docs(protocol): $exists means HAS A VALUE, not key presence

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code

[os-project-manager]

Part of #13539 — deliberately not a closing keyword. This PR carries 3 of that card's 6 sites. The two published skill sites ship separately as #13577 (governed surface, human merge), and the packages/spec/src/data/filter.zod.ts JSDoc site belongs to the domain:spec seat. Merging this must not close the card while those remain.

Session, for durable attribution: https://claude.ai/code/session_01Pk26oZ12t5N1hwGW1m1MgC

What was false

Three hand-written protocol docs described $exists as a key-presence test:

• content/docs/protocol/objectql/query-syntax.mdx, operator table — `$exists` | Field exists (NoSQL)
• content/docs/protocol/objectql/query-syntax.mdx, Null Checks example — // Field exists (NoSQL)
• content/docs/protocol/kernel/http-protocol.mdx, operator list — `$exists` - Field existence check

That stopped being true when the has-value alignment landed at 9dac1ae017 (PR #13529). These are not stale wording — they are false about shipped behaviour, and the corpus does not execute, so no gate could go red on them.

What the operator actually does — established from code, not from another document

$exists asks whether the field has a value (!= null). It is the exact inverse of $null, it is portable rather than NoSQL-specific, and a stored null counts as no value on every backend:

• packages/drivers/driver-mongodb/src/mongodb-filter.ts, case '$exists' → put(value === true ? '$ne' : '$eq', null). MongoDB's own $exists is never emitted; that spelling would have been key presence and would have kept a null-valued row.
• packages/drivers/driver-sql/src/sql-driver.ts, case '$exists' → whereNotNull / whereNull.
• packages/objectql/src/having-filter.ts, case '$exists' → value !== undefined && value !== null.
• packages/spec/src/data/filter-logic-conformance.ts states the settled semantic verbatim: "$exists means "has a value" (!= null), never key-presence" — 非否定路径上的 $ne / $nin / $notContains:driver-sql 排除 NULL 行,driver-memory / formula 返回它们(#5146 只裁定了 $not) #5298 leg 3 / $exists 的比较值不是布尔时,三个后端面给出三个答案,driver-memory 自己的两个面在 'yes' 上就已分叉 —— 实测,$null(#5347)的同族另一轴 #5369, landed in PR fix(drivers,analytics,formula): $ne / $nin / $notContains 在 $not 之外也 NULL-safe (#5298) #5962.

Executed rather than merely read: 14/14 green on packages/drivers/driver-mongodb/src/mongodb-exists-has-value-translation.test.ts, which pins translateFilter({name: {$exists: true}}) to {name: {$ne: null}} and asserts that document is exactly what $null emits.

The change

Both table/list rows now read "Field has a value — the inverse of $null" with the IS NOT NULL / IS NULL compilation stated, the example comment says what the block actually selects, and a callout under the Null Checks example says plainly that this is not a key-presence test and that MongoDB never sees an $exists.

content/docs/data-modeling/queries.mdx already read "Field has a value" and is deliberately untouched — the card names it as the target wording, not as another site to sweep. A blanket find-and-replace would have broken the one line already telling the truth.

content/docs/references/data/filter.mdx is generated from the Zod JSDoc and was not hand-edited; check:docs confirms 230 generated files still in sync.

Gates — run locally at head b81efc88f

The family was derived from the real diff with node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, not recalled: 26 families, 25 green, 1 not measurable locally.

EXIT=0 :: pnpm check:doc-authoring
EXIT=0 :: pnpm check:doc-anchors
EXIT=0 :: pnpm check:docs-single-h1
EXIT=0 :: pnpm check:docs-redirects
EXIT=0 :: pnpm check:docs-audit-scope
EXIT=0 :: pnpm check:role-word
EXIT=0 :: node scripts/check-docs-section-name.mjs
EXIT=0 :: node scripts/check-doc-frontmatter.mjs
EXIT=0 :: pnpm --filter @objectstack/spec run check:docs
(230 generated files in sync with packages/spec)
EXIT=0 :: pnpm --filter @objectstack/spec run check:skill-examples
(260 prose examples type-check across 3 surfaces)
EXIT=0 :: pnpm --filter @objectstack/spec run check:liveness / check:empty-state /
check:strictness-ledger / check:variant-docs / check:yaml-examples
EXIT=0 :: pnpm --filter @objectstack/lint run check:doc-formula-expressions / check:doc-security-posture
EXIT=0 :: 6 further derived families

check:skill-examples matters here specifically: the edited Null Checks block carries an os:check marker, so its TypeScript is type-checked against the live spec — the edit is inside that checked block and it still compiles.

node scripts/check-test-completeness.mjs exits 3 = PREREQUISITE NOT MET by its own text (it grades a saved turbo run test log and none was named): recorded as NOT MEASURED, not as a red.

ESLint was not run repo-wide: its population, read from eslint.config.mjs itself, is **/*.{ts,tsx,mts,cts,js,jsx,mjs,cjs} and this diff is .mdx only, so zero of these files are in that gate's population in either direction.

No changeset: this PR releases nothing from any package, which is the repo's live convention for a pure content/docs/** change — measured, the last 20 pure-docs commits on main carry zero changesets. skip-changeset is applied.

Clause ② carrier

needs:contract-review is attached to this PR and to #13539, under the standing instruction that the card's clause-② determination is unchanged and the same write that creates a PR re-attaches both carriers. Measured honestly: this diff touches zeropackages/spec/** paths, so the path limb is not fired by these bytes — the carrier rides the card's determination, not this diff. ⛔ Not self-cleared here.

Why no gate caught it, tested rather than assumed

The card's claim is that no gate could. That holds as stated: os:check type-checks the block's TypeScript but is blind to a comment inside it, and no gate reads prose for semantic agreement with the driver it describes. It does not hold as a limit — a lexical pin over content/docs/** + skills/** (the roots check:role-word already walks) would catch exactly this class. That is a new validation surface and belongs on its own card, not smuggled into a wording repair.

Generated by Claude Code

---

Generated by Claude Code