Baselinr Chat - Available Tools

The chat agent has access to the following tools for querying your data quality monitoring system.

Profiling Tools

query_recent_runs

Query recent profiling runs to see what tables have been profiled.

Parameters:

Name	Type	Required	Description
table	string	No	Filter by table name
schema	string	No	Filter by schema name
limit	integer	No	Maximum runs to return (default: 10)
days	integer	No	Look back days (default: 7)
status	string	No	Filter by status (completed, failed)

Example questions:

• "What tables have been profiled recently?"
• "Show me runs for the orders table"
• "List failed profiling runs from the last week"

get_table_profile

Get detailed profile for a specific table including all column metrics.

Parameters:

Name	Type	Required	Description
table	string	Yes	Table name
run_id	string	No	Specific run ID (latest if not provided)

Example questions:

• "Show me the profile for the customers table"
• "What metrics do we have for orders?"
• "Get detailed info about the products table"

Drift & Anomaly Tools

query_drift_events

Query drift detection events to find significant data changes.

Parameters:

Name	Type	Required	Description
table	string	No	Filter by table name
column	string	No	Filter by column name
severity	string	No	Filter by severity (low, medium, high)
days	integer	No	Look back days (default: 7)
limit	integer	No	Maximum events to return (default: 20)

Example questions:

• "Show me high severity drift events"
• "What drift has occurred in the customers table?"
• "Are there any drift issues with the email column?"

query_anomalies

Query anomaly detection events to find unusual data patterns.

Parameters:

Name	Type	Required	Description
table	string	No	Filter by table name
column	string	No	Filter by column name
metric	string	No	Filter by metric (null_rate, mean, distinct_count)
days	integer	No	Look back days (default: 7)
limit	integer	No	Maximum anomalies to return (default: 20)

Example questions:

• "What anomalies have been detected?"
• "Show me null rate anomalies in the orders table"
• "Are there any issues with the customer_id column?"

Trend & History Tools

get_column_history

Get historical trend for a specific column metric over time.

Parameters:

Name	Type	Required	Description
table	string	Yes	Table name
column	string	Yes	Column name
metric	string	Yes	Metric to track (mean, null_ratio, distinct_count)
days	integer	No	Days of history (default: 30)

Example questions:

• "Show me the trend for order_amount mean over the last month"
• "How has the null rate for email changed?"
• "Get the history of distinct_count for customer_id"

compare_runs

Compare two profiling runs to see what changed between them.

Parameters:

Name	Type	Required	Description
table	string	Yes	Table name
run_id_1	string	No	First run ID (latest if not provided)
run_id_2	string	No	Second run ID (second latest if not provided)

Example questions:

• "Compare the last two runs for orders"
• "What changed between the most recent profiles of customers?"
• "Show me the differences in the products table"

Discovery Tools

search_tables

Search for tables by name or schema pattern.

Parameters:

Name	Type	Required	Description
query	string	Yes	Search query (matches table or schema name)
limit	integer	No	Maximum results (default: 20)

Example questions:

• "Find tables with 'order' in the name"
• "Search for customer-related tables"
• "What tables are in the analytics schema?"

get_lineage

Get data lineage (upstream sources and downstream dependents) for a table.

Parameters:

Name	Type	Required	Description
table	string	Yes	Table name
schema	string	No	Schema name
direction	string	No	Lineage direction (upstream, downstream, both)
max_depth	integer	No	Maximum depth to traverse (default: 2)

Example questions:

• "What are the upstream sources for orders?"
• "Show me downstream tables that depend on customers"
• "Get the lineage for the transactions table"

Tool Usage Tips

1. Tools are called automatically - You don't need to specify which tool to use; the agent chooses based on your question

2. Multiple tools per query - The agent may use several tools to answer a complex question

3. Context is maintained - The agent remembers previous tool results when answering follow-up questions

4. Caching - Tool results are cached within a session for faster subsequent queries

5. Verbose mode - Use /verbose to see which tools are being called

Extending Tools

The tool system is extensible. See the developer documentation for how to add custom tools for your specific use cases.