Skip to main content

Chat with your catalog

Atlan MCP gives any connected AI client—Claude, Cursor, GitHub Copilot, ChatGPT, Gemini—direct access to your Atlan catalog. Ask questions in plain language and the AI calls the right Atlan MCP tools automatically: searching for assets, tracing lineage, updating metadata, managing glossary terms, enforcing data quality, and more.

This page is a prompt reference. Each workflow shows which MCP tools the AI reaches for and includes a ready-to-copy example prompt. Paste it directly into your AI client to try it.

info

Make sure Atlan MCP is configured before running any prompts. For setup, see Set up Atlan MCP.

Data discovery​

Use Atlan MCP to find data assets, trace dependencies, and explore lineage from chat—without switching to the catalog UI. Run these prompts in any MCP-aware client to surface relevant tables, dashboards, glossary terms, and their relationships.

Assess lineage and downstream impact​

Understand how a change to a dataset affects dashboards, reports, or downstream assets—evaluate dependencies before applying schema or logic changes.

How it works
1
Trace downstream assets and their dependencies
traverse_lineage
2
Locate impacted reports, dashboards, and models
semantic_search
Example prompt
For the table orders_summary, list all downstream assets and identify impacted dashboards.

Review SQL changes with lineage context​

Assess the lineage impact of SQL changes during pull requests or code reviews—surface affected assets, risk level, and owners before a change is merged.

How it works
1
Evaluate downstream dependencies and impact scope
traverse_lineage
2
Identify and analyze impacted assets
semantic_search
Example prompt
Analyze this SQL diff for potential downstream impact using Atlan MCP.
Return a summary with impacted assets, risk level, and owners.

Analyze data lineage and impact​

Trace asset dependencies, identify impacted objects, and perform impact analysis—upstream or downstream, at any depth.

How it works
1
Visualize upstream and downstream relationships between data assets
traverse_lineage
Example prompt
Get all downstream assets for the table CUSTOMER_TRANSACTIONS.

Asset management​

Update certifications, descriptions, README content, and ownership across your catalog using natural language. These prompts help analysts and stewards keep asset metadata accurate without opening the Atlan UI.

Find assets by certification and popularity​

Find assets by certification status, popularity score, or usage frequency to identify trusted or frequently used datasets.

How it works
1
Filter assets by certification status, popularity scores, and usage metrics
semantic_search
Example prompt
Find the top 10 high-popularity tables related to customers that are certified.

Update and enrich asset metadata​

Update certifications, add descriptions, and modify README documentation to keep metadata accurate and consistent across the catalog.

How it works
1
Modify asset certifications, descriptions, and README documentation
update_assets
Example prompt
Mark the table CUSTOMER_TRANSACTIONS as certified (VERIFIED) and add the description 'Contains customer purchase history'.

Data quality rules​

Create, tune, schedule, and remove data quality rules directly from chat. Use these prompts to enforce null checks, freshness windows, row count thresholds, and custom SQL rules on critical tables and columns.

Create data quality rules​

Define null checks, row count checks, freshness checks, or custom SQL-based rules on critical tables and columns directly from chat.

How it works
1
Define new quality rules on tables, views, and columns
create_dq_rules
Example prompt
Create a null check rule on the EMAIL column in the CUSTOMERS table. The null count should be less than 5 and mark it as URGENT priority.

Adjust rule thresholds and priorities​

Update thresholds, priorities, and conditions on existing rules as data patterns evolve—reduce false positives without losing real coverage.

How it works
1
Modify thresholds, alert priorities, and conditions on existing rules
update_dq_rules
Example prompt
Update the null count threshold for the EMAIL column rule on the CUSTOMERS table to 50 instead of 5.

Schedule quality checks​

Set cron-based schedules for data quality rules to align checks with ETL jobs or off-peak windows.

How it works
1
Configure cron-based schedules for rules on tables, views, and Snowflake dynamic tables
schedule_dq_rules
Example prompt
Schedule all data quality rules on the CUSTOMERS table to run daily at 2 AM UTC.

Remove deprecated rules​

Delete data quality rules that are outdated, noisy, or replaced by better logic to keep your rule set focused.

How it works
1
Remove outdated or noisy data quality rules by GUID
delete_dq_rules
Example prompt
Delete all data quality rules that were created for testing last week on the CUSTOMERS table.

Business glossary​

Create glossaries, organize categories, and define business terms at scale. These prompts let domain owners and stewards align technical assets with business language directly from their AI client.

Explore glossary relationships​

Browse business glossary terms, their definitions, and how they connect to datasets and dashboards.

How it works
1
Map linked datasets and dashboards to glossary terms
semantic_search
2
Fetch business term definitions and their relationships
get_asset
Example prompt
Find glossary terms related to 'Customer Churn' and show how they connect to datasets or dashboards.

Create glossaries, categories, and terms​

Define new glossaries, add category hierarchies, and populate business terms to maintain shared understanding across teams.

How it works
1
Define new glossaries with descriptions and metadata
create_glossaries
2
Add category hierarchies
3
Populate individual business terms
Example prompt
Create a new glossary named 'Customer Data Business Graph'. Add a category 'Customer Demographics' and a verified term 'Customer Lifetime Value (CLV)'.

Tag management​

Apply and remove classification tags such as PII or Confidential to govern sensitive data. These prompts add tags with optional downstream propagation across lineage and remove them when classifications no longer apply.

Apply classification tags​

Tag assets with labels like PII or Confidential and propagate them automatically to downstream lineage.

How it works
1
Look up the correct tag name before applying
resolve_metadata
2
Apply classification tags with optional downstream propagation
add_atlan_tags
Example prompt
Add the PII tag to the CUSTOMERS table and propagate it to all downstream assets.

Remove classification tags​

Remove a classification tag from an asset when it no longer applies—for example, after data has been anonymized.

How it works
1
Remove a specific classification tag from an asset
remove_atlan_tag
Example prompt
The ORDERS_ARCHIVE table no longer contains PII data. Remove the PII tag from it.

Custom metadata​

Set, extend, and clear custom metadata values to capture governance, compliance, and quality context that standard Atlan attributes don't cover. These prompts work with both individual asset values and the schema-level metadata sets.

Set custom metadata values​

Populate custom metadata attributes—such as a Data Quality Score or Governance Status—on specific assets.

How it works
1
Look up the correct custom metadata set and attribute names
resolve_metadata
2
Set attribute values on the target asset
update_custom_metadata
Example prompt
Set the Data Quality Score to 95 on the ORDERS table in the Governance custom metadata set.

Create and extend custom metadata sets​

Define new custom metadata sets or extend existing ones with additional attributes for structured governance, compliance, or quality tracking.

How it works
1
Define a new set with typed attributes
create_custom_metadata_set
2
Extend an existing set with new attributes
add_attributes_to_cm_set
Example prompt
Create a Governance custom metadata set with a Status (string) attribute and an Owner (string) attribute.

Clear custom metadata from assets​

Clear custom metadata values from an asset when the context is no longer relevant—for example, after a migration.

How it works
1
Clear all values from a specific custom metadata set on an asset
remove_custom_metadata
Example prompt
Clear all values from the Data Quality custom metadata set on the ORDERS_LEGACY table.

Announcements​

Post and remove announcements on assets to communicate deprecations, incidents, or important context to anyone viewing the asset in Atlan. Use these prompts to broadcast time-sensitive governance messages.

Add announcements​

Post informational, warning, or issue announcements on assets to flag deprecations, ongoing incidents, or important notices.

How it works
1
Add an announcement with a type (info, warning, or issue), title, and message
manage_announcements
Example prompt
Add a warning announcement to the PAYMENTS_V1 table: 'This table is being deprecated. Please migrate to PAYMENTS_V2 by June 30.'

Remove announcements​

Remove an announcement from an asset once the issue is resolved or the notice is no longer relevant.

How it works
1
Remove an existing announcement from a specific asset
manage_announcements
Example prompt
Remove the issue announcement from the SALES_SUMMARY view. The pipeline has been fixed and data is complete.

Asset lifecycle​

Archive, restore, and purge assets to manage the end-of-life state of deprecated data. These prompts keep your catalog clean while preserving audit trails for assets that are no longer in active use.

Archive deprecated assets​

Archive assets that are no longer in active use but need to be retained for audit or historical purposes.

How it works
1
Archive one or more assets by GUID or qualified name
manage_asset_lifecycle
Example prompt
Archive the following staging tables — STAGING_ORDERS_2023, STAGING_CUSTOMERS_2023, STAGING_PAYMENTS_2023. They have been replaced by their 2024 equivalents.

Restore or purge assets​

Restore an archived asset that's needed again, or permanently purge assets that no longer belong in the catalog.

How it works
1
Restore an archived asset or purge it permanently
manage_asset_lifecycle
Example prompt
Restore the CUSTOMERS_ARCHIVE table. It needs to be available again for the compliance team.