docs(skills): objectstack-data factual sweep (1/N) — field-types, relationships, indexing (#13687)

Six measured-false behavioral claims corrected against the implementation:

- field-types.md: an unknown field key is REFUSED at parse (unrecognized_keys),
  not "silently stripped"; same for an authored vectorConfig.
- field-types.md: a wrong-cased date token ({yyyy}) parses as a {field}
  reference and renders EMPTY — it is not emitted literally.
- relationships.md: roll-up summary fields DO work over a lookup child.
- relationships.md: master_detail required is forced only under
  sharingModel: 'controlled_by_parent'; elsewhere it is a lint warning.
- indexing.md: foreign keys are NOT auto-indexed.

Token ratchet: all three files shrink (-6 / -5 / -1 = -12).

Part of #13675


Claude-Session: https://claude.ai/code/session_01EXxTW8mvPBhoHxmyPZ63de

Co-authored-by: Claude <noreply@anthropic.com>

Author: zhuangjianguo

skills/objectstack-data/rules/field-types.md

@@ -3,8 +3,8 @@
 Quick reference for choosing the right field type from 49 available options.
 
 > **Config columns list only real `FieldSchema` keys.** Per-type display knobs
-> beyond these do **not** exist — an unknown field key is silently stripped at
-> parse (dead metadata), so don't invent config like `theme`, `rows`, or
+> beyond these do **not** exist — an unknown field key is REFUSED at parse
+> (`unrecognized_keys`), so don't invent `theme`, `rows`, or
 > `fileAttachmentConfig`. Source of truth:
 > `node_modules/@objectstack/spec/src/data/field.zod.ts`.
 
@@ -343,7 +343,7 @@ The `format` is literal text interleaved with `{...}` tokens:
 1. **Every `{field}` you interpolate must be `required: true`** and set before the record is created. An empty interpolated field makes the record number generation *throw* (the compile lint flags a non-existent field as an error, an optional one as a warning).
 2. **Put a delimiter between adjacent variable tokens** — `{section}-{zone}{000}`, not `{section}{zone}{000}`. Without one, `('AB','C')` and `('A','BC')` both render prefix `ABC` and share a counter (to keep numbers unique). The literal separator keeps distinct groups apart.
 3. **Pad width is a MINIMUM, not a cap.** `{000}` → `001`…`999`, then `1000` (it grows, never wraps). Size it for readability, not as a ceiling.
-4. **Only known tokens are interpolated.** Date tokens are **case-sensitive and exact** (`{YYYY}`, not `{yyyy}` or `{YYYY-MM}`). An unrecognized `{...}` is emitted **literally** into the number — `{ YYYY }` (spaces) renders the text `{ YYYY }`.
+4. **Only known tokens are interpolated.** Date tokens are **case-sensitive** — `{yyyy}` parses as a `{field}` reference and renders **empty** (rule 1 applies). Only a spelling no field could have — `{ YYYY }`, `{YYYY-MM}` — is emitted **literally**.
 
 ### Vector (AI Embeddings)
 
@@ -354,8 +354,8 @@ The `format` is literal text interleaved with `{...}` tokens:
 }
 ```
 
-There is **no** `vectorConfig` block — an authored `vectorConfig` is silently
-stripped (dead metadata). `dimensions` is the flat field-level key.
+There is **no** `vectorConfig` block — an authored `vectorConfig` is refused at
+parse. `dimensions` is the flat field-level key.
 
 ## Incorrect vs Correct
 

skills/objectstack-data/rules/indexing.md

@@ -6,8 +6,7 @@ Guide for creating efficient database indexes in ObjectStack.
 
 ObjectStack automatically creates indexes for:
 - Primary keys (`id`)
-- Foreign keys (lookup/master_detail fields)
-- Unique constraints
+- Field-level `unique` — **not** foreign keys: declare those
 
 **Only declare non-default values.** `unique` defaults to `false` — omit it when using the default.
 
@@ -83,7 +82,7 @@ Notes an author has to know:
 
 ### ✅ Always Index
 
-1. **Foreign keys** — Automatic, but verify
+1. **Foreign keys** — declare them; never automatic
 2. **Filter fields** — Columns used in WHERE clauses
 3. **Sort fields** — Columns used in ORDER BY
 4. **Unique constraints** — Enforce uniqueness at DB level
@@ -328,7 +327,7 @@ SHOW INDEX FROM your_table;
 
 ## Best Practices
 
-1. **Index foreign keys** — Always (automatic in ObjectStack)
+1. **Index foreign keys** — always; declare each one
 2. **Composite for common queries** — Combine frequently filtered columns
 3. **Order matters** — Most selective field first
 4. **Partial for subsets** — but build it in a migration, not a declaration

skills/objectstack-data/rules/relationships.md

@@ -6,23 +6,21 @@ Guide for modeling relationships between objects using `lookup`, `master_detail`
 
 | Type | Lifecycle | Required | Sharing | Roll-ups | Use Case |
 |:-----|:----------|:---------|:--------|:---------|:---------|
-| `lookup` | Independent | Optional by default | Independent | Not available | "Related to" |
-| `master_detail` | Coupled (cascade delete) | Always required | Inherits parent | Supported via `summary` | "Owned by" |
+| `lookup` | Independent | Optional by default | Independent | Supported via `summary` | "Related to" |
+| `master_detail` | Coupled (cascade delete) | Forced only under `controlled_by_parent`; else lint-warned | Inherits parent | Supported via `summary` | "Owned by" |
 | `tree` | Self-reference | Optional | N/A | Not available | Hierarchical |
 
 ## When to Use lookup vs master_detail
 
 ### Use `lookup` When:
 - Child record can exist independently
-- Parent deletion should not affect child (`deleteBehavior: 'set_null'`)
-- No roll-up aggregations needed
+- Parent deletion should only clear the FK (`deleteBehavior: 'set_null'`, the default)
 - Relationship is optional
 - **Example:** `task.assigned_to → user` (task can exist without assignment)
 
 ### Use `master_detail` When:
 - Child record is meaningless without parent
 - Parent deletion should cascade to children
-- Need roll-up summaries (count, sum, min, max, avg)
 - Relationship is mandatory
 - **Example:** `invoice_line_item.invoice_id → invoice` (line items belong to invoice)