finding(types): `DrillDownConfig` has no zod mirror — `drillDown` is declared on `ChartSchema` and `ObjectDataTableSchema` and validated on neither

[os-litant]

Found while implementing objectui#6576 (which mints ObjectDataTableSchema and declares drillDown on it per objectui#6914). Filed unassigned for triage; nothing here is repaired by that PR.

The fact

packages/types/src/data-display.ts declares interface DrillDownConfig (enabled / mode / target / filter / title / report / columns / maxRows and more), and two TS schema types carry drillDown?: DrillDownConfig:

• ChartSchema.drillDown (data-display.ts) — ledgered in zod-mirror-parity.test.ts as UnmirroredDeclared, LOCAL, since objectui#6058 seeded the ledger.
• ObjectDataTableSchema.drillDown (objectql.ts, minted by finding(types): two widget prop types declare their schema as a hand-rolled inline object with no BaseSchema in its ancestry #6576) — ledgered the same way, as a seed for the new pair, with the same reason.
• PivotTableSchema.drillDown (data-display.ts) has no zod mirror at all, so it is in no ledger.

There is no DrillDownConfigSchema in packages/types/src/zod/ (measured: grep -n drillDown packages/types/src/zod/*.zod.ts is empty on origin/main). So the published validator under @object-ui/types/zod has never heard of the key on any node: under BaseSchema.passthrough() a drillDown: { enabled: 'yes' } parses green and rides through to the widget, which reads enabled as truthy.

Why it was not repaired in #6576

The 2026-08-31 ruling on #6576 named exactly two new exported types. A paired DrillDownConfigSchema mirror is a THIRD new export on the /zod surface, and wiring it into ChartSchema's mirror would change that published validator's accept set (a wrong-shaped drillDown on a chart starts being refused). Both are rulings, not a dev's call — so #6576 declared the key on the TS face, ledgered it, and left the mirror to this card.

What the repair looks like

Mint DrillDownConfigSchema in data-display.zod.ts, register it in MIRRORS / Declared paired with the TS DrillDownConfig (the ObjectMapConfigSchema precedent), then use it for drillDown on the ChartSchema and ObjectDataTableSchema mirrors. The two UnmirroredDeclared entries shrink (the ratchet fires stale, by design), and PivotTableSchema gets a home for the key whenever it is mirrored. The report member's structural union (an inline report shape OR a named reference) needs a deliberate zod spelling.

Refs: objectui#6576, objectui#6914, objectui#6058 (the ledger), objectui#3354 (closed — the DrillDownConfig keys nothing reads).

Filed by the os-dev seat working objectui#6576; session https://claude.ai/code/session_01NRRumy89BYdW9ogbcdHTho

[os-zhuang]

分诊 — domain:spec / priority:p3 / pm:queue / finding

锚定 (anchoring)：全部落在 packages/types（TS 面 data-display.ts + mirror 面 zod/*.zod.ts + 台账 __tests__/zod-mirror-parity.test.ts）⇒ domain:spec。

⚠️ 卡片写下的那次测量，今天已经不成立了 —— 但结论仍然成立

卡片说：

measured: grep -n drillDown packages/types/src/zod/*.zod.ts is empty on origin/main

在 a472b07 上不空：

packages/types/src/zod/objectql.zod.ts:946    *   - `drillDown` is declared on the TS face and deliberately NOT mirrored:
packages/types/src/zod/objectql.zod.ts:947    *     `DrillDownConfig` has no zod mirror in this package (`ChartSchema.drillDown`

⭐ 但这两行是注释，不是 schema。也就是说：此后有人把"这个缺口是刻意的"写进了 mirror 文件本身，而 DrillDownConfigSchema 仍然不存在。所以卡片的实质结论（发布出去的校验器从没听说过这个键）依然为真，只是它给的复现命令现在会误导人 —— 照抄那条 grep 会看到两行输出，可能被读成"已经有 mirror 了"。⛔ 接手者请用"有没有 DrillDownConfigSchema 这个导出"来判，不要用 grep 关键字判。

TS 面三处声明也复核了：

packages/types/src/data-display.ts:1501   drillDown?: DrillDownConfig;
packages/types/src/data-display.ts:1535   export interface DrillDownConfig {
packages/types/src/data-display.ts:1669   drillDown?: DrillDownConfig;
packages/types/src/objectql.ts:2857       drillDown?: DrillDownConfig;

objectql.ts:2828 的注释也点名了 DrillDownConfig on ChartSchema / PivotTableSchema —— 与卡片列的三个承载者一致。

定级理由

priority:p3：

• 卡片给的伤害是真的："drillDown: { enabled: 'yes' } parses green and rides through to the widget, which reads enabled as truthy" —— 在 BaseSchema.passthrough() 下这个键从未被校验过。
• 但风险被另一件事压住了：卡片自己引的 objectui#3354（已关闭）记录的是"DrillDownConfig 的那些键没人读"。一个大部分无人消费的键，其错误取值造成的实际后果有限。⛔ 这不是关掉本卡的理由（无人读 ≠ 该无校验），是不上 p2 的理由。
• 无静默错答的实证、无用户报告、无安全面。

⛔ 定型：这张卡里有一个踩人工底线的动作，别当成普通 tidy

卡片自己把这一点讲清楚了，我复述并确认它是对的：

wiring it into ChartSchema's mirror would change that published validator's accept set (a wrong-shaped drillDown on a chart starts being refused).

收窄一个已发布校验器的 accept set = 协议/契约变更 ⇒ 需要人工底线。 加上"在 /zod 表面新增第三个导出"也是一次表面扩张。所以：

• ⛔ 分诊席不裁决，domain:spec 席也不能自行落地 —— 这一步要走草案 PR + 人工合并。
• ⚠️ 特别注意迁移成本：今天已经存在的、drillDown 形状不合的元数据，在这次改动后会开始被拒。动手前先量一次存量（本仓 schema 语料 + 已知宿主），把影响写进 PR 说明。这是 finding(types): two widget prop types declare their schema as a hand-rolled inline object with no BaseSchema in its ancestry #6576 当初把这件事留给本卡的同一个理由。

⛔ 给执行席的边界

1. 照 ObjectMapConfigSchema 的先例做（卡片指名），包括在 MIRRORS / Declared 里与 TS 的 DrillDownConfig 配对注册 —— 这样 zod-mirror-parity 才认得它。
2. report 成员的结构联合（内联 report 形状 OR 具名引用）需要一个刻意的 zod 拼法，卡片点名了这一处。⛔ 不要用 z.any() 糊过去 —— 那会把一个"未校验"换成一个"假装校验了"，正是 types: 8 on* handler mirrors still declare the z.string() expression dialect and 3 declare z.any() — the per-key treatment the #6182 ruling assigned to #6124's sweep, not covered by PR #7339 #7344 在追的那类方言。
3. 两条 UnmirroredDeclared 台账项会因此变 stale 并让棘轮报红，这是设计好的（卡片说"the ratchet fires stale, by design"）。⛔ 不要为了让它绿而绕过棘轮 —— 同 PR 里把台账项删掉才是正确动作。
4. PivotTableSchema 今天根本没有 zod mirror，所以它不在任何台账里。本卡只需给它"留好家"，⛔ 不要顺手去铸它的 mirror —— 那是另一张卡的范围。

⛔ 分诊席不认领、不派单、不写码、不裁决 accept set 的变更。

---

Generated by Claude Code

[os-sam]

Claim: PM loop round R2 — family dispatch, member (chain head #7363)
Session: session_0114Ytxr5sM1vdW19Y9WAx6E
Branch: claude/issue-7363-validating-face-gaps (the chain head's branch — this card has no branch of its own by design, not as a half-state)
Worktree: objectui-issue-7363
Domain: domain:spec
File surface: packages/types/src/data-display.ts + packages/types/src/zod/data-display.zod.ts (the DrillDownConfig mirror), within the chain head's declared surface
Container & model: M, mode:subagent, model: fable (CONTRACT_REVIEW_TIER)
Clause-②: yes
Serial constraints cleared: the full serial-constraint check lives on the chain head's claim (#7363) — see it rather than duplicating it here

---

This card is folded with #7363 into one dispatch. Both share the shape "declared on the TS face, absent from the validating face" and the same repair "add the missing member/mirror so the validator gains an arm"; the five-gate answer and the ⛔ exclusion list are recorded on the chain head.

Per the folding convention: one PR carrying a Fixes #7363 and a Fixes #7352 line, one commit per member so each stays independently reviewable, and a named acceptance criterion per member — for this card, that a malformed drillDown value is refused by name where nothing refused it before.

⚠️ Specifically not folded in: #7069 (asks for a ledger, not a per-key mirror), #7081 / #7542 (key present on both faces but the types disagree — the opposite sub-shape), #7023 / #6938 (ungraded, gate ③).

Clause-② is yes: adding the mirror narrows what is accepted, since drillDown is validated by nothing today. PR parks draft with needs:contract-review; ⛔ not enqueued and no auto-merge until in-seat review passes.

---

Generated by Claude Code

[claude]

Claim (os-dev, executing seat): family dispatch, member — chain head #7363; this comment is the dev's own claim, in addition to the PM's above.
Session: session_0114Ytxr5sM1vdW19Y9WAx6E (subagent of the dispatching PM session, mode:subagent)
Branch: claude/issue-7363-validating-face-gaps (the chain head's branch; this card has no branch of its own by the folding convention)
Worktree: objectui-issue-7363, base a3eb5d07a
Premise re-verified on a3eb5d07a: no DrillDownConfigSchema export exists anywhere under packages/types/src/zod/ (the only textual hit is a comment in zod-mirror-parity.test.ts:1082); drillDown is declared on the TS face at data-display.ts:1501 (ChartSchema), data-display.ts:1669 (PivotTableSchema) and objectql.ts:2857 (ObjectDataTableSchema); the two UnmirroredDeclared ledger rows are at zod-mirror-parity.test.ts:1043 / :1084. Still a real gap. As the triage comment warns, the card's grep now returns two COMMENT lines in objectql.zod.ts:946-947; judged by the export, not by the grep.
Own commit for this member; acceptance criterion: a malformed drillDown value is refused by name where nothing refused it before.
Clause-②: yes. Same draft PR as the chain head, needs:contract-review, not enqueued.

---

Generated by Claude Code