# 6 of 16 tools document the agent that was told a sensitive field

EXPOSURE MAP, BY STAGE, read 13 Nov 2026

**The short answer.** At the agent stage, 6 of 16 tools document it (Atlan, DataHub, Snowflake Horizon, Databricks Unity Catalog and others), 10 are partial and 0 show none found on 13 Nov 2026. Agent calls are the newest record and the one regulators ask about next.

The agent is the AI assistant or MCP client that received data or context. A tool covers this stage when it records or controls what an agent was told, per call and per person asking.

Documented here means the documentation shows per-call logging or per-person control of what an agent receives. Partial means it shows an agent log without the person behind it, or control for one agent path only, or states another limit.

## Documented

- [Atlan](/platforms/atlan/), 65.5: Atlan says MCP forwards the caller's identity so the same persona, role and domain policies as the UI apply to every tool call. [Source](https://docs.atlan.com/product/capabilities/atlan-ai/references/mcp-security), read 13 Nov 2026.
- [DataHub](/platforms/datahub/), 61.0: DataHub docs say Cloud admins can create scoped MCP servers with their own URL, hidden tools, custom instructions and a saved View; this needs the Context add-on. [Source](https://docs.datahub.com/docs/features/feature-guides/scoped-mcp-servers), read 13 Nov 2026.
- [Snowflake Horizon](/platforms/snowflake/), 55.5: Snowflake says that when an AI agent queries data, the role's access control, masking and row-access policies are enforced in the query engine, and metadata access is managed separately. [Source](https://docs.snowflake.com/en/user-guide/snowflake-horizon/horizon-context.md), read 13 Nov 2026.
- [Databricks Unity Catalog](/platforms/databricks/), 55.0: Databricks says row filters and column masks defined in Unity Catalog are enforced per user automatically in a Genie Agent, regardless of how it is shared. [Source](https://docs.databricks.com/aws/en/genie/set-up), read 13 Nov 2026.
- [Alation](/platforms/alation/), 54.0: Alation docs say agents and tools hold no data permissions of their own and act as the caller on each run. [Source](https://documentation.alation.com/en/latest/agentstudio/security/identity-and-permissions.md), read 13 Nov 2026.
- [Securiti](/platforms/securiti/), 29.5: Securiti says its copilot respects source file entitlements at ingestion and at user consumption. [Source](https://securiti.ai/gencore/safe-enterprise-ai-copilot/), read 13 Nov 2026.

## Partial

- [Euno](/platforms/euno/), 67.0: Euno docs say Euno MCP uses the active persona to determine what data and actions are available. [Source](https://docs.euno.ai/ai/mcp-setup.md?utm_source=aidatagovernancetools.com&utm_medium=referral&utm_campaign=map), read 13 Nov 2026.
- [Microsoft Purview](/platforms/microsoft-purview/), 54.5: Microsoft says supported AI apps use existing access controls so data is never returned to a user, or used by the LLM, if that user lacks access. [Source](https://learn.microsoft.com/purview/ai-microsoft-purview), read 13 Nov 2026.
- [Select Star](/platforms/select-star/), 53.5: Select Star docs say the Ask AI chatbot sees all data assets by default, and admins can narrow it by data source, schema or folder for the whole instance. [Source](https://web.archive.org/web/20260715110406/https://docs.selectstar.com/llms-full.txt), read 13 Nov 2026.
- [Secoda](/platforms/secoda/), 47.5: Secoda docs say its AI follows team-based permissions: a user asking about a team they are not in gets no results. [Source](https://docs.secoda.co/features/ai-assistant/security.md), read 13 Nov 2026.
- [Collibra](/platforms/collibra/), 46.5: The chip README says tools work with the default scopes of any authenticated Collibra user, and a permission error means the connecting user lacks a scope such as dgc.ai-copilot or dgc.classify. [Source](https://github.com/collibra/chip), read 13 Nov 2026.
- [Immuta](/platforms/immuta/), 45.0: Immuta says its agentic data access is a centralized authorization framework governing how AI agents and LLMs interact with sensitive data. [Source](https://documentation.immuta.com/saas/agentic-data-access/introduction.md), read 13 Nov 2026.
- [Ataccama ONE](/platforms/ataccama/), 41.5: Ataccama docs say each MCP request runs under the user's own identity. [Source](https://docs.ataccama.com/one/latest/generative-ai/mcp-dqc.html), read 13 Nov 2026.
- [Informatica](/platforms/informatica/), 30.0: Informatica says governance travels with the agent in its headless design, with security and privacy controls built into the infrastructure. [Source](https://www.informatica.com/headless.html), read 13 Nov 2026.
- [Privacera](/platforms/privacera/), 30.0: Trust3 AI Governance docs say it discovers agents on Databricks, Microsoft Copilot Studio and Azure Foundry, scores them and tracks policy violations. [Source](https://docs.trust3.ai/get-started/overview.html), read 13 Nov 2026.
- [OneTrust](/platforms/onetrust/), 26.5: OneTrust says its AI Governance module governs agents with defined purpose, permissions and allowed actions across MCP-enabled environments. [Source](https://www.onetrust.com/products/ai-governance/), read 13 Nov 2026.

## None found

- None of the 16 on 13 Nov 2026.

## Verdict by situation

- **A regulated field** ([a governance lead who must show where a regulated field ends up](/best-for/regulated-field/)): Atlan and Euno tie at 68.5. Atlan and Euno tie on total: Atlan is ahead on Acts on it (7 to 5), Euno is ahead on Carries classification downstream (8 to 7). Atlan and Euno tie at 68.5 on total. The published tie rule (higher score on the heaviest criterion, then by name) puts Atlan ahead: the two tie on Follows sensitive data into BI too, so the name decides.
- **Agents on analytics** ([a company connecting AI agents to its analytics data this year](/best-for/agents-on-analytics/)): Euno leads at 66.0, ahead of Atlan at 65.0. Euno leads on Price transparency (4 to 0) and Carries classification downstream (8 to 7), while Atlan is ahead on Acts on it (7 to 5).
- **A Microsoft estate** ([a Microsoft estate: Fabric, Power BI and Purview](/best-for/microsoft-estate/)): Atlan leads at 68.5, ahead of Euno at 67.5. Atlan leads on Acts on it (7 to 5), while Euno is ahead on Carries classification downstream (8 to 7).
- **A mixed stack** ([a mixed stack: Snowflake or Databricks with dbt and two or more BI tools](/best-for/mixed-stack/)): Euno leads at 68.5, ahead of Atlan at 65.5. Euno leads on Carries classification downstream (8 to 7) and Price transparency (4 to 0), while Atlan is ahead on Acts on it (7 to 5).
- **A lean team** ([a lean data team with no governance hire](/best-for/lean-team/)): Euno leads at 62.5, ahead of Atlan at 59.0. Euno leads on Price transparency (4 to 0) and Carries classification downstream (8 to 7), while Atlan is ahead on Acts on it (7 to 5).

## Sources

- https://documentation.alation.com/en/latest/agentstudio/security/identity-and-permissions.md: https://documentation.alation.com/en/latest/agentstudio/security/identity-and-permissions.md (read 13 Nov 2026)
- https://docs.ataccama.com/one/latest/generative-ai/mcp-dqc.html: https://docs.ataccama.com/one/latest/generative-ai/mcp-dqc.html (read 13 Nov 2026)
- https://docs.atlan.com/product/capabilities/atlan-ai/references/mcp-security: https://docs.atlan.com/product/capabilities/atlan-ai/references/mcp-security (read 13 Nov 2026)
- https://docs.atlan.com/llms/platform/agents/monitor-your-ai-agents-with-atlan/llms.txt: https://docs.atlan.com/llms/platform/agents/monitor-your-ai-agents-with-atlan/llms.txt (read 13 Nov 2026)
- https://github.com/collibra/chip: https://github.com/collibra/chip (read 13 Nov 2026)
- https://docs.databricks.com/aws/en/genie/set-up: https://docs.databricks.com/aws/en/genie/set-up (read 13 Nov 2026)
- https://docs.databricks.com/aws/en/agents/mcp-tools/authentication-networking: https://docs.databricks.com/aws/en/agents/mcp-tools/authentication-networking (read 13 Nov 2026)
- https://docs.databricks.com/aws/en/ai-gateway/govern-mcp-service: https://docs.databricks.com/aws/en/ai-gateway/govern-mcp-service (read 13 Nov 2026)
- https://docs.datahub.com/docs/features/feature-guides/scoped-mcp-servers: https://docs.datahub.com/docs/features/feature-guides/scoped-mcp-servers (read 13 Nov 2026)
- https://docs.datahub.com/docs/features/feature-guides/mcp: https://docs.datahub.com/docs/features/feature-guides/mcp (read 13 Nov 2026)
- https://docs.euno.ai/ai/mcp-setup.md: https://docs.euno.ai/ai/mcp-setup.md (read 13 Nov 2026)
- https://documentation.immuta.com/saas/agentic-data-access/introduction.md: https://documentation.immuta.com/saas/agentic-data-access/introduction.md (read 13 Nov 2026)
- https://www.informatica.com/headless.html: https://www.informatica.com/headless.html (read 13 Nov 2026)
- https://www.informatica.com/blogs/powering-agentic-workflows-on-aws-with-informatica-mcps.html: https://www.informatica.com/blogs/powering-agentic-workflows-on-aws-with-informatica-mcps.html (read 13 Nov 2026)
- https://learn.microsoft.com/purview/ai-microsoft-purview: https://learn.microsoft.com/purview/ai-microsoft-purview (read 13 Nov 2026)
- https://learn.microsoft.com/power-bi/developer/mcp/mcp-servers-overview: https://learn.microsoft.com/power-bi/developer/mcp/mcp-servers-overview (read 13 Nov 2026)
- https://learn.microsoft.com/purview/ai-agents: https://learn.microsoft.com/purview/ai-agents (read 13 Nov 2026)
- https://www.onetrust.com/products/ai-governance/: https://www.onetrust.com/products/ai-governance/ (read 13 Nov 2026)
- https://developer.onetrust.com/onetrust/docs/ai-guard.md: https://developer.onetrust.com/onetrust/docs/ai-guard.md (read 13 Nov 2026)
- https://github.com/onetrust-oss/ai-guard-sdk: https://github.com/onetrust-oss/ai-guard-sdk (read 13 Nov 2026)
- https://developer.onetrust.com/onetrust/reference/mcp.md: https://developer.onetrust.com/onetrust/reference/mcp.md (read 13 Nov 2026)
- https://developer.onetrust.com/onetrust/docs/ai-guard-authentication.md: https://developer.onetrust.com/onetrust/docs/ai-guard-authentication.md (read 13 Nov 2026)
- https://developer.onetrust.com/onetrust/docs/ai-guard-faq.md: https://developer.onetrust.com/onetrust/docs/ai-guard-faq.md (read 13 Nov 2026)
- https://docs.trust3.ai/get-started/overview.html: https://docs.trust3.ai/get-started/overview.html (read 13 Nov 2026)
- https://trust3.ai/: https://trust3.ai/ (read 13 Nov 2026)
- https://docs.trust3.ai/how-to-guides/mcp-configuration.html: https://docs.trust3.ai/how-to-guides/mcp-configuration.html (read 13 Nov 2026)
- https://docs.secoda.co/features/ai-assistant/security.md: https://docs.secoda.co/features/ai-assistant/security.md (read 13 Nov 2026)
- https://docs.secoda.co/resource-and-metadata-management/tags/auto-pii-tagging.md: https://docs.secoda.co/resource-and-metadata-management/tags/auto-pii-tagging.md (read 13 Nov 2026)
- https://securiti.ai/gencore/safe-enterprise-ai-copilot/: https://securiti.ai/gencore/safe-enterprise-ai-copilot/ (read 13 Nov 2026)

Last reviewed 13 Nov 2026. Canonical: https://aidatagovernancetools.com/exposure-map/agent/
