@ghentcdh/crouton-core 0.0.1-alpha.87 → 0.0.1-alpha.89

package/agent-docs/1.getting-started/index.md

@@ -10,7 +10,8 @@ Two ways to start with crouton: **create a new project** from scratch, or **add
 
 ## @ghentcdh/create-crouton
 
-Scaffold a complete crouton project — a NestJS backend with Prisma, pre-configured datasource, Docker Compose, and optionally an Nx monorepo layout with a Vue frontend.
+Scaffold a complete crouton project — a NestJS backend with Prisma, pre-configured datasource, Docker Compose, and
+optionally an Nx monorepo layout with a Vue frontend.
 
 ```sh
 npm create @ghentcdh/crouton my-app
@@ -40,21 +41,21 @@ node packages/create-crouton/dist/index.js my-test-app
 
 ### Flags
 
-| Flag | Default | Description |
-| --- | --- | --- |
-| `--nx` | prompt | Use Nx monorepo layout (backend + frontend + shared libs). |
-| `--no-nx` | — | Skip Nx layout (scaffold a regular single-app project). |
-| `--no-frontend` | — | Skip Vue frontend generation (Nx layout only). |
-| `--no-postgres` | — | Skip PostgreSQL setup (Prisma schema, adapter, Docker). |
-| `--sample` | — | Include a sample resource. |
-| `--pm <manager>` | prompt | Package manager: `pnpm`, `npm`, `yarn`, or `bun`. |
-| `--no-install` | — | Skip dependency installation. |
-| `--no-git` | — | Skip `git init`. |
-| `--no-docker` | — | Skip Docker files (Dockerfile, compose.yml). |
-| `--prefix <name>` | prompt | Subfolder prefix for apps/config (e.g. `split`). Implies Nx layout. |
-| `--db-url <url>` | prompt | Database connection URL. |
-| `-y, --yes` | — | Accept all defaults (non-interactive). |
-| `--force` | — | Overwrite existing files. |
+| Flag              | Default | Description                                                         |
+|-------------------|---------|---------------------------------------------------------------------|
+| `--nx`            | prompt  | Use Nx monorepo layout (backend + frontend + shared libs).          |
+| `--no-nx`         | —       | Skip Nx layout (scaffold a regular single-app project).             |
+| `--no-frontend`   | —       | Skip Vue frontend generation (Nx layout only).                      |
+| `--no-postgres`   | —       | Skip PostgreSQL setup (Prisma schema, adapter, Docker).             |
+| `--sample`        | —       | Include a sample resource.                                          |
+| `--pm <manager>`  | prompt  | Package manager: `pnpm`, `npm`, `yarn`, or `bun`.                   |
+| `--no-install`    | —       | Skip dependency installation.                                       |
+| `--no-git`        | —       | Skip `git init`.                                                    |
+| `--no-docker`     | —       | Skip Docker files (Dockerfile, compose.yml).                        |
+| `--prefix <name>` | prompt  | Subfolder prefix for apps/config (e.g. `split`). Implies Nx layout. |
+| `--db-url <url>`  | prompt  | Database connection URL.                                            |
+| `-y, --yes`       | —       | Accept all defaults (non-interactive).                              |
+| `--force`         | —       | Overwrite existing files.                                           |
 
 ### Layouts
 
@@ -113,7 +114,8 @@ my-app/
 
 #### Nx monorepo with prefix
 
-When using `--prefix split`, apps and generated code live under a subfolder. This is useful when the Nx workspace also contains non-crouton projects.
+When using `--prefix split`, apps and generated code live under a subfolder. This is useful when the Nx workspace also
+contains non-crouton projects.
 
 ```
 my-app/
@@ -138,7 +140,8 @@ The CLI runs these steps automatically (unless skipped via flags):
 
 1. **`git init`** — initialises a repository with an initial commit.
 2. **`<pm> install`** — installs all dependencies.
-3. **`crouton update resources`** — introspects the database and generates resource CRUD (best-effort; warns on failure if DB is not running).
+3. **`crouton update resources`** — introspects the database and generates resource CRUD (best-effort; warns on failure
+   if DB is not running).
 
 Then follow the printed next-steps:
 
@@ -157,7 +160,8 @@ crouton update resources --prefix split
 
 ## @ghentcdh/add-crouton
 
-Add crouton to an **existing** NestJS project — regular or Nx. This command is additive: it never overwrites existing files.
+Add crouton to an **existing** NestJS project — regular or Nx. This command is additive: it never overwrites existing
+files.
 
 ```sh
 npx @ghentcdh/add-crouton
@@ -167,32 +171,32 @@ npx @ghentcdh/add-crouton --cwd /path/to/project
 
 ### Flags
 
-| Flag | Default | Description |
-| --- | --- | --- |
-| `--cwd <dir>` | `.` | Project directory. |
-| `--backend <app>` | auto-detect | Backend app name (Nx only). |
-| `--frontend <app>` | auto-detect | Frontend app name (Nx only). |
-| `--no-frontend` | — | Skip frontend detection. |
-| `--pm <manager>` | auto-detect | Package manager. |
-| `--no-install` | — | Skip dependency installation. |
-| `--no-docker` | — | Skip Docker file generation. |
-| `-y, --yes` | — | Accept all defaults (non-interactive). |
+| Flag               | Default     | Description                            |
+|--------------------|-------------|----------------------------------------|
+| `--cwd <dir>`      | `.`         | Project directory.                     |
+| `--backend <app>`  | auto-detect | Backend app name (Nx only).            |
+| `--frontend <app>` | auto-detect | Frontend app name (Nx only).           |
+| `--no-frontend`    | —           | Skip frontend detection.               |
+| `--pm <manager>`   | auto-detect | Package manager.                       |
+| `--no-install`     | —           | Skip dependency installation.          |
+| `--no-docker`      | —           | Skip Docker file generation.           |
+| `-y, --yes`        | —           | Accept all defaults (non-interactive). |
 
 ### What it does
 
-1. **Detects project type** — checks for `nx.json` at or above the current directory to distinguish Nx workspaces from regular projects.
-2. **Discovers apps** (Nx only) — scans `apps/` subdirs, classifies them as backend (`@nestjs/core` in deps) or frontend (`vue`/`vite` in deps), and prompts which to use.
+1. **Detects project type** — checks for `nx.json` at or above the current directory to distinguish Nx workspaces from
+   regular projects.
+2. **Discovers apps** (Nx only) — scans `apps/` subdirs, classifies them as backend (`@nestjs/core` in deps) or frontend
+   (`vue`/`vite` in deps), and prompts which to use.
 3. **Detects package manager** — from lockfile presence (pnpm-lock.yaml, package-lock.json, yarn.lock, bun.lockb).
-4. **Scans dependencies** — reports missing backend deps (`@ghentcdh/crouton-api`, `@ghentcdh/crouton-core`, `prisma`, etc.).
+4. **Scans dependencies** — reports missing backend deps (`@ghentcdh/crouton-api`, `@ghentcdh/crouton-core`, `prisma`,
+   etc.).
 5. **Writes config files** (if absent):
-   - `crouton.json` — with detected `resourcesDir` and `dataSourcesDir`
-   - `crouton.enums.json` — empty enum registry
-   - `data-sources/default/` — default datasource with Prisma schema and config
-   - `.env.example` — with `DATABASE_URL` placeholder
+    - `crouton.json` — with detected `resourcesDir` and `dataSourcesDir`
+    - `crouton.enums.json` — empty enum registry
+    - `data-sources/default/` — default datasource with Prisma schema and config
+    - `.env.example` — with `DATABASE_URL` placeholder
 6. **Installs dependencies** — runs `<pm> install` if missing deps were detected.
 
-After running `@ghentcdh/add-crouton`, wire `CroutonApiModule` into your NestJS app module — see [Backend setup](backend.md).
-
-## Keeping agent files current
-
-After upgrading crouton, run `crouton migrate` to update your `resource.json` files **and** refresh the Claude skill, Cursor rule, and `AGENTS.md` block to the new version. See [Migrations](../5.tooling/migrate.md#agent-files) for details.
+After running `@ghentcdh/add-crouton`, wire `CroutonApiModule` into your NestJS app module —
+see [Backend setup](backend.md).

package/agent-docs/2.resources/actions.md

@@ -141,3 +141,43 @@ The `condition` object takes a `field`, an operator, and (except for the existen
 ## Display options
 
 Table actions support `icon`, `label`, and `tooltip`; row actions support `label`. The HTTP `method` of a procedure action defaults to `post`.
+
+## Toolbar status feedback
+
+When a procedure action runs, the CRUD toolbar shows an inline status message. No toasts are shown — feedback stays next to the table.
+
+| State | Icon | Default text |
+|-------|------|--------------|
+| In progress | ⏳ | *Processing…* |
+| Success | ✓ | Server `message`, or *Done* |
+| Error | ✗ | Server `message`, or *Action failed* |
+| Network error | ✗ | *Something went wrong, please try again* |
+
+Override the default messages per action with optional fields on a procedure action:
+
+```json
+{
+  "layout": {
+    "table": {
+      "rowActions": [
+        {
+          "id": "publish",
+          "label": "Publish",
+          "procedure": "publish",
+          "pendingMessage": "Publishing…",
+          "successMessage": "Published successfully",
+          "errorMessage": "Could not publish"
+        }
+      ]
+    }
+  }
+}
+```
+
+| Field | Purpose |
+|-------|---------|
+| `pendingMessage` | Shown immediately when the button is clicked. |
+| `successMessage` | Shown on success. Falls back to the server `message`, then *Done*. |
+| `errorMessage` | Shown when the server returns `success: false`. Falls back to the server `message`, then *Action failed*. |
+
+Success messages auto-clear after 3 s; error messages stay for 5 s.

package/agent-docs/2.resources/columns.md

@@ -16,21 +16,21 @@
 
 Both forms support the same options:
 
-| Option                                            | Description                                                                                                                                                                                                                                                                                                                 |
-|---------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
-| `idField`                                         | Marks the id column                                                                                                                                                                                                                                                                                                         |
-| `type`                                            | Data type — a shorthand (`"string"`, `"integer"`, `"boolean"`, `"date"`, …) or a JSON Schema fragment for nested shapes. Optional on a prisma resource (derived from the Zod model); **required on every column of a `kind: "custom"` resource**, see [Custom resources](custom-resource.md#column-types)                   |
-| `label` / `hideLabel`                             | Display label, or hide it                                                                                                                                                                                                                                                                                                   |
-| `hiddenInTable` / `hiddenInForm` / `hiddenInView` | Visibility per context when **no layout is defined for that context**. When `layout.<ctx>` exists it is the full list — only listed columns are rendered, and these flags are ignored. On a `oneToMany` relation column, `hiddenInTable` also drops it from the `_count` subquery `findAll` issues — that count only ever fed a table cell |
-| `createable` / `updateable`                       | Whether the field is written on create/update                                                                                                                                                                                                                                                                               |
-| `required`                                        | Whether the **form** demands a value. Overrides the Zod model in either direction on a prisma resource (`true` adds, `false` removes, omitted defers to the model); the only way to mark a field required on a `kind: "custom"` resource. Ignored on the id column and on fields that are neither createable nor updateable |
-| `displayKey`                                      | Nested field to display (e.g. `author.name`)                                                                                                                                                                                                                                                                                |
-| `showInLookup` *(deprecated)*                     | **Use `lookup.labels` at the resource root instead** (run `crouton migrate`). Sets the display label when this resource appears as an autocomplete option in another resource. Does **not** affect `?q=` search.                                                                                                             |
-| `unique`                                          | Validate the value is unique: `true`, or `{ scope, caseInsensitive, message }`. Checks as-you-type and maps a DB unique violation to this field. Needs a matching `@unique` constraint. See [Unique fields](unique.md)                                                                                                       |
-| `column`                                          | Source column name when it differs from the key/`id` (defaults to `id`)                                                                                                                                                                                                                                                    |
-| `enum`                                            | Name of a shared enum in `crouton.enums.json`; its `{ value, label }[]` is injected into `fieldInput.options.values`                                                                                                                                                                                                        |
-| `extend`                                          | Path to another `resource.json` whose columns are expanded as virtual sub-columns under this key (with a per-sub-column `columns` override map)                                                                                                                                                                             |
-| `fieldInput`                                      | Form control configuration, see below                                                                                                                                                                                                                                                                                       |
+| Option                                            | Description                                                                                                                                                                                                                                                                                                                                |
+|---------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| `idField`                                         | Marks the id column                                                                                                                                                                                                                                                                                                                        |
+| `type`                                            | Data type — a shorthand (`"string"`, `"integer"`, `"boolean"`, `"date"`, …) or a JSON Schema fragment for nested shapes. Optional on a prisma resource (derived from the Zod model); **required on every column of a `kind: "custom"` resource**, see [Custom resources](custom-resource.md#column-types)                                  |
+| `label` / `hideLabel`                             | Display label, or hide it                                                                                                                                                                                                                                                                                                                  |
+| `hiddenInTable` / `hiddenInForm` / `hiddenInView` | Visibility per context, used **only when that context has no `layout.<ctx>.items`** — they are the fallback that builds the layout. As soon as `layout.<ctx>` lists field items, those items are the full list: only listed columns render and these flags are ignored. On a `oneToMany` relation column, `hiddenInTable` (when acting as the table fallback) also drops it from the `_count` subquery `findAll` issues |
+| `createable` / `updateable`                       | Whether the field is written on create/update                                                                                                                                                                                                                                                                                              |
+| `required`                                        | Whether the **form** demands a value. Overrides the Zod model in either direction on a prisma resource (`true` adds, `false` removes, omitted defers to the model); the only way to mark a field required on a `kind: "custom"` resource. Ignored on the id column and on fields that are neither createable nor updateable                |
+| `displayKey`                                      | Nested field to display (e.g. `author.name`)                                                                                                                                                                                                                                                                                               |
+| `showInLookup` *(deprecated)*                     | **Use `lookup.labels` at the resource root instead**. Sets the display label when this resource appears as an autocomplete option in another resource. Does **not** affect `?q=` search.                                                                                                                                                   |
+| `unique`                                          | Validate the value is unique: `true`, or `{ scope, caseInsensitive, message }`. Checks as-you-type and maps a DB unique violation to this field. Needs a matching `@unique` constraint. See [Unique fields](unique.md)                                                                                                                     |
+| `column`                                          | Source column name when it differs from the key/`id` (defaults to `id`)                                                                                                                                                                                                                                                                    |
+| `enum`                                            | Name of a shared enum in `crouton.enums.json`; its `{ value, label }[]` is injected into `fieldInput.options.values`                                                                                                                                                                                                                       |
+| `extend`                                          | Path to another `resource.json` whose columns are expanded as virtual sub-columns under this key (with a per-sub-column `columns` override map)                                                                                                                                                                                            |
+| `fieldInput`                                      | Form control configuration, see below                                                                                                                                                                                                                                                                                                      |
 
 ::: warning Moved to `layout.table`
 The following column-level options are **deprecated** and will be removed. Use `layout.table` instead:
@@ -42,7 +42,8 @@ The following column-level options are **deprecated** and will be removed. Use `
 - `showWhen` / `hideWhen` / `disabledWhen` → `layout.table.items[{ id, showWhen, ... }]`
 - `fieldTable` → `layout.table.items[{ id, type, options }]`
 
-See [Layout — Table layout](layout.md#table-layout-layouttable) and [migrate-to-next-alpha.md](/migrate-to-next-alpha) for migration details.
+See [Layout — Table layout](layout.md#table-layout-layouttable) and [migrate-to-next-alpha.md](/migrate-to-next-alpha)
+for migration details.
 :::
 
 ## Free-text search (`?q=`)
@@ -53,16 +54,23 @@ Declare which columns are included in free-text search via `operations.findAll.s
 {
   "operations": {
     "findAll": {
-      "search": ["title", "authorId"]
+      "search": [
+        "title",
+        "authorId"
+      ]
     }
   }
 }
 ```
 
-- **Multiple entries** are ORed together: `?q=tolkien` matches books whose `title` contains "tolkien" OR whose author name contains "tolkien".
-- **`manyToOne` relation columns** (e.g. `authorId`) automatically resolve to the related resource's first visible display field (e.g. `author.name`), so you never search against a raw foreign key string.
-- **`showInLookup`** is a separate concern — it controls the label shown when this resource appears as an autocomplete option somewhere else. Do not use it to drive `?q=` search; use `operations.findAll.search` instead.
-- **Default**: when `findAll: true` and no explicit `search` is declared, all simple (non-id, non-relation, non-date) columns are searched.
+- **Multiple entries** are ORed together: `?q=tolkien` matches books whose `title` contains "tolkien" OR whose author
+  name contains "tolkien".
+- **`manyToOne` relation columns** (e.g. `authorId`) automatically resolve to the related resource's first visible
+  display field (e.g. `author.name`), so you never search against a raw foreign key string.
+- **`showInLookup`** is a separate concern — it controls the label shown when this resource appears as an autocomplete
+  option somewhere else. Do not use it to drive `?q=` search; use `operations.findAll.search` instead.
+- **Default**: when `findAll: true` and no explicit `search` is declared, all simple (non-id, non-relation, non-date)
+  columns are searched.
 
 ## Nested object and array columns
 
@@ -75,15 +83,21 @@ A column's `type` may be a full JSON Schema fragment, which is how you describe
     "type": {
       "type": "object",
       "properties": {
-        "id": { "type": "string" },
-        "name": { "type": "string" }
+        "id": {
+          "type": "string"
+        },
+        "name": {
+          "type": "string"
+        }
       }
     }
   },
   "tags": {
     "type": {
       "type": "array",
-      "items": { "type": "string" }
+      "items": {
+        "type": "string"
+      }
     }
   }
 }
@@ -129,10 +143,10 @@ form-open time, not persisted until the user saves.
 
 #### Dynamic default tokens
 
-| Token      | Resolves to                                              |
-|------------|----------------------------------------------------------|
-| `"$now"`   | Current datetime as an ISO 8601 string                   |
-| `"$today"` | Current date as an ISO 8601 date string (`"2025-03-14"`) |
+| Token      | Resolves to                                                         |
+|------------|---------------------------------------------------------------------|
+| `"$now"`   | Current datetime as an ISO 8601 string                              |
+| `"$today"` | Current date as an ISO 8601 date string (`"2025-03-14"`)            |
 | `"$user"`  | Current user object (requires `defaults` in `CroutonPlugin` config) |
 
 **Using `$user`** requires passing the current user when setting up the plugin:
@@ -163,7 +177,8 @@ onMounted(async () => {
 
 ## Field variants — per-context overrides
 
-A single `fieldInput` drives the form, the read-only view, and the table cell. To render differently in a specific context, use layout item overrides:
+A single `fieldInput` drives the form, the read-only view, and the table cell. To render differently in a specific
+context, use layout item overrides:
 
 - **form** ← `fieldInput`
 - **view** ← `fieldInput` (override per-column via `layout.view.items[{ id, type, options }]`)
@@ -178,19 +193,32 @@ A single `fieldInput` drives the form, the read-only view, and the table cell. T
       "fieldInput": {
         "format": "relation",
         "resource": "./author/resource.json",
-        "options": { "display": "autocomplete", "displayKey": "name" }
+        "options": {
+          "display": "autocomplete",
+          "displayKey": "name"
+        }
       }
     }
   },
   "layout": {
     "view": {
       "items": [
-        { "id": "author", "options": { "display": "link" } }
+        {
+          "id": "author",
+          "options": {
+            "display": "link"
+          }
+        }
       ]
     },
     "table": {
       "items": [
-        { "id": "author", "options": { "displayKey": "shortName" } }
+        {
+          "id": "author",
+          "options": {
+            "displayKey": "shortName"
+          }
+        }
       ]
     }
   }
@@ -208,7 +236,12 @@ Add `customComponent` to the table layout item's `options` to override the cell
   "layout": {
     "table": {
       "items": [
-        { "id": "section_number", "options": { "customComponent": "moveUpDown" } }
+        {
+          "id": "section_number",
+          "options": {
+            "customComponent": "moveUpDown"
+          }
+        }
       ]
     }
   }
@@ -236,12 +269,15 @@ Read-only columns derived from a raw SQL subquery at query time. Add a `computed
 }
 ```
 
-| Field        | Type     | Description                                                                 |
-|--------------|----------|-----------------------------------------------------------------------------|
-| `sql`        | `string` | Raw SQL subquery. Use `main` as the alias for the parent table row.         |
-| `type`       | `string` | Cast type for the result: `"number"` (default), `"string"`, or `"boolean"`. |
+| Field  | Type     | Description                                                                 |
+|--------|----------|-----------------------------------------------------------------------------|
+| `sql`  | `string` | Raw SQL subquery. Use `main` as the alias for the parent table row.         |
+| `type` | `string` | Cast type for the result: `"number"` (default), `"string"`, or `"boolean"`. |
 
-All standard column fields (`label`, `hiddenInTable`, `hiddenInView`, `position`, `fieldInput`, …) are valid alongside `computed`.
+All standard column fields (`label`, `hiddenInTable`, `hiddenInView`, `position`, `fieldInput`, …) are valid alongside
+`computed`. When a context declares a layout, list the computed column's id in `layout.table.items` /
+`layout.view.items` to place it there (it is hidden otherwise); with no layout it follows `hiddenInTable` /
+`hiddenInView` as before.
 
 ::: warning
 Computed columns are not available on `kind: "custom"` resources — they run raw SQL against a real table.
@@ -249,12 +285,18 @@ Compute the value in `findAll` instead.
 :::
 
 ::: warning Deprecated: `calculatedColumns`
-The top-level `calculatedColumns` array is removed. Run `crouton migrate` to move each entry to `columns[id].computed`.
+The top-level `calculatedColumns` array is removed — computed columns live on
+`columns[id].computed`. Legacy top-level entries no longer accept `hiddenInTable` /
+`hiddenInView`; without an explicit layout such a column is always rendered.
 :::
 
 ## Includes
 
-Eagerly load relations with the list/detail queries. Declare them per operation so `findAll` and `findOne` can differ:
+Eagerly load relations with the list/detail queries. Declare them per operation so `findAll` and `findOne` can differ.
+
+### Dot-path strings (recommended for simple nesting)
+
+Use `"a.b"` shorthand instead of the verbose object form when you don't need `orderBy`:
 
 ```json
 {
@@ -262,13 +304,43 @@ Eagerly load relations with the list/detail queries. Declare them per operation
     "findAll": {
       "include": [
         "author",
-        { "relation": "chapters", "orderBy": { "position": "asc" }, "include": ["sections"] }
+        "chapters.sections"
       ]
-    },
-    "findOne": {
+    }
+  }
+}
+```
+
+Multiple dot-path entries for the same top-level relation are **deep-merged**:
+
+```json
+"include": ["a.b", "a.c"]
+// → { a: { include: { b: true, c: true } } }
+```
+
+Mix dot paths and object form freely — they merge on the same relation:
+
+```json
+"include": [
+  "chapters.sections",
+  { "relation": "chapters", "orderBy": { "position": "asc" } }
+]
+// → chapters included with sections nested + orderBy applied
+```
+
+### Object form (when you need options)
+
+```json
+{
+  "operations": {
+    "findAll": {
       "include": [
         "author",
-        { "relation": "chapters", "orderBy": { "position": "asc" }, "include": ["sections"] }
+        {
+          "relation": "chapters",
+          "orderBy": { "position": "asc" },
+          "include": ["sections"]
+        }
       ]
     }
   }
@@ -277,15 +349,19 @@ Eagerly load relations with the list/detail queries. Declare them per operation
 
 When `findOne.include` is absent, crouton falls back to `findAll.include`.
 
-| Field      | Type             | Description                                                        |
-|------------|------------------|--------------------------------------------------------------------|
-| `relation` | `string`         | Prisma relation field name                                         |
-| `orderBy`  | `object`         | Optional sort: `{ field: 'asc' \| 'desc' }` for included records  |
-| `include`  | `array`          | Nested includes                                                    |
+| Field      | Type     | Description                                                       |
+|------------|----------|-------------------------------------------------------------------|
+| `relation` | `string` | Prisma relation field name                                        |
+| `orderBy`  | `object` | Optional sort: `{ field: 'asc' \| 'desc' }` for included records |
+| `include`  | `array`  | Nested includes (string, dot-path, or object)                     |
 
 When a relation column has a `sort` option (see [Relations](relations.md#sorting-related-records)), the loader
 automatically injects the corresponding `orderBy` into the include clause — no manual configuration needed.
 
+`?include=` is limited to relation columns the **table layout** renders (via `layout.table.items`); without a
+table layout it falls back to the `hiddenInTable` column flag. A relation the table never shows cannot be included.
+
 ::: warning Deprecated: root `include`
-The top-level `include` field is removed. Run `crouton migrate` to move it to `operations.findAll.include` / `operations.findOne.include`.
+The top-level `include` field is removed. Move it to `operations.findAll.include` /
+`operations.findOne.include`.
 :::

package/agent-docs/2.resources/field-inputs/autocomplete.md

@@ -39,4 +39,30 @@ Autocomplete with optional free-text and create. Three option shapes are support
 }
 ```
 
+## Excluding already-selected values
+
+Use `exclude` to hide values the user already chose — useful for many-to-many chip lists or sibling field constraints.
+
+```jsonc
+"fieldInput": {
+  "type": "autocomplete",
+  "options": {
+    "resource": "./author.resource",
+    "labelKey": "name",
+    "valueKey": "id",
+    "exclude": {
+      "values": "{form.coAuthors}",   // {form.path} placeholder or a literal array
+      "key": "id",                     // filter field — defaults to valueKey
+      "selected": true                 // also exclude this field's own current value
+    }
+  }
+}
+```
+
+- **`values`** — a `{form.path}` placeholder (resolved at fetch time) or a literal `unknown[]`. Objects in the array are mapped through `valueKey` to extract the scalar.
+- **`key`** — the server-side field used for the `not_in` filter. Defaults to `valueKey` (or `"id"`).
+- **`selected`** — when `true`, the control's own current value is also excluded. Automatically applied by `RelationInline` for many-to-many chips.
+
+For resource-backed autocomplete the exclude list is sent as a `filter=<key>:[…]:not_in` query param (skipped above 200 values). For remote URIs, add `{exclude}` to the URI template. Client-side filtering always runs as a fallback.
+
 <!-- @include: ./_generated/autocomplete.md -->