docs: add number type, fix n8n node structure, correct SDK ordering

Cowork 3P · Cowork 3P · commit c33c57b1ae69 · 2026-06-06T14:00:23.000+08:00
diff --git a/src/content/docs/developer-guide/worker-definition/input-schema.md b/src/content/docs/developer-guide/worker-definition/input-schema.md
@@ -327,3 +327,4 @@ Developers can logically group multiple configuration items by using specific fi
 1. **Write clear descriptions**: Make sure `description` is clear and accurate. This helps your script get discovered by more target users.
 2. **Set sensible defaults**: A reasonable `default` value lets users run the script immediately and greatly lowers the barrier to entry.
 3. **Validate required fields**: For parameters without which the script cannot run, such as login cookies or the main URL, be sure to set `"required": true`.
+4. **Max results naming**: If your Worker accepts a maximum-results parameter, use the field name `max_results`. This is the conventional name recognized by the platform and downstream integrations.
diff --git a/src/content/docs/developer-guide/worker-definition/output-schema.md b/src/content/docs/developer-guide/worker-definition/output-schema.md
@@ -44,7 +44,7 @@ Each column definition contains the following properties:
 | Property       | Required | Description                                                                                   |
 | -------------- | -------- | --------------------------------------------------------------------------------------------- |
 | **name**       | Yes      | Column identifier. Must match the key name used in `push_data`. Must be unique.              |
-| **type**       | Yes      | Data type of the column. Supported values: `string`, `integer`, `boolean`, `array`, `object` |
+| **type**       | Yes      | Data type of the column. Supported values: `string`, `number`, `integer`, `boolean`, `array`, `object` |
 | **description**| No       | Column description. Displayed as the column header label in the UI.                          |
 
 ## Relationship with SDK
diff --git a/src/content/docs/developer-guide/worker-definition/sdk-modules.md b/src/content/docs/developer-guide/worker-definition/sdk-modules.md
@@ -185,6 +185,7 @@ res, err := coresdk.Result.SetTableHeader(ctx, headers)
 * **key**: Unique identifier used in code (recommended lowercase with underscores)
 * **format**: Data type, supported values:
   * `"text"`: String / text
+  * `"number"`: Number (integer or float)
   * `"integer"`: Integer
   * `"boolean"`: Boolean (true / false)
   * `"array"`: List / array
@@ -237,7 +238,7 @@ for i, news := range newsData {
 
 ##### Important Notes
 
-* Header definition and data pushing **can be done in either order**
+* Header definition **must be done before** the first data push
 * Data keys **must exactly match** the header keys (case-sensitive)
 * Data must be pushed **one record at a time**
 * Logging after each push is recommended
diff --git a/src/content/docs/integrations/workflows-and-notifications/n8n.md b/src/content/docs/integrations/workflows-and-notifications/n8n.md
@@ -9,12 +9,12 @@ Use [n8n](https://n8n.io/) to build automated workflows that trigger CoreClaw Wo
 
 ## How it works
 
-The CoreClaw n8n integration provides a dedicated community node (`n8n-nodes-coreclaw`) with built-in actions for common operations:
+The CoreClaw n8n integration provides a dedicated community node (`n8n-nodes-coreclaw`) with four resources and built-in actions for common operations:
 
-- **Start a Worker** — Run any Worker from the CoreClaw Store
-- **Check run status** — Poll execution progress
-- **Get results** — Retrieve output data
-- **Abort a run** — Cancel an in-progress execution
+- **Scraper** — Search the marketplace, get scraper details, and run a scraper
+- **Run** — Get status, get results, export results, get logs, abort, or rerun an execution
+- **Task** — Run a pre-configured saved task
+- **Account** — Get account info (balance, traffic, plan)
 
 You can also use the **HTTP Request** node to call the CoreClaw REST API directly for advanced use cases.
 
@@ -36,15 +36,15 @@ For n8n Cloud users, installation is straightforward — search and add the node
 2. Open the **nodes panel** (click the **+** button on the canvas).
 3. Search for **CoreClaw** in the community node registry.
 
-Search for CoreClaw node in n8n
+![Search for CoreClaw node in n8n](@/assets/docs/n8n-1.png)
 
-1. Click **Install node** to add the CoreClaw node to your instance.
+4. Click **Install node** to add the CoreClaw node to your instance.
 
-Install n8n-nodes-coreclaw package
+![Install n8n-nodes-coreclaw package](@/assets/docs/n8n-2.png)
 
 After installation, you can find the CoreClaw node under **Community Nodes** in the nodes panel.
 
-CoreClaw node in Community Nodes list
+![CoreClaw node in Community Nodes list](@/assets/docs/n8n-3.png)
 
 ### Create credentials
 
@@ -53,14 +53,14 @@ Before using the CoreClaw node, you need to create a credential with your CoreCl
 1. In n8n, go to **Credentials** → **Add Credential**.
 2. Search for **CoreClaw API** and select it.
 
-Create CoreClaw API credential
+![Create CoreClaw API credential](@/assets/docs/n8n-4.png)
 
-1. Enter a name for the credential (e.g., "CoreClaw Production").
-2. In the **API Key** field, paste the API key you copied from the [CoreClaw Console](https://console.coreclaw.com/settings/integrations).
+3. Enter a name for the credential (e.g., "CoreClaw Production").
+4. In the **API Key** field, paste the API key you copied from the [CoreClaw Console](https://console.coreclaw.com/settings/integrations).
 
-Enter API key and save credential
+![Enter API key and save credential](@/assets/docs/n8n-5.png)
 
-1. Click **Save** to store the credential.
+5. Click **Save** to store the credential.
 
 You can now use this credential in any CoreClaw node in your workflows.
 
@@ -87,59 +87,130 @@ Follow the same steps as [Create credentials](#create-credentials) above to set
 
 ## CoreClaw node actions
 
-The CoreClaw node provides several built-in actions for common operations.
+The CoreClaw node is organized by **resource** (Scraper, Run, Task, Account). Select a resource, then choose the operation you want to perform.
 
-### Start a Worker
+### Scraper resource
 
-Run any Worker from the CoreClaw Store.
+#### Search
 
+Search the CoreClaw marketplace for ready-to-run scrapers by keyword.
+
+| Field   | Description                                                            |
+| ------- | ---------------------------------------------------------------------- |
+| **Query** | Keyword matched against scraper title / description / tags             |
+| **Limit** | Max number of results to return (1–100, default: 50)                   |
+
+#### Get Details
+
+Fetch the full spec of a scraper: current version, system defaults, custom input schema, and README.
+
+| Field          | Description                                                            |
+| -------------- | ---------------------------------------------------------------------- |
+| **Scraper**    | Pick from the marketplace list, or paste a slug directly               |
+
+#### Run
+
+Start an asynchronous scraper run with custom parameters.
 
 | Field                | Description                                                              |
 | -------------------- | ------------------------------------------------------------------------ |
-| **Worker Slug**      | The unique identifier of the Worker (e.g., `01KNXSHE0Y7DZKF1N8B1EMFX35`) |
-| **Version**          | The Worker version to run (leave empty for latest)                       |
-| **Input Parameters** | Worker-specific input fields (varies by Worker)                          |
-| **Proxy Region**     | Optional proxy region for the run                                        |
+| **Scraper**          | Pick from the marketplace list, or paste a slug directly                 |
+| **Version**          | Scraper version string (required). Obtain from **Get Details** → version |
+| **Custom Parameters**| Scraper-specific input parameters as JSON (schema from Get Details)      |
+| **System Parameters**| Optional JSON overrides for cpus, memory, timeout, max charge, traffic  |
 | **Callback URL**     | Optional webhook URL for async notifications                             |
 
+:::caution
+**Version is required.** The node does not support "leave empty for latest". Always obtain the correct version string from **Get Details** first.
+:::
 
 Get the `scraper_slug` (Worker Slug) from the Worker page in the [CoreClaw Console](https://console.coreclaw.com/store) or from the API (`GET /api/scraper?slug=<scraper_slug>`).
 
-### Check Run Status
+### Run resource
 
-Poll the status of a running or completed Worker execution.
+#### Get
 
+Get the current execution status of a run (status, started_at, duration, cost).
 
 | Field        | Description                                        |
 | ------------ | -------------------------------------------------- |
-| **Run Slug** | The run identifier returned when starting a Worker |
-
+| **Run Slug** | The run identifier returned when starting a scraper or task |
 
 Status codes: `1` Ready, `2` Running, `3` Succeeded, `4` Failed, `5` Aborting.
 
-### Get Results
+#### Get Many
+
+List the user's historical scraper runs with pagination and filters.
+
+| Field          | Description                                         |
+| -------------- | --------------------------------------------------- |
+| **Return All** | Whether to return all results or only up to a limit |
+| **Limit**      | Max number of results to return (1–200, default: 50) |
+| **Filters**    | Filter by status and/or scraper slug                |
 
-Retrieve the output data from a completed run.
+#### Get Results
 
+Get paginated result records from a completed run.
 
-| Field          | Description                              |
-| -------------- | ---------------------------------------- |
-| **Run Slug**   | The run identifier                       |
-| **Page Index** | Page number for pagination (default: 1)  |
-| **Page Size**  | Number of results per page (default: 20) |
+| Field          | Description                                         |
+| -------------- | --------------------------------------------------- |
+| **Run Slug**   | The run identifier                                  |
+| **Return All** | Whether to return all results or only up to a limit |
+| **Limit**      | Max number of results to return (1–500, default: 50)|
 
+#### Export Results
 
-For large datasets, use the **Export Results** action to download a JSON or CSV file.
+Export a run's full result set as a downloadable CSV or JSON file.
 
-### Abort a Run
+| Field          | Description                                                            |
+| -------------- | ---------------------------------------------------------------------- |
+| **Run Slug**   | The run identifier                                                   |
+| **Format**     | `csv` (human-readable, opens in Excel) or `json` (preserves nesting)   |
+| **Filter Keys**| Comma-separated field keys to include. Leave empty for all fields.     |
 
-Cancel an in-progress Worker execution.
+#### Get Logs
 
+Fetch execution logs from a run for debugging or understanding failures.
+
+| Field        | Description                 |
+| ------------ | --------------------------- |
+| **Run Slug** | The run identifier          |
+
+#### Abort
+
+Cancel an in-progress scraper run.
 
 | Field        | Description                 |
 | ------------ | --------------------------- |
 | **Run Slug** | The run identifier to abort |
 
+#### Rerun
+
+Re-run a previous run with the exact same parameters.
+
+| Field          | Description                                                            |
+| -------------- | ---------------------------------------------------------------------- |
+| **Run Slug**   | The run identifier to rerun                                          |
+| **Callback URL**| Optional webhook URL for async notifications                           |
+
+### Task resource
+
+#### Run
+
+Run a pre-configured saved task from the CoreClaw console. Task parameters are stored with the task itself, so no custom input is needed.
+
+| Field          | Description                                                            |
+| -------------- | ---------------------------------------------------------------------- |
+| **Task Slug**  | Saved task identifier from the CoreClaw Console → Tasks page           |
+| **Callback URL**| Optional webhook URL for async notifications                           |
+
+### Account resource
+
+#### Get Info
+
+Get account info: balance, traffic usage, and plan expiry.
+
+No parameters required.
 
 ---
 
@@ -148,13 +219,13 @@ Cancel an in-progress Worker execution.
 Here's a typical n8n workflow using CoreClaw:
 
 1. **Trigger** — Schedule Trigger (e.g., every day at 9 AM) or Webhook
-2. **CoreClaw: Start a Worker** — Run a web scraper with target URLs
+2. **CoreClaw: Scraper → Run** — Run a web scraper with target URLs
 3. **Wait** — Wait 30 seconds for the run to progress
-4. **CoreClaw: Check Run Status** — Poll until status is `3` (Succeeded)
+4. **CoreClaw: Run → Get** — Poll until status is `3` (Succeeded)
 5. **IF** — Check if status equals `3`
   - **True** → Continue to get results
   - **False** → Loop back to Wait
-6. **CoreClaw: Get Results** — Retrieve the scraped data
+6. **CoreClaw: Run → Get Results** — Retrieve the scraped data
 7. **Downstream nodes** — Send to Google Sheets, Slack, database, etc.
 
 ---
@@ -165,7 +236,6 @@ For operations not covered by the CoreClaw node, use the **HTTP Request** node t
 
 ### Configuration
 
-
 | Field             | Value                                            |
 | ----------------- | ------------------------------------------------ |
 | Method            | `POST` (most endpoints)                          |
@@ -175,10 +245,8 @@ For operations not covered by the CoreClaw node, use the **HTTP Request** node t
 | Header Value      | Your CoreClaw API key                            |
 | Body Content Type | `JSON`                                           |
 
-
 ### Common endpoints
 
-
 | Action                  | Method | Endpoint                           |
 | ----------------------- | ------ | ---------------------------------- |
 | Get Worker schema       | `GET`  | `/api/scraper?slug=<scraper_slug>` |
@@ -189,7 +257,6 @@ For operations not covered by the CoreClaw node, use the **HTTP Request** node t
 | Export results (file)   | `POST` | `/api/v1/run/result/export`        |
 | Abort a run             | `POST` | `/api/v1/scraper/abort`            |
 
-
 Full API reference: [API Integration](/api/integration/).
 
 ---
@@ -212,8 +279,6 @@ Full API reference: [API Integration](/api/integration/).
 2. Check **Settings → Community Nodes** — the node should be listed there.
 3. If using n8n Cloud, ensure verified community nodes are enabled in the Cloud Admin Panel.
 
-
-
 **Invalid API key error**
 
 1. Verify the API key in the [CoreClaw Console](https://console.coreclaw.com/settings/integrations).
@@ -229,8 +294,6 @@ curl -X POST "https://openapi.coreclaw.com/api/v1/account/info" \
 
 A successful response contains `code: 0`.
 
-
-
 **Worker-specific input fields**
 
 Each Worker has different input parameters. To find the correct fields:
@@ -246,5 +309,4 @@ Or call the API:
 curl "https://openapi.coreclaw.com/api/scraper?slug=YOUR_SCRAPER_SLUG"
 ```
 
-The response contains `data.parameters.custom.properties` — each entry maps to an input field.
-
+The response contains `data.parameters.custom.properties` — each entry maps to an input field.
diff --git a/src/content/docs/zh-cn/developer-guide/worker-definition/input-schema.md b/src/content/docs/zh-cn/developer-guide/worker-definition/input-schema.md
@@ -329,3 +329,4 @@ sidebar:
 1. **清晰的提示**：description 务必清晰准确，这将有利于您的脚本被更多目标用户检索到。
 2. **设置默认值**：合理的 default 可以让用户直接点击运行，极大地降低使用门槛。
 3. **校验必填**：对于没有它脚本就无法运行的参数（如登录 Cookie、主链接），一定要设置 `"required": true`。
+4. **最大结果数命名**：如果 Worker 需要限制最大返回条数，字段名应使用 `max_results`。这是平台及下游集成所识别的标准命名。
diff --git a/src/content/docs/zh-cn/developer-guide/worker-definition/output-schema.md b/src/content/docs/zh-cn/developer-guide/worker-definition/output-schema.md
@@ -44,7 +44,7 @@ sidebar:
 | 属性           | 是否必填 | 说明                                                       |
 | -------------- | -------- | ---------------------------------------------------------- |
 | **name**       | 是       | 列标识符。必须与 `push_data` 中使用的键名一致。必须唯一。 |
-| **type**       | 是       | 列的数据类型。支持：`string`、`integer`、`boolean`、`array`、`object` |
+| **type**       | 是       | 列的数据类型。支持：`string`、`number`、`integer`、`boolean`、`array`、`object` |
 | **description**| 否       | 列描述。在 UI 中作为列标题标签显示。                       |
 
 ## 与 SDK 的关系
diff --git a/src/content/docs/zh-cn/developer-guide/worker-definition/sdk-modules.md b/src/content/docs/zh-cn/developer-guide/worker-definition/sdk-modules.md
@@ -185,6 +185,7 @@ res, err := coresdk.Result.SetTableHeader(ctx, headers)
 * **key**：代码中使用的唯一标识（建议小写+下划线）
 * **format**：数据类型，支持以下值：
   * `"text"`：字符串/文本
+  * `"number"`：数值（整数或浮点数）
   * `"integer"`：整数
   * `"boolean"`：布尔值（true / false）
   * `"array"`：列表/数组
@@ -237,7 +238,7 @@ for i, news := range newsData {
 
 ##### 注意事项
 
-* 表头定义和数据推送**顺序不限**
+* 表头定义**必须在第一次推送数据之前完成**
 * 数据的 key **必须与表头 key 完全一致**（区分大小写）
 * 数据必须**逐条推送**
 * 建议每次推送后记录日志
diff --git a/src/content/docs/zh-cn/integrations/workflows-and-notifications/n8n.md b/src/content/docs/zh-cn/integrations/workflows-and-notifications/n8n.md
