@clickzetta/cz-cli-darwin-arm64 0.5.15 → 0.5.17

package/bin/skills/lakehouse-doc-en/references/create-catalog-connection.md

@@ -251,3 +251,4 @@ DROP CONNECTION my_conn;
 - [Create External Schema](create-external-schema.md) — Mount an external database into the current workspace
 - [Create Storage Connection](create-storage-connection.md) — Storage Connection required by HMS
 - [Accessing Iceberg Tables in Snowflake Open Catalog](query-snowflake-open-catalog-iceberg-table.md)
+- [Databricks Unity Catalog Federated Query Practice](databricks-external-catalog-practice.md) — Complete path from Databricks-side configuration to query verification

package/bin/skills/lakehouse-doc-en/references/create-dynamic-table.md

@@ -262,7 +262,7 @@ When refreshing a partitioned table, the `partition_spec` must be specified in t
   ```sql
   set dt.args.day='2024-11-13';
   set dt.args.hour='23';
-  REFRESH DYNAMIC TABLE dt PARTITION (day='2024-11-13', hour='23');
+  REFRESH DYNAMIC TABLE dt PARTITION (day='2024-11-13', hour='23'); 
   ```
 
 * **Invalid specification**: Skip an intermediate level (e.g., skip hour and directly specify min).
@@ -270,7 +270,7 @@ When refreshing a partitioned table, the `partition_spec` must be specified in t
   -- Error example: hour is skipped
   set dt.args.day='2024-11-13';
   set dt.args.min='30';
-  REFRESH DYNAMIC TABLE dt PARTITION (day='2024-11-13', min='30');
+  REFRESH DYNAMIC TABLE dt PARTITION (day='2024-11-13', min='30'); 
   ```
 
 
@@ -412,8 +412,8 @@ CREATE DYNAMIC TABLE doc_dt_user_purchase_analysis (
 )
 COMMENT 'Real-time user purchase behavior analysis table'
 REFRESH INTERVAL 1 HOUR VCLUSTER default
-AS
-SELECT
+AS 
+SELECT 
     u.user_id, u.user_name, u.city,
     COUNT(o.order_id) AS total_orders,
     SUM(o.amount) AS total_amount

package/bin/skills/lakehouse-doc-en/references/create-external-catalog.md

@@ -1,6 +1,6 @@
-# Create Hive External Catalog
+# Create External Catalog
 
-Steps to create Hive External Catalog
+Steps to create a Hive External Catalog:
 
 1. [Create Storage Connection](create-storage-connection.md): First, you need to create a storage connection to access the object storage service.
 2. [Create Catalog Connection](create-catalog-connection.md): Use the storage connection information and Hive Metastore address to create a Catalog Connection.
@@ -13,11 +13,11 @@ CREATE EXTERNAL CATALOG catalog_name
 ```
 **Parameter Description**
 
-catalog\_api\_connection: The name of the catalog connection. Currently, only HIVE is supported. [Refer to creating catalog connection](create-catalog-connection.md)
+* **catalog_api_connection**: The name of the Catalog Connection. Supports HMS (Hive), Databricks Unity Catalog, Iceberg REST Catalog, and other types. [Refer to creating Catalog Connection](create-catalog-connection.md)
 
-### Example
+### Examples
 
-Example 1: Hive ON OSS
+**Case 1: Hive ON OSS (Read and Write Supported)**
 
 * Create storage connection
 ```SQL
@@ -57,9 +57,10 @@ SHOW SCHEMAS IN my_external_catalog;
 -- Verify reading data, STORAGE CONNECTION permission will be used when reading data
 SELECT * FROM my_external_catalog.my_schema.my_table;
 ```
-Case 2: Hive ON COS
+
+**Case 2: Hive ON COS (Read and Write Supported)**
+
 * Create storage connection
-```
 ```sql
 CREATE STORAGE CONNECTION catalog_storage_cos 
   TYPE COS
@@ -81,12 +82,6 @@ type hms
 hive_metastore_uris='xxxx:9083'    
 storage_connection='catalog_storage_cos';
 ```
-```SQL
-CREATE CATALOG CONNECTION if not exists catalog_api_connection    
-type hms    
-hive_metastore_uris='xxxx:9083'    
-storage_connection='catalog_storage_oss';
-```
 * Create Catalog
 ```SQL
 CREATE EXTERNAL CATALOG my_external_catalog
@@ -109,7 +104,8 @@ SHOW SCHEMAS IN my_external_catalog;
 -- Verify reading data, STORAGE CONNECTION permission will be used when reading data
 SELECT * FROM my_external_catalog.my_schema.my_table;
 ```
-Case 3: Hive ON S3
+
+**Case 3: Hive ON S3 (Read and Write Supported)**
 
 * Create storage connection
 ```
@@ -155,7 +151,8 @@ SHOW SCHEMAS IN my_external_catalog;
 -- Verify reading data, STORAGE CONNECTION permission will be used when reading data
 SELECT * FROM my_external_catalog.my_schema.my_table;
 ```
-Case 4: Hive ON HDFS (Read Support)
+
+**Case 4: Hive ON HDFS (Read Support)**
 
 * **Create Storage Connection**
 
@@ -196,25 +193,91 @@ Case 4: Hive ON HDFS (Read Support)
   SELECT * FROM my_external_catalog.my_schema.my_table;
   ```
 
+> Lakehouse supports connecting to Hive Metastore and HDFS clusters with Kerberos authentication enabled.
+>
+> **Step 0: Prerequisites**
+>
+> 1. **Kerberos environment**: Ensure your Hive Metastore and HDFS cluster have Kerberos authentication enabled.
+> 2. **Authentication files**: Prepare the following required Kerberos authentication files:
+>
+>   * `krb5.conf`: Kerberos configuration file
+>   * `hive.keytab`: Keytab file for the service principal
+>
+> 3. **Network connectivity**: Ensure Singdata Lakehouse can access the Kerberos KDC and Hive Metastore services.
+>
+> **Step 1**: Upload the Kerberos config file and keytab file to a Volume using a [client tool](comprehensive_guide_to_ingesting_dbv_sql_put.md):
+>
+> ```
+> -- upload krb5.conf
+> PUT '/etc/krb5.conf' TO USER VOLUME FILE 'krb5.conf';
+>
+> -- upload keytab 
+> PUT '/path/to/hive.keytab' TO USER VOLUME FILE 'hive.keytab';
+> ```
+>
+> **Step 2**: Create a Catalog Connection with Kerberos authentication using the following syntax:
+>
+> ```
+> CREATE CATALOG CONNECTION conn_hms_kerberos_auth
+> TYPE HMS
+> AUTH_TYPE = 'kerberos'
+> KERBEROS_CLIENT_PRINCIPAL = 'hive/localhost@YOUR-REALM.COM'
+> KERBEROS_SERVICE_PRINCIPAL = 'hive/localhost@YOUR-REALM.COM'
+> KERBEROS_KRB5_CONFIG_PATH = 'volume:user//~/krb5.conf'
+> KERBEROS_KEYTAB_PATH = 'volume:user//~/hive.keytab'
+> HIVE_METASTORE_URIS = 'thrift://your-hms-host:9083'
+> STORAGE_CONNECTION = 'oss_conn_hz';
+> ```
+>
+> | Parameter Name | Required | Description | Example |
+> | -------------- | -------- | ----------- | ------- |
+> | AUTH_TYPE | Yes | Authentication type, set to 'kerberos' | 'kerberos' |
+> | KERBEROS\_CLIENT\_PRINCIPAL | Yes | Kerberos client principal | 'hive/localhost\@CZ.COM' |
+> | KERBEROS\_SERVICE\_PRINCIPAL | Yes | Kerberos service principal | 'hive/localhost\@CZ.COM' |
+> | KERBEROS\_KRB5\_CONFIG\_PATH | Yes | Path to krb5.conf file | 'volume://vol\_name/krb5.conf' |
+> | KERBEROS\_KEYTAB\_PATH | Yes | Path to keytab file | 'volume://vol\_name/hive.keytab' |
+>
+> Connection parameters:
+>
+> `TYPE:` The connection type, set to `HMS`
+>
+> `HIVE_METASTORE_URIS:` The Hive Metastore service address `'thrift://host:9083'`
+>
+> `STORAGE_CONNECTION:` The associated storage connection name `'your_storage_conn'`
+>
+> ^
+>
+> **Step 3**: Create an External Schema based on the Kerberos-authenticated Catalog Connection
+>
+> ```
+> CREATE EXTERNAL SCHEMA ext_kerberos_hms_sch 
+> CONNECTION conn_hms_kerberos_auth 
+> OPTIONS ('schema'='default');
+> ```
+
+^
 
 # Create Databricks External Catalog
 
-Steps to create Databricks External Catalog
+Steps to create Databricks External Catalog:
 
-1. [Create Catalog Connection](create-catalog-connection.md): Store Databricks' Unity Catalog, connection authentication information.
+1. [Create Catalog Connection](create-catalog-connection.md): Store Databricks' Unity Catalog connection authentication information.
 2. Create External Catalog: Use Catalog Connection to create an external Catalog to access external data in the data lake.
 
 ## Syntax
 ```SQL
 CREATE EXTERNAL CATALOG catalog_name
-    CONNECTION catalog_api_connection;
+    CONNECTION catalog_api_connection
     OPTIONS ('catalog'='catalog_name');
 ```
 **Parameter Description**
-- **`catalog_name`**  ：The name of the external Catalog. This name is used to identify the Catalog, and it must be unique and comply with naming conventions.                                       
-- **`CONNECTION catalog_api_connection`** ：Specifies the connection to the external Catalog. `catalog_api_connection` is the name of a pre-created connection used to access the external Catalog.   
-- **`OPTIONS ('catalog'='catalog_name')`**  : Specifies the configuration options for the external Catalog. `'catalog'='catalog_name'`: Indicates the name of the external Catalog, `catalog_name` is the name of the target Catalog.
+- **`catalog_name`**: The name of the external Catalog. This name is used to identify the Catalog, and it must be unique and comply with naming conventions.
+- **`CONNECTION catalog_api_connection`**: Specifies the connection to the external Catalog. `catalog_api_connection` is the name of a pre-created connection used to access the external Catalog.
+- **`OPTIONS ('catalog'='catalog_name')`**: Specifies the configuration options for the external Catalog. `'catalog'='catalog_name'`: Indicates the name of the external Catalog, `catalog_name` is the name of the target Catalog.
+
+^
 
 ## Related Guides
 
-- [Federated Query](sql_external_catalog_guide.md): Complete examples for querying external data sources such as Hive and Databricks through External Catalog.
+- [Federated Query](SQL_External_Catalog_Guide.md): Complete examples for querying external data sources such as Hive and Databricks through External Catalog.
+- [Databricks Unity Catalog Federated Query Practice](databricks-external-catalog-practice.md) — Complete examples from Databricks-side configuration, permissions, common errors, and federated query scenarios

package/bin/skills/lakehouse-doc-en/references/create-table-ddl.md

@@ -19,6 +19,7 @@ CREATE TABLE [ IF NOT EXISTS ] table_name
     [SORTED BY (column_name [ ASC | DESC ])] 
     [INTO num_buckets BUCKETS] 
 ]
+[ ROW FILTER filter_function ON (column_name,...) ]
 [ COMMENT 'table_comment' ]
 [PROPERTIES('data_lifecycle'='day_num')];
 
@@ -31,6 +32,7 @@ CREATE TABLE [ IF NOT EXISTS ] table_name
 column_name column_type 
 { NOT NULL |
   PRIMARY KEY|
+  UNIQUE|
   IDENTITY[(seed)]|
   GENERATED ALWAYS AS ( expr ) |
   DEFAULT default_expression |
@@ -75,6 +77,26 @@ CREATE    TABLE pk_table (
           PRIMARY key (id, pt)
           ) PARTITIONED BY (pt);
 ```
+#### Unique Key (UNIQUE)
+
+* **Unique Key (UNIQUE)** is used to declare that the values of a column or a combination of columns are unique. Unlike primary keys, UNIQUE is a **declarative constraint**: in the default mode (`DISABLE NOVALIDATE RELY`), uniqueness is not enforced at write time. It is primarily used to declare data semantics to the query optimizer to improve query optimization. UNIQUE columns allow NULL values. A table can have multiple UNIQUE constraints, and they can only be specified at table creation time. Both column-level and table-level (including composite) forms are supported. For detailed behavior, see [Unique Key Introduction](unique-key.md).
+
+```
+-- Column-level
+CREATE TABLE uk_table (id int UNIQUE, col string);
+
+-- Table-level (single column)
+CREATE TABLE uk_table (id int, col string, UNIQUE (id));
+
+-- Table-level (composite)
+CREATE TABLE uk_table (a int, b int, UNIQUE (a, b));
+
+-- With modifier (default is DISABLE NOVALIDATE RELY)
+CREATE TABLE uk_table (id int UNIQUE NORELY, col string);
+```
+
+> ⚠️ If you need to enforce deduplication at write time, use a primary key (PRIMARY KEY); do not rely on the UNIQUE constraint.
+
 #### Auto-increment Column (IDENTITY\[(seed)])
 
 * **IDENTITY\[(seed**)]: Supports specifying auto-increment. It cannot guarantee that the values in the sequence are continuous (gapless), nor can it guarantee that the sequence values are allocated in a specific order. This is because other concurrent inserts may occur in the table. These limitations are part of the design to improve performance. For specific usage, refer to the [IDENTITY Column documentation](identity-column.md)
@@ -168,6 +190,27 @@ CREATE TABLE sales_data (
 ) 
 SORTED BY (sale_date DESC);
 ```
+### ROW FILTER (Row-Level Security)
+
+> [Preview Release] This feature is currently in an invite-only preview. Contact our technical support team for assistance.
+
+* **ROW FILTER**: Binds a filter function that returns `BOOLEAN` to a table. The system automatically applies it during queries and DML operations — only rows for which the function returns `true` are visible to the current operation. Commonly used for multi-tenant isolation and row-level data access control by user/role. The filter function can use security context functions such as `current_user()` and `current_roles()` to dynamically filter based on the current login identity. Columns listed in `ON (...)` are passed in order as arguments to the filter function; their types and count must match the function definition. It is recommended to reference filter functions using schema-qualified names. See [Row-Level Security (Row Filter)](row-filter.md).
+
+```sql
+-- Create a filter function: each user can only see rows where owner equals their login name
+CREATE FUNCTION my_schema.owner_only(owner STRING) RETURNS BOOLEAN
+AS owner = current_user();
+
+-- Bind a row filter at table creation
+CREATE TABLE my_schema.docs (
+    id      INT,
+    owner   STRING,
+    content STRING
+) ROW FILTER my_schema.owner_only ON (owner);
+```
+
+You can also bind or remove a row filter on an existing table using `ALTER TABLE ... SET ROW FILTER ...` and `ALTER TABLE ... DROP ROW FILTER`.
+
 ### PROPERTIES
 
 You can set table-level properties via `PROPERTIES` at table creation time, and also modify them via `ALTER TABLE ... SET PROPERTIES`.
@@ -370,6 +413,8 @@ Users can create indexes on multiple columns when creating a table. Indexes can
 | [Partitioned Table Usage Guide](notes-and-guidelines-for-partition-tables.md) | Usage patterns for partition design, partition pruning, and dynamic partitions |
 | [Bucketing (CLUSTERED BY)](cluster-table-guide.md) | Hash bucketing to improve Join and aggregation performance |
 | [Primary Key](primary-key.md) | Primary key table design, supporting CDC real-time write deduplication |
+| [Unique Key](unique-key.md) | Declarative UNIQUE constraint for query optimization |
+| [Row-Level Security (Row Filter)](row-filter.md) | Bind a filter function to a table to control row-level data visibility |
 
 **Indexing**
 

package/bin/skills/lakehouse-doc-en/references/creating_alicloud_privatelinkendpoint.md

@@ -2,7 +2,7 @@
 
 When you need to access the Singdata Lakehouse service through the internal network within the cloud provider's VPC, you need to create an endpoint in Alibaba Cloud that connects to the Singdata Lakehouse endpoint service. Then replace the domain name of the JDBC connection or API accessing Singdata Lakehouse with the domain name of the endpoint.
 
-:-: ![](.topwrite/assets/privatelink_customer_access_LH_example.png)
+
 
 ## Steps
 
@@ -10,7 +10,6 @@ The following operations need to be completed in the Alibaba Cloud console envir
 
 1\. **Add to Whitelist**. First, please confirm that you have added your cloud platform account to the whitelist.
 
-:-: ![](.topwrite/assets/image_1733063925523.png =817)
 
 To ensure that Lakehouse can normally read the status of your endpoint and enhance the security of obtaining endpoint information within your cloud service platform, please create an independent access control role in the cloud service platform, and authorize and add the external ID.
 
@@ -18,19 +17,18 @@ To ensure that Lakehouse can normally read the status of your endpoint and enhan
 
 On the Alibaba Cloud "RAM Access Control" - "Roles" page, select the access control role whose ARN you want to add, copy the values of ARN and externalID respectively, return to the Singdata Lakehouse page, and in the add whitelist popup, copy the above two items into the corresponding options.
 
-:-: ![](.topwrite/assets/image_1733069144028.png =796)
+
 
 ^
 
-:-: ![](.topwrite/assets/image_1733069236299.png =366)
 
 2\. **Create Endpoint**. On the "Endpoints" page of the Alibaba Cloud console, click the "Create Endpoint" button. Select the same region as Lakehouse in the "Region" field, and select "Other Endpoint Services" on the page. Copy the endpoint service name of the region from the Lakehouse service and fill it in on this page.
 
-:-: ![](.topwrite/assets/image_1733063937656.png =819)
+
 
 ^
 
-:-: ![](.topwrite/assets/image_1733063950478.png =815)
+
 
 3\. **Allow Endpoint Connection**. After creation, refresh the "Endpoints" page in Lakehouse, and you will see the endpoint you created in the list. Click "Allow Connection" to complete the network interconnection configuration based on the private network connection.
 

package/bin/skills/lakehouse-doc-en/references/creating_alicloud_privatelinkservice.md

@@ -2,7 +2,7 @@
 
 When you need to use Singdata Lakehouse to access services within your cloud vendor's VPC (such as a self-built MySQL database), you first need to create an endpoint service on Alibaba Cloud and configure the service to be accessed as the service resource of the endpoint service.
 
-:-: ![](.topwrite/assets/private_link_access_customer_example.png)
+
 
 ## Steps
 
@@ -12,20 +12,17 @@ The following operations need to be completed in the Alibaba Cloud console envir
 
 2\. In the Alibaba Cloud console, click PrivateLink - Endpoint Services in the left navigation bar, and select the same region as the current Lakehouse service instance to create the endpoint service. You can find the region and availability zone of the current service instance at the top of the "Create Endpoint Service" popup in the Lakehouse service. Please create the endpoint service according to the region and availability zone of the current service instance, otherwise, Lakehouse will not be able to create the corresponding endpoint and establish PrivateLink.
 
-:-: ![](.topwrite/assets/image_1733063729452.png =381)
 
 3\. Configure other options for creating the endpoint service according to the actual situation of your service on Alibaba Cloud. Please note, select "Service Provider" for the "Service Payer" option, otherwise, it will affect the normal creation of the connection.
 
-:-: ![](.topwrite/assets/image_1733063748752.png =791)
+
 
 4\. After the endpoint service is created, click the name of the endpoint to enter the endpoint service details page, and add the Singdata cloud account ID to the whitelist.
 
-:-: ![](.topwrite/assets/image_1733063764046.png =757)
+
 
 ^
 
-:-: ![](.topwrite/assets/image_1733063781421.png =367)
 
-5\. After completing the above operations, return to the Lakehouse page. In the "Create Endpoint Service" popup, fill in the endpoint service instance ID (found on the Alibaba Cloud endpoint service details page) and name created in the above steps, and click "OK".
 
-:-: ![](.topwrite/assets/image_1733063795296.png =557)
+5\. After completing the above operations, return to the Lakehouse page. In the "Create Endpoint Service" popup, fill in the endpoint service instance ID (found on the Alibaba Cloud endpoint service details page) and name created in the above steps, and click "OK".

package/bin/skills/lakehouse-doc-en/references/creating_tencentcloud_privatelinkendpoint.md

@@ -2,7 +2,6 @@
 
 When you need to access the Singdata Lakehouse service through the internal network within the cloud provider's VPC, you need to create an endpoint in Tencent Cloud that connects to the Singdata Lakehouse endpoint service. Then replace the domain name of the JDBC connection or API accessing Singdata Lakehouse with the domain name of the endpoint.
 
-:-: ![](.topwrite/assets/privatelink_customer_access_LH_example.png)
 
 ## Steps
 
@@ -10,7 +9,7 @@ The following operations need to be completed in the Alibaba Cloud console envir
 
 1\. **Add to Whitelist**. First, please confirm that you have added your cloud platform account to the whitelist.
 
-![](.topwrite/assets/image_1732869126126.png)
+
 
 To ensure that Lakehouse can correctly read the status of your endpoint and enhance the security of obtaining endpoint information within your cloud service platform, please create an independent access control role within the cloud service platform, and authorize and add an external ID.
 
@@ -20,16 +19,12 @@ On the Tencent Cloud access control role list page, click the role name you want
 
 After pasting, click "OK".
 
-:-: ![](.topwrite/assets/image_1733067670318.png =826)
-
-:-: ![](.topwrite/assets/image_1733067744128.png =826)
 
 ^
 
-:-: ![](.topwrite/assets/image_1733067854956.png =447)
+
 
 2\. **Create a New Endpoint**. On the endpoint page, click the "New" button, select "Other Account" for "Peer Account Type". Then, in "Peer Account UIN" and "Peer Endpoint Service ID", respectively fill in the "Lakehouse UID" and "Endpoint Service ID" copied from the Lakehouse page. Click "OK".
 
-:-: ![](.topwrite/assets/image_1733063999441.png =455)
 
 3\. **Allow Endpoint Connection**. After creation, refresh the "Endpoint" page in Lakehouse, and you will see the endpoint you created in the list. Click "Allow Connection" to complete the network interconnection configuration based on the private connection.

package/bin/skills/lakehouse-doc-en/references/creating_tencentcloud_privatelinkservice.md

@@ -2,18 +2,14 @@
 
 When you need to use Singdata Lakehouse to access services within your cloud vendor's VPC (such as a self-built MySQL database), you first need to create an endpoint service on Tencent Cloud and configure the service to be accessed as a service resource for that endpoint service.
 
-:-: ![](.topwrite/assets/private_link_access_customer_example.png)
+
 
 ## Steps
 
 1\. On the Tencent Cloud Network Connection - Endpoint Service page, select the same region as the current Lakehouse service instance to create an endpoint service. You can find the region and availability zone of the current service instance at the top of the "Create Endpoint Service" popup in the Lakehouse service. Please create the endpoint service according to the region and availability zone of the current service instance, otherwise, Lakehouse will not be able to create the corresponding endpoint and establish PrivateLink.
 
-:-: ![](.topwrite/assets/image.png =412)
 
 2\. After completing the creation of the endpoint service, click on the endpoint service name to enter the details page. Switch to the "Whitelist" tab, click the "Add" button, and add the UIN of the Lakehouse service. You can copy the "Lakehouse UID" from the information popup of the newly created endpoint service in the Lakehouse service and fill it in.
 
-:-: ![](.topwrite/assets/image.png =412)
 
 3\. After completion, please copy your endpoint service ID and paste it into the Lakehouse creation page, then click "Confirm" to complete the creation of the endpoint service.
-
-:-: ![](.topwrite/assets/image_1733063879303.png =408)

package/bin/skills/lakehouse-doc-en/references/cz-cli-agent.md

@@ -55,18 +55,30 @@ If you use an enterprise bot such as Hermes to host an AI Agent, it is recommend
 
 `cz-cli agent run` requires an LLM to be configured before use. LLM configuration is independent of the Lakehouse connection profile and is stored in the `[llm.*]` section of `~/.clickzetta/profiles.toml`.
 
-### Method 1: Use the Singdata Built-in LLM (Recommended for New Users)
+### Method 1: Use the Singdata Built-in LLM (Recommended, provided by AI Gateway)
 
-Complete configuration in one step using the CLI connection string:
+The Singdata built-in LLM is provided uniformly through [AI Gateway](AIGateway.md) — no separate model API Key required. Complete configuration in one step using the CLI connection string:
 
 ```bash
 cz-cli setup --credential <CLI connection string>
 ```
 
-The CLI connection string is available on the LakehouseMCP page in the account console (see the [Installation and Configuration Guide](setup_cz_cli.md)).
+The CLI connection string is available on the LakehouseMCP page in the account console (see the [Installation and Configuration Guide](setup_cz_cli.md)). The connection string already contains the AI Gateway endpoint and authentication information — ready to use once configured.
 
 ### Method 2: Connect an External LLM
 
+Supports any OpenAI-compatible interface. If your enterprise already has an AI Gateway, simply enter its endpoint:
+
+**Use AI Gateway (enterprise self-hosted or Singdata-managed)**
+
+```bash
+cz-cli agent llm add my-gateway \
+  --provider openai-compatible \
+  --base-url https://<your-instance>.singdata.com/gateway/v1 \
+  --api-key <AI-Gateway-API-Key> \
+  --use
+```
+
 **OpenAI / GPT**
 
 ```bash
@@ -88,35 +100,28 @@ cz-cli agent llm add my-relay \
 
 ### Verify and Manage
 
-```bash
-```
-
 View the currently active LLM and all configurations:
 
 ```bash
 cz-cli agent llm show
-
 ```
 
 List all configured LLMs:
 
 ```bash
 cz-cli agent llm list
-
 ```
 
 Test connectivity:
 
 ```bash
 cz-cli agent llm test my-openai
-
 ```
 
 Switch the active LLM:
 
 ```bash
 cz-cli agent llm use my-openai
-
 ```
 
 Remove an LLM configuration:

package/bin/skills/lakehouse-doc-en/references/cz-cli-datasource.md

@@ -2,42 +2,34 @@
 
 The `datasource` command browses and tests external data sources configured in Studio (MySQL, PostgreSQL, Kafka, etc.). It is commonly used to validate connections before configuring sync tasks:
 
-```bash
-```
-
 List all configured data sources:
 
 ```bash
 cz-cli -p prod datasource list
-
 ```
 
 Test data source connectivity:
 
 ```bash
 cz-cli -p prod datasource test <datasource_name>
-
 ```
 
 Browse databases/schemas in a data source:
 
 ```bash
 cz-cli -p prod datasource catalogs <datasource_name>
-
 ```
 
 Browse tables in a specific database:
 
 ```bash
 cz-cli -p prod datasource objects <datasource_name> <catalog_name>
-
 ```
 
 View metadata for a specific table:
 
 ```bash
 cz-cli -p prod datasource describe <datasource_name> <catalog_name> <table_name>
-
 ```
 
 Preview sample data from a table: