# Collibra MCP for AI Agents MCP

> Collibra MCP connects your AI agent directly to the Collibra data intelligence platform. It lets you search, inspect, and manage an organization's entire data catalog via natural language conversation. You can retrieve metadata for specific assets, list all available communities, or even create new data records without needing to navigate complex UIs.

## Overview
- **Category:** security-compliance
- **Price:** Free
- **Tags:** data-governance, metadata-management, data-catalog, compliance, data-assets, enterprise-data

## Description

Collibra helps organizations build trust in their data by providing a centralized intelligence platform. With this MCP, your AI agent gets direct access to that deep catalog knowledge. Instead of spending hours clicking through dozens of tabs just to find out who owns 'Customer ID' or what classifications it has, you ask the question and get an answer instantly. You can list all domains, check asset relationships, or pull up detailed governance policies for any piece of data. This capability means data stewards and compliance teams operate faster, making manual audits a thing of the past. Vinkius hosts this MCP, giving your AI client access to Collibra's full suite of tools right alongside other enterprise services.

## Tools

### create_asset
Creates a new record for a data asset within the Collibra catalog.

### get_asset
Pulls all detailed information regarding one specific, known data asset.

### get_community_details
Retrieves comprehensive details for a single, identified community within the platform.

### list_asset_types
Returns an exhaustive list of all types of assets available in the system.

### list_assets
Generates a full listing of data assets across the entire catalog.

### list_communities
Provides a list of all organizational communities available in Collibra.

### list_domain_types
Returns an exhaustive list of all types of data domains that can be used for classification.

### list_domains
Generates a full listing of all organizational domains within the platform.

### list_statuses
Provides a list of all available status tags that can be applied to data assets.

### search_assets
Searches for metadata about multiple assets based only on the asset's name.

## Prompt Examples

**Prompt:** 
```
What are the top 3 communities and how many assets are in each?
```

**Response:** 
```
**Data Catalog Summary**

Here's a breakdown of the largest communities:

1. **Engineering:** (15 Domains, 1,204 Assets)
   *Focus: Raw data pipelines and ETL definitions.*
2. **Business Intelligence:** (8 Domains, 532 Assets)
   *Focus: Reporting layers and aggregated metrics.*
3. **Data Governance:** (12 Domains, 845 Assets)
   *Focus: Policy documents and business terms.

You can dive deeper into any of these domains.
```

**Prompt:** 
```
Show me the owner and classification for 'Customer Transaction History'.
```

**Response:** 
```
**Asset Details Found**

| Field | Value |
| :--- | :--- |
| **Name** | Customer Transaction History |
| **Domain** | Financial Data |
| **Owner** | Jane Doe (jane.doe@corp.com) |
| **Classification** | Restricted - PII/PCI |

This asset belongs to the Finance community and must adhere to PCI-DSS guidelines.
```

**Prompt:** 
```
I need a list of all possible data domain types we use.
```

**Response:** 
```
**Available Domain Types:**

The system supports these classifications:

*   **Customer Identity**: For all PII related assets.
*   **Financial Metrics**: Assets tied to revenue, costs, and budgets.
*   **Operational Logs**: Data from application usage and activity tracking.
*   **Product Inventory**: Cataloging physical or digital goods.

These types help keep our data governance consistent across the enterprise.
```

## Capabilities

### Search and Retrieve Metadata
Finds data assets by name, type, or domain and pulls all associated metadata into conversation.

### Map Organizational Structure
Lists all communities and domains to provide a comprehensive view of the entire governance hierarchy.

### Inspect Asset Details
Retrieves full attributes, ownership roles, and relationships for any specific data asset identifier.

### Create New Data Assets
Allows the agent to programmatically generate a new record in the Collibra catalog when needed.

### List Available Domain Types
Provides an exhaustive list of classification standards, ensuring data is categorized correctly upon creation or review.

## Use Cases

### Checking compliance status for sensitive data
A Compliance Analyst needs to prove that all 'PII' assets are correctly classified. The agent uses the tool to retrieve detailed information about specific assets, verifying their assigned governance policies and relationships against regulatory standards.

### Discovering unknown data sources for a project
A Data Engineer starts a new model and needs input tables. They ask the agent to list all available assets in 'Data Engineering' community, narrowing down potential sources without manually browsing hundreds of entries.

### Structuring an internal data governance wiki
A Data Steward wants to document best practices for a new department. They use the tool to list all domains and communities available in Collibra, creating a structured map that guides future users on where to store governed knowledge.

### Validating data definitions before deployment
A team needs to ensure two separate tables both refer to the same 'Customer ID' definition. They use the agent to get asset details for both, instantly confirming they share consistent attributes and responsibilities.

## Benefits

- Saves time on audits: Instead of navigating complex UI paths to verify data lineage, you can use the agent to check specific asset relationships instantly.
- Quickly understand ownership: Use the tool to retrieve detailed information about any single asset, immediately showing who is responsible for it and what its attributes are.
- Map your entire structure: You can list all communities and domains, giving you a high-level view of where data assets reside without needing administrative access to every section.
- Automate documentation: Easily look up table/column definitions or ownership from chat. Data engineers get actionable metadata instantly, speeding up development time.
- Build new records fast: Need to log a new piece of governed data? The `create_asset` tool lets you programmatically add assets without manual form filling.

## How It Works

The bottom line is that your AI client treats the complex data catalog like a simple search engine, giving you instant answers instead of endless clicks.

1. First, add the Collibra integration to your AI client's toolset and provide the required instance details.
2. Next, tell your agent what you need. You can ask it to search for a specific asset or list all available communities using plain language.
3. The MCP executes the request against the data platform and sends back structured metadata, which your agent then presents in conversational format.

## Frequently Asked Questions

**How does Collibra MCP help me find data ownership?**
It quickly finds and reports the owner (Data Steward) for any asset you reference. You no longer have to hunt through departmental contacts; the metadata gives you a direct answer, saving hours of manual investigation.

**Can I use Collibra MCP to map out my entire data structure?**
Yes. By listing all available communities and domains, your agent provides a clear, high-level map of the entire data catalog. This is vital for understanding scope before starting any major project.

**Is Collibra MCP useful for compliance audits?**
Absolutely. It allows you to verify specific asset classifications and relationships instantly, providing auditable proof that governance policies are consistently applied across your data assets without manual checks.

**What if I need to add a new data asset record?**
You can use the MCP to create new records directly in Collibra. This means you don't have to manually fill out forms; your agent handles the structured entry, keeping your catalog clean and up-to-date.

**Can I search for data assets using natural language?**
Yes. You just ask your AI agent what you're looking for—by name or type—and it translates that into a metadata query, bringing the relevant results to you in plain conversation.