leading comment names the reversal, and the sibling's text was left as it is (it is outside this card's file surface).
• checkViewFilterRuleTextComparand's docblock, carve-out 1, says an omitted comparand "is left to whatever judges absence". That stays true: the shape arm now judges it. Not edited.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

2 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 60 of 215 client-bound route-ledger rows — the other 155 have no registrar path: tail to select them, so pages documenting THEIR client methods cannot appear above, on this or any run. Of those 155: 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; 100 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 — 136 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 2c1011b01bc071c545f72f2761647b8d9ab56375 → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 8a983b94d1df26cec68edd780f828452c2df5c7b — the merge of head ab0104d0020f8f88ecc62828ed857827e5c28873 into base 2c1011b01bc071c545f72f2761647b8d9ab56375, 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 8a983b94d1df26cec68edd780f828452c2df5c7b && git checkout 8a983b94d1df26cec68edd780f828452c2df5c7b
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin 2c1011b01bc071c545f72f2761647b8d9ab56375 ab0104d0020f8f88ecc62828ed857827e5c28873 && git checkout -B drift-repro 2c1011b01bc071c545f72f2761647b8d9ab56375 && git merge --no-ff ab0104d0020f8f88ecc62828ed857827e5c28873

node scripts/docs-audit/affected-docs.mjs --json 2c1011b01bc071c545f72f2761647b8d9ab56375

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

[objectstack-fleet]

Contract review

Served-tier: 47/47 CONTRACT_REVIEW_TIER
Head-sha: 22a14a10384056741dda7a2d8e2626901bf8fd56

Isolated at-tier reviewer subagent, run by the domain:spec seat-4 session on the maintainer's instruction in that session (「帮我处理」, with the landing route chosen there); every one of its 47 transcript turns served at the tier the constant names. Adopted by the seat 2026-09-24T05:38Z. The record below is the reviewer's, unedited except the two header lines.

① Derived judgments

Ruling applied (card #19751): triage comment 2026-09-23T00:33Z graded class (b) Bug ("declared ≠ enforced"); the dispatch comment 2026-09-23T10:26Z ordered step 1 "measure the producer first ... if it does [save a valueless row], stop and report", and step 2 "only if the producer reading is clean: refuse an absent value on value-taking operators at authoring time, with a prescription in the file's wording style. The unary operators stay valueless." Surface: view.zod.ts region checkViewFilterRuleValueShape + the ViewFilterRuleSchema block, tests beside it, regenerated files, an ADR-0087 entry if required, .changeset/.

Refs: git rev-parse origin/main = c164186; git rev-parse refs/review/pr-19861 = 22a14a1; merge-base 4112752; two commits (325052f, 22a14a1); six files, matching pulls/19861/files.

1. Scalar arm, accept-set change. Base packages/spec/src/ui/view.zod.ts:744 at 4112752 reads if (value === undefined) return; ahead of the valueless-operator check at :745. PR ref view.zod.ts:758-778: the valueless check is answered first, then if (value === undefined) adds one custom issue at path ['value'] and returns. VIEW_FILTER_OPERATORS at PR ref :254-264 lists 20 operators; minus in/not_in, between, and the four in VIEW_FILTER_VALUELESS_OPERATORS (:538-540) leaves the 13 scalar operators. Effect: 13 operators with value omitted or explicitly undefined: ACCEPT to REFUSE. Right against the ruling.
2. Unary four (is_empty/is_not_empty/is_null/is_not_null): absent value still accepted, answered before the new arm. Right ("stay valueless"). Pinned both directions at PR-ref test view-filter-rule-value-shape.test.ts:354-366.
3. in/not_in/between: no change; their arms already refused absence (pinned :315-324 with their own wording).
4. value: null on a scalar operator: unchanged, still accepted (null is not undefined; falls through to the array check). Not in the card; correct not to touch.
5. One issue per rule: checkViewFilterRuleTextComparand at PR ref :846 keeps if (value === undefined) return;, so icontains + absent yields one issue (sweep at test :293-303). Its docblock :820-823 "an omitted comparand is left to whatever judges absence" remains true.
6. Prescription wording (PR ref :770-778): leads with the runtime's sentence "Filter comparand for operator X on field Y is undefined.", names operator and field, prescribes writing the value or one of the four valueless operators (joined from VIEW_FILTER_VALUELESS_OPERATORS), closes with "400 INVALID_FILTER"; no tracker id (pinned :349-352). In the file's style, like the list/range arms.
7. Runtime untouched: no parseFilterAST, no packages/rest lowering, no FILTER_TEXT_CASES row, no objectui edit — honours the card's ⛔ on growing the conformance table.
8. Step 1 (producer) verified, not trusted. git show 4112752ec3:.objectui-sha and git show refs/review/pr-19861:.objectui-sha both print 87af769e9a3ee28ace099fdd653d3ebd79fe82e2 (the PR body's sha). At that sha in the objectui repo: packages/app-shell/src/views/viewFilterFold.ts:198-200 takesValue && isMissingValue(...) continue; packages/components/src/custom/filter-builder.tsx:952-965 isFilterValueComplete returns false on value == null; packages/app-shell/src/views/ObjectDataPage.tsx:141 ViewFilterRuleSchema.safeParse per rule. Clean, so step 2 was authorised. See ③ for main's newer pin.
9. ADR-0087: new semantic entry packages/spec/src/migrations/entries/semantic/18.view-filter-rule-absent-value-refused.ts, id view-filter-rule-absent-value-refused, surface/replacement/reason/acceptanceCriteria present, no D2 conversion (nothing to infer). registry.ts diff inserts the entry before the sibling view-filter-rule-scalar-operator-array-refused with the leading comment carried, which is what packages/spec/scripts/build-migration-registry.ts:253 (sort by major then id) and :179-183 (leading // run copied) produce. Step-18 entries do not project: git show origin/main:packages/spec/spec-changes.json | grep scalar-operator-array-refused and the same over docs/protocol-upgrade-guide.md print nothing, so no regeneration was owed.
10. Tests: the two accepted-side pins at base view-filter-rule-value-shape.test.ts:190-191 moved to the refused side; new block :270-379 sweeps every value-taking operator, the alias path, explicit undefined, and a carrier (ListView.filter at filter.1.value).
11. Fixtures: git grep at origin/main over examples/, packages/, apps/ (non-test), content/docs, docs, *.md/*.mdx for single-line scalar-operator rule literals with no value prints nothing. CI at head: commits/22a14a1038…/check-runs total_count 35, 32 success, 3 skipped (Console Pin Gate, Build Docs, Packed-tarball opt-in), 0 failure; Check Changeset (the job that runs check-adr-0087-registration.mjs, pr-automation.yml:939-940 at origin/main) success.
12. Required and missing: nothing under the ruling. Beyond the claim: one test file, ③.

② Semver level

Decision: minor. The changeset says patch.

Rule text applied:

• AGENTS.md at origin/main :1084-1085: "The declaration is Clause-②: yes|no plus at most one arm from the closed pair (widening)/(narrowing): yes takes at least minor, (narrowing) is BREAKING".
• scripts/check-changeset-no-major.mjs at origin/main, header: "During the launch window we ship breaking changes as minor" and "Until then it is NOT the carrier ... a breaking change ships as minor, so the bump level tells a consumer nothing about whether the release breaks them. The mandatory information carriers for breaking-ness in the meantime are the BREAKING banner ... and the ADR-0087 migration-ledger disposition".

The test set by the maintainer: does the change refuse input the published schema accepted before? Yes. The published surface accepted an absent value on the thirteen operators by six independent readings at 4112752 / origin/main: (a) the published type — .optional() at base view.zod.ts:894-895, unchanged, so z.input carries value?: for every operator; (b) the repo's own accept-set pin, base view-filter-rule-value-shape.test.ts:188-191, under "what stays accepted ... never stricter than the query path": "value is optional and absence is not a shape. The scalar arm must not turn an omitted comparand into a refusal"; (c) the base code comment view.zod.ts:741-744 recording absence as a carve-out "the query path itself makes"; (d) the pending changeset .changeset/19514-view-filter-rule-scalar-arm-and-icontains-comparand.md:27 at origin/main: "an omitted value still parses (value is optional)", and :35 "a view rule's value is optional, so an absent comparand is left unjudged there"; (e) the shipped ledger prose registry.ts:13340 (sibling entry, in dist through MIGRATIONS_BY_MAJOR): "An omitted value is still an omitted value"; (f) every generated reference table lists value as optional (content/docs/references/ui/view.mdx:320, page.mdx:44, component.mdx:192). Against them stands only the .describe() shape sentence "every other operator takes a scalar" and the JSDoc parenthetical at :888 "(optional for unary operators like is_empty, is_null)" — text that coexisted with (a)-(f) and never said an absent value is refused. "Already declared invalid" needs unambiguous published text; this is not that. The accept set of a published authoring surface narrows: BREAKING, carried in the launch window as minor + BREAKING banner + ADR-0087 marker.

The PR's own changeset agrees on the direction — **BREAKING** (line 13) and Clause-②: no (narrowing) (line 46), which the gate reads as [BREAKING+clause-②-narrowing] (readClause2Line: arm token (narrowing) per scripts/pm/clause2-line.mjs:286-288; signals at check-adr-0087-registration.mjs:632-640) — and only the level disagrees with the convention. The dev's reading of WHICH LEVEL (pr-automation.yml:753-760: "a fix( that changes no public surface stays patch") does not reach this act: the accept set of an exported, published schema is the public surface, and that paragraph grades widenings and names the breaking-ness carriers; it nowhere licenses patch for a declared BREAKING change. Precedent on this same check, file and key: .changeset/19514-...md:2 "@objectstack/spec": minor, body "Shipped as minor under the repo's launch-window convention for accept-set narrowings".

Marker vs scripts/check-adr-0087-registration.mjs: the changeset's HTML-comment marker adr-0087: registered view-filter-rule-absent-value-refused (line 48) matches readDisposition's registered form, names one id that resolves at HEAD and is new in the diff (entry file added), so the registered category's own check passes. Exactly one marker. Clause-②: no (narrowing) is declared (value no, arm narrowing), not the malformed no (widening) pair. Summary fix(spec): without !: the banner is the carrier; fine.

Blocking: .changeset/19751-view-filter-rule-absent-value-refused.md:2 "@objectstack/spec": patch contradicts this finding; must read minor. Line 13's sentence "The level is patch because nothing widens; during the launch window breaking-ness is carried by this banner ..." ships verbatim into CHANGELOG.md and would state the convention falsely; rewrite it (the 19514 wording is the template). The PR body's "⚠️ Level" paragraph then follows.

③ Boundary flags

• Dev-declared deviation 1 — file outside the claim: packages/spec/src/data/filter-icontains-parse-door.test.ts (claim: view.zod.ts region + tests beside it). Its base pin :204-209 asserted rule().success true for icontains + absent, exactly the removed carve-out, and would be red; the edit keeps the door's own half (refused once, in the absent-value arm's words, never the table's). Non-blocking.
• Dev-declared deviation 2 — PR body Clause-②: no (the claim's, verbatim) while the changeset carries no (narrowing). The arm is optional by grammar (clause2-line.mjs:80-85); the ADR-0087 gate reads the changeset's line, and the level axis reads the PR body's bare no and stands down. Non-blocking, but the PR-level declaration understates the direction the changeset declares.
• Prose about to ship false: pending changeset .changeset/19514-...md:27,35 at origin/main still tells consumers an omitted value "still parses (value is optional)" and an absent comparand "is left unjudged". It is unreleased, so it compiles into the same CHANGELOG.md release as this changeset and contradicts it there. .changeset/ is inside the claim's surface; one sentence in that pending file, or a sentence here naming that the carve-out finding(spec): ViewFilterRuleSchema accepts two shapes every consumer refuses, and ObjectGridProps.defaultFilters is z.unknown() so nothing gates it at all — the protocol half of objectui#9050's ruling C′ #19514's changeset recorded is reversed, closes it. Non-blocking (not yet shipped).
• Shipped ledger prose now stale: registry.ts:13340 sibling entry "An omitted value is still an omitted value". The new entry's leading comment names the reversal; this matches the repo's own precedent (the sibling recorded its reversal of view-filter-rule-value-shaped-by-operator as a new entry, not an amendment). Non-blocking.
• Pin moved under the reading: origin/main:.objectui-sha is 62597c588072636e9c30ea35b3d89b1e46fd765d (bumped by chore(objectui): bump the console pin to 62597c588072 (carries the fail-closed quick_actions requiredPermissions fix) with 46 re-measured citations, the FormField.span describe re-point and the lockstep re-record #19832 after the branch point); the PR's producer reading names 87af769e9a3e. Re-read at 62597c58: viewFilterFold.ts:198-200, filter-builder.tsx:952-965, ObjectDataPage.tsx:141 are line-identical. The pin is weight-bearing and the reading holds at both. Non-blocking.
• .describe() untouched: it still states only the shape ("takes a scalar") and never that an absent value is refused; generated reference pages keep value as optional. Not false (the unary carve-out is real), but enforcement is now stricter than the describe's explicit words, and the card's surface included the value key and docblock. Non-blocking.
• Console Pin Gate skipped by path filter (ci.yml:2337); no export removed or retyped (.optional() unchanged), so the pinned sibling's build is not at stake. Non-blocking.
• PR is draft; pulls/19861 reads mergeable_state: clean; no review threads or review comments exist; the one issue comment is the docs-drift bot.

Implemented-by: claude/issue-19751-filter-rule-absent-value
Reviewed-by: session_019c3Hi6ZMU1p6m6aA6Bz45d

VERDICT: FAIL

Blocking items:

1. Changeset level patch at .changeset/19751-view-filter-rule-absent-value-refused.md:2 contradicts ② (minor): an accept-set narrowing declared BREAKING + (narrowing) ships as minor under the launch-window convention (check-changeset-no-major.mjs header; AGENTS.md :1085; precedent .changeset/19514-...md:2). Change the frontmatter to minor and rewrite line 13's "The level is patch because nothing widens" sentence so the shipped CHANGELOG text does not misstate the convention; update the PR body's Level paragraph to match.

---

Generated by Claude Code

[objectstack-fleet]

Contract review

Served-tier: 99/99 CONTRACT_REVIEW_TIER
Head-sha: ab0104d0020f8f88ecc62828ed857827e5c28873

Isolated at-tier reviewer subagent, run by the domain:spec seat-4 session; every one of its 99 transcript turns served at the tier the constant names. Re-review after the takeover patch round (the maintainer's 「你接手派补丁轮」); it supersedes the FAIL record 5808364674. Adopted by the seat 2026-09-24T14:10Z. The record below is the reviewer's, unedited except the two header lines.

① Derived judgments

Ruling applied (card #19751): triage 5786917202 graded class (b) Bug, "the fix pulls the accept set back to what the describe already declares", Clause-② no; dispatch 5793158767 ordered step 1 (measure the producer, stop if it saves a valueless row) and step 2 (refuse an absent value on value-taking operators at authoring time, unary operators stay valueless); takeover claim 5814327629 (updated 2026-09-24T13:55:52Z) restates the original surface and confines the patch round to .changeset/19751-view-filter-rule-absent-value-refused.md, with Clause-②: no (narrowing).

Refs: git rev-parse origin/main = 2c1011b; refs/review/pr-19861 = ab0104d; merge-base 4112752; three commits (325052f, 22a14a1, ab0104d); six files, matching pulls/19861/files. git diff --numstat 22a14a1038 ab0104d002 = the changeset only, +4/-2, so every package byte is the one the prior record judged; re-verified rather than trusted:

1. Scalar arm, accept-set change. PR ref packages/spec/src/ui/view.zod.ts:769 answers the four valueless operators first, :770-781 then adds one custom issue at path ['value'] on value === undefined and returns; base :744 at 4112752 read if (value === undefined) return;. VIEW_FILTER_OPERATORS (:254-264) is 20; minus in/not_in, between and the four in VIEW_FILTER_VALUELESS_OPERATORS (:538-540) leaves the 13 scalar operators: ACCEPT to REFUSE. Right against the ruling.
2. Unary four stay valueless (answered before the arm); pinned both directions at view-filter-rule-value-shape.test.ts:356-368 (accepted exactly when in the valueless set).
3. in/not_in/between: unchanged, their own arms and wording (test :307-327).
4. value: null on a scalar operator: still accepted (null is not undefined; falls to the array check at :782). Changeset line 23 states this truthfully.
5. One issue per rule: checkViewFilterRuleTextComparand keeps if (value === undefined) return; at :846; the sweep at test :295-305 asserts exactly one value-path issue for every value-taking operator.
6. Prescription (:772-778): the runtime's leading sentence with operator and field, both repairs, the valueless names joined from the private set, closes with 400 INVALID_FILTER; no tracker id (test :351-354). In the file's style like the list and range arms.
7. Runtime, packages/rest lowering, FILTER_TEXT_CASES, objectui and .objectui-sha untouched. Honours the card's ⛔ on growing the table.
8. Step 1 (producer) verified at BOTH pins, 87af769e9a3e (branch and merge-base .objectui-sha) and 62597c5880 (current origin/main:.objectui-sha): objectui packages/app-shell/src/views/viewFilterFold.ts:198-200 drops a value-taking row when isMissingValue is true and :205 sets rule.value only when defined; :108-113 isMissingValue is the negation of isFilterValueComplete; packages/components/src/custom/filter-builder.tsx:952-965 returns false on value == null, '', []; packages/app-shell/src/views/ObjectDataPage.tsx:140-141 omits an undefined value and safeParses each rule. Line-identical at both shas. Adversarial extra: objectui's VALUELESS_FILTER_OPERATORS (viewFilterFold.ts:84-88) also holds the builder spellings exists/notExists, which have no alias in the spec's table (view.zod.ts:327-330); such a row folds to an operator the z.enum already refused before this PR, so it is not a working flow this change breaks.
9. ADR-0087: new entry packages/spec/src/migrations/entries/semantic/18.view-filter-rule-absent-value-refused.ts, id view-filter-rule-absent-value-refused, surface / replacement / reason / acceptanceCriteria present, no D2 conversion. registry.ts inserts it before the sibling view-filter-rule-scalar-operator-array-refused with the leading comment carried, which is what packages/spec/scripts/build-migration-registry.ts:253 (sort by major, then id) and :179-183/:261 (leading // run copied) produce; the builder is unchanged between the merge-base and origin/main. Neither id projects into packages/spec/spec-changes.json or docs/protocol-upgrade-guide.md at origin/main (git grep -c on both prints nothing), so no regeneration was owed.
10. Tests: the two accepted-side pins moved (test :203-210 comment records it); new block :272-381 sweeps the 13, the alias path, explicit undefined, and a carrier (ListView.filter at filter.1.value). packages/spec/src/data/filter-icontains-parse-door.test.ts:204-215 now asserts absence is refused once, in the arm's words and never the table's.
11. Fixtures at current origin/main 2c1011b (53 commits past the branch point): a git grep for a single-line { field, operator: scalar } literal with no value over examples/, apps/, packages/*/src, content/docs, docs, md/mdx/json/yml (non-test) prints nothing; the end-of-line operator: 'x' form (value on a later line) prints nothing; every one of the scalar-operator rule literals in examples/ and content/docs carries value. Nothing that landed on main emits the shape.
12. CI at head: commits/ab0104d002/check-runs total_count 42: 37 success, 5 skipped (Auto Label, Build Docs, Check PR Size, Console Pin Gate, Packed-tarball smoke opt-in; the first and third also have a success row), 0 failure. Check Changeset ran twice on this head: 107639840188 (started 12:51:43Z, after the push) and 107664890606 (started 13:56:03Z, after the 13:55:55Z body edit; pr-automation.yml:8 subscribes edited), both success. Required contexts all success.
13. Required and missing under the ruling: nothing.

② Semver level

Decision: minor. The changeset agrees: .changeset/19751-view-filter-rule-absent-value-refused.md:2 "@objectstack/spec": minor (patch diff 22a14a1..ab0104d changes patch to minor).

Rule text: AGENTS.md at origin/main :1084-1085 ("(narrowing) is BREAKING"); scripts/check-changeset-no-major.mjs header :46-48 "During the launch window we ship breaking changes as minor" and :66-72 "Until then it is NOT the carrier ... a breaking change ships as minor ... The mandatory information carriers ... are the BREAKING banner ... and the ADR-0087 migration-ledger disposition". Precedent on the same check and key: .changeset/19514-view-filter-rule-scalar-arm-and-icontains-comparand.md:2 minor, :7 "Shipped as minor under the repo's launch-window convention for accept-set narrowings".

• BREAKING banner: line 7 opens **BREAKING**, the signal breakingDeclaration reads (check-adr-0087-registration.mjs:631).
• Clause-②: no (narrowing) line: changeset line 42 Clause-②: no (narrowing) — no key is added ...; scripts/pm/clause2-line.mjs readArmToken (:281-290) takes the exact (narrowing token and leaves trailing prose alone, so it reads declared, value no, arm narrowing (signal clause-②-narrowing, adr gate :640); not the malformed no (widening). PR body line 3 reads Clause-②: no (narrowing) line-leading, and claim 5814327629 (amended 13:55:52Z) reads the same, verbatim: the dev's open question resolved as option B.
• ADR-0087 marker vs scripts/check-adr-0087-registration.mjs: the HTML-comment marker on line 44 reads adr-0087: registered view-filter-rule-absent-value-refused; readDisposition (:1923-1935) accepts registered plus one id; exactly one marker in the body; the id resolves to the entry file added in this diff (fresh). Summary fix(spec): without !: the banner carries.
• Level axis: with the arm present the axis no longer stands down (check-changeset-no-major.mjs:1640-1646: a bare no stands down UNLESS it carries the narrowing arm; :1608 offends on patch for a package whose packages/**/src/** moved; self-test :2962 a declared narrowing graded minor is clean). @objectstack/spec src moved and is graded minor, so the verdict is clean; the 13:56:03Z Check Changeset run on this head, which read the edited body, concluded success. That confirms the dev's offline --event reading (patch head exit 1, this head exit 0) without any gate re-run here.
• Rewritten convention sentence (line 7): "Shipped as minor under the repo's launch-window convention for accept-set narrowings: during the launch window a breaking change ships as minor, so the level alone does not signal the break — this banner and the ADR-0087 disposition below carry it." TRUE against header :46-48 and :66-72. The false "The level is patch because nothing widens" sentence is gone.
• Added finding(spec): ViewFilterRuleSchema accepts two shapes every consumer refuses, and ObjectGridProps.defaultFilters is z.unknown() so nothing gates it at all — the protocol half of objectui#9050's ruling C′ #19514 sentence (line 17): "its statements that an omitted value still parses (value is optional) and that an absent icontains comparand is left unjudged on a view rule no longer hold ... refused once, with the message above." TRUE: origin/main:.changeset/19514-...md:27 "an omitted value still parses (value is optional)" and :35 (in the icontains section) "a view rule's value is optional, so an absent comparand is left unjudged there"; "once" is pinned by test :295-305 and the parse-door pin; the message block sits at lines 13-15. The 19514 file is untouched: git diff origin/main refs/review/pr-19861 -- .changeset/19514-...md prints nothing, it is absent from pulls/19861/files, and it is still pending in origin/main:.changeset/.
• files[] at PR ref ships dist and src/**/*.zod.ts, so the narrowing publishes.

③ Boundary flags

Prior record's non-blocking notes, status at this head:

• Deviation 1 (filter-icontains-parse-door.test.ts outside the claim's file list): unchanged; the takeover claim restates the surface as "the view-filter-rule tests" and this is one, living in src/data/. Non-blocking, stands.
• Deviation 2 (PR body bare no under a changeset no (narrowing)): CLOSED. Claim amended, body copied verbatim, level axis armed and green on the post-edit run.
• Pending finding(spec): ViewFilterRuleSchema accepts two shapes every consumer refuses, and ObjectGridProps.defaultFilters is z.unknown() so nothing gates it at all — the protocol half of objectui#9050's ruling C′ #19514 prose contradiction: CLOSED by line 17 of this changeset; the 19514 file not edited, as the record allowed.
• Shipped ledger prose (registry.ts sibling "An omitted value is still an omitted value"): unchanged; the new entry's leading comment names the reversal (registry diff :152-157), matching the sibling's own precedent. Non-blocking, stands.
• Pin moved (origin/main:.objectui-sha 62597c5880 vs the PR's 87af769e9a3e): re-read at 62597c5880 above, the three producer sites are line-identical. Non-blocking, holds.
• .describe() and the ViewFilterRuleSchema block untouched (no hunk of the four, :573, :645, :690, :737, reaches the value key at :926-936); the describe already says every other operator takes a scalar; generated pages keep optional. Non-blocking, stands.
• Console Pin Gate skipped by path filter; no export moved. Non-blocking, stands.

New at this head, all non-blocking:

• Main is 53 commits ahead. git merge-tree --write-tree origin/main refs/review/pr-19861 is clean (exit 0). Main's view.zod.ts hunks since the merge-base are :51 (an import) and :1015 onward, none inside the check function or its docblock. Main's registry.ts gained five step-18 entries and modified seven; every new id sorts before view-, so the PR's insertion point is the builder's own; the registry is regenerated at merge per the claim, and a byte-exact regeneration was not run here (gate families are not re-run by this review). Non-blocking.
• PR body "Gates" paragraph counts "35 check runs, 32 success, 3 skipped" for this head; at review time the same head reads 42/37/5 because the edited re-run added rows; still 0 failure. Stale count, not a false one. Non-blocking.
• Line-number correction in the dev report accepted: the prior record's "line 13/46/48" were diff-output lines; the file lines are 7/42/44 at this head and its ":2" was a file line. Non-blocking.
• Wording: "the finding(spec): ViewFilterRuleSchema accepts two shapes every consumer refuses, and ObjectGridProps.defaultFilters is z.unknown() so nothing gates it at all — the protocol half of objectui#9050's ruling C′ #19514 entry" in line 17 names the pending changeset, while finding(spec): ViewFilterRuleSchema accepts two shapes every consumer refuses, and ObjectGridProps.defaultFilters is z.unknown() so nothing gates it at all — the protocol half of objectui#9050's ruling C′ #19514's ledger entry is view-filter-rule-scalar-operator-array-refused; the two quoted statements identify the changeset unambiguously, so a CHANGELOG reader is not misled. Non-blocking.
• objectui builder vocabulary exists/notExists sits in its valueless set with no spec alias; pre-existing, refused at the operator key before this PR, outside the card. Non-blocking, not this PR's.
• PR is draft; mergeable_state: clean; no review threads or review comments; card assignee os-litant; the closing-keyword and claim-branch checks are success; none of the six paths is a governed Tier H or Tier S surface.

Implemented-by: claude/issue-19751-filter-rule-absent-value
Reviewed-by: session_019c3Hi6ZMU1p6m6aA6Bz45d

VERDICT: PASS

---

Generated by Claude Code