docs(skills): lint-rules table states both tiers of relationship/master-detail-required (#22370)

Fixes #22111

Clause-②: no

The published `objectstack-data` skill's lint-rules table said
`relationship/master-detail-required` is a `warning` everywhere. Since
the lint change that landed with the parent card (#9139, PR #22109),
that is false on one class of object. This PR changes that one table row
so both cells are true in both cases, in the table's own terse style.

## The row

Before (`skills/objectstack-data/references/lint-rules.md:13`):

```
| `relationship/master-detail-required` | warning | a `master_detail` that isn't `required` (a detail can't exist without its master) |
```

After:

```
| `relationship/master-detail-required` | warning; error under `controlled_by_parent` | a `master_detail` that isn't `required` — or, under `controlled_by_parent`, is `readonly`/`system` |
```

## What the rule does on `main` (verified in code, each fact at its
line)

- `packages/lint/src/data-model-rules.ts:256` — one rule id,
`relationship/master-detail-required`, on both tiers.
- `:863` — the tier is chosen by `obj.sharingModel ===
'controlled_by_parent'`.
- `:324-325`, `:344` — on a `controlled_by_parent` object the finding is
`severity: 'error'` whenever `required !== true`, or `readonly: true`,
or `system: true` (so `required: true` + `readonly: true` and `required:
true` + `system: true` are refused there).
- `:866-868` — on every other object only `required !== true` fires, at
`severity: 'warning'`, unchanged.
- Pinned by
`packages/lint/src/data-model-rules.master-detail-required.test.ts:52-57`
(the three unsafe shapes at `error`) and `:111-131` (`private` /
`public_read` / `public_read_write` / unset: still a `warning`;
`readonly`/`system` draw nothing there).

## Budget — the file sat at its token ceiling

`scripts/check-skills-token-ratchet.mjs` holds this file at 970 tokens
and it was at 970/970 (headroom 0), so a longer row cannot land without
deleting content in the same file. The payment is the intro clause "not
just naming/labels but the relationship/master-detail/roll-up patterns",
which duplicated the table's own heading "Data-model rules (in addition
to naming/label/i18n)"; the row also drops its parenthetical rationale.
Re-wrapped so the file keeps its line count.

| reading | before | after |
|---|---|---|
| lines | 53 | 53 (net 0) |
| bytes | 3880 | 3856 |
| tokens (`ceil(bytes/4)`, ceiling 970) | 970 (headroom 0) | 964
(headroom 6) |
| the row, bytes | 136 | 191 |

`scripts/pm/check-skill-line-ratchet.mjs` does not cover the published
`skills/` root (its header says so by design), so the token ratchet is
the only ratchet on this file; it stays green and its ceiling row is
untouched.

## Scope check inside the skill

The only other sentence in `skills/objectstack-data/**` that speaks to
this rule's severity is `rules/relationships.md:16` — "Forced only under
`controlled_by_parent`; else lint-warned". That is true as written (the
builder forces `required` under `controlled_by_parent`; the lint warning
is the other case), so it is left alone.

## Changeset

`skills/**` is in no released package's `files[]` (no `package.json`
under `packages/` names it; the catalog ships through `npx skills add`
from this repo), so this diff publishes nothing from any released
package: no `.changeset/*.md`, and the `skip-changeset` label is the
repo's skip form.

## Gates

Derived from the merge base at `6dc260aae` with `node
scripts/pm/dispatch-gates.mjs --commands --repo
objectstack-ai/objectstack` (24 families), each run with its exit code
captured before any pipe, then reconciled with `--ran`: `24 derived, 24
run, 0 NOT-MEASURED, 0 UNRUN`.

- 23 of 24 exit 0 on the first pass. `pnpm --filter @objectstack/lint
run check:doc-formula-expressions` exited 3 (PREREQUISITE NOT MET:
`@objectstack/formula` and `@objectstack/lint` not built); after the
gate's own fix line, `pnpm exec turbo run build
--filter=@objectstack/formula --filter=@objectstack/lint`, under the
verify lock (`VERDICT command-exit 0`, held 161s, waited 0s), it exits 0
— "22 record-scoped formula example(s) across 471 files / 1384 TS blocks
judged clean".
- `node scripts/check-skills-token-ratchet.mjs` exit 0 —
"skills/objectstack-data/references/lint-rules.md is 964 tokens (ceiling
970; headroom 6)"; its `--self-test` exit 0.
- `pnpm --filter @objectstack/spec run check:skill-refs` exit 0 (it
generates `references/_index.md` from the Zod map, not from this file,
so nothing to regenerate).
- `pnpm check:nul-bytes` exit 0; a control-byte scan over the edited
file finds none.
- Artifact-roster block (the 52 families the derivation scores silent
for every card): 47 exit 0. Three are PR-context guards that refuse
without `PR_NUMBER`/`PR_BODY` — `check:partof-closing-keyword` re-run
with this body exits 0; `check-closing-target-claim` and
`check-single-claim-paths` are run against the PR once it exists, result
in the report comment on the card. Two need a whole-tree `dist`
(`check:dts-closure`, `check:published-readme-exports`: exit 3,
PREREQUISITE NOT MET) — NOT MEASURED locally; a skills-only diff cannot
move either roster and CI runs both.
- `pnpm lint` (repo-level eslint) was not run locally; the diff is one
`.md` file. CI owns that run.

## Acceptance notes

- The card's landing text assumed a line ratchet on `skills/**`; there
is none — the binding gate was the token ratchet at zero headroom, which
is why a one-cell change carries an intro-clause deletion.
- Governed surface, Tier H (`skills/**`): this PR stays draft; landing
waits for an authorized approval and the seat's contract-tier review of
the skills hunk.

## 维护者速读(草稿)

- **改了什么:** 已发布的 `objectstack-data` 技能里,lint 规则表中
`relationship/master-detail-required` 这一行。原来只写 `warning`;现在写明:默认
`warning`,在 `controlled_by_parent` 对象上是 `error`,且在那里 `readonly`/`system`
的主从引用也会被拒。
- **为什么改:** `os lint` 已经按两档执行(PR #22109 落地)。技能文本还在教"到处都只是建议",按它写出的
`controlled_by_parent` 明细对象会被 `os lint` 拒绝。文件已顶到 token
上限(970/970),所以多出来的字用删掉一句与表头重复的引言来付,行数不变。
- **风险与代价(含回滚):** 纯文本改动,不碰代码、不发包、无 changeset。回滚即还原这一个文件的一次提交。
- **席位意见:**
- **你要做的:** 看一眼第 13 行这一行表述是否认可,认可就给一个批准。

---
_Generated by [Claude
Code](https://claude.ai/code/session_01RPo7FUd6bSnAfkWMAKi848)_

Co-authored-by: objectstack-fleet[bot] <noreply@anthropic.com>

Author: objectstack-fleet[bot]

skills/objectstack-data/references/lint-rules.md

@@ -1,16 +1,16 @@
 ## Linting & Generation Quality
 
-`os lint` checks the data model against the conventions in this skill —
-not just naming/labels but the relationship/master-detail/roll-up patterns. Run
-it after authoring or generating metadata. Severities: `error` (structural,
-fails the command), `warning` (likely-wrong choice), `suggestion` (nudge).
+`os lint` checks the data model against the conventions in this skill. Run it
+after authoring or generating metadata.
+Severities: `error` (structural, fails the command), `warning` (likely-wrong
+choice), `suggestion` (nudge).
 
 Data-model rules (in addition to naming/label/i18n):
 
 | Rule | Severity | Catches |
 |---|---|---|
 | `relationship/missing-reference` | error | lookup/master_detail without a `reference` target |
-| `relationship/master-detail-required` | warning | a `master_detail` that isn't `required` (a detail can't exist without its master) |
+| `relationship/master-detail-required` | warning; error under `controlled_by_parent` | a `master_detail` that isn't `required` — or, under `controlled_by_parent`, is `readonly`/`system` |
 | `relationship/delete-behavior` | suggestion | `master_detail` without an explicit `deleteBehavior` |
 | `relationship/line-items-inline-edit` | suggestion | a `*_line`/`*_item` master_detail child without `inlineEdit` |
 | `relationship/line-item-should-be-master-detail` | suggestion | a line-item-shaped child using `lookup` instead of `master_detail` |