tn updates

Author: PatrickFarley

articles/ai-foundry/responsible-ai/agents/transparency-note.md

@@ -94,7 +94,7 @@ Developers can configure a rich ecosystem of  knowledge sources to enable an Age
 * **File Search** (a built-in retrieval-augmented generation (RAG) tool to process and search through private data in Azure AI Search, Azure Blob Storage, and local files)
 * **Grounding with Bing Search** (a web search tool that uses Bing Search to extract information from the web)
 * **SharePoint** (built-in tools that connect an organization’s internal documents in SharePoint for grounded responses)
-* **Fabric AI Skills** (a built-in tool to chat with structured data on Microsoft Fabric using generative AI)
+* **Fabric Data Agent** (a built-in tool to chat with structured data on Microsoft Fabric using generative AI)
 * **Bring your licensed data** (a tool that enables grounding using proprietary data accessed using a licensed API key obtained by the developer from the data provider, for example, TripAdvisor)
 
 Agents simplify secure data access to SharePoint and Fabric AI Skills through on-behalf-of (OBO) authentication, allowing the Agent to access only the SharePoint or Fabric files for which the user has permissions.
@@ -107,6 +107,7 @@ Developers can connect an Agent to external systems, APIs, and services through
 * **Azure Logic Apps** (a cloud-based PaaS tool that enables automated workflows using 1,400+ built-in connectors)
 * **Azure Functions** (a tool that enables an Agent to execute serverless code for synchronous, asynchronous, long-running, and event-driven actions)
 * **OpenAPI 3.0 specified tools** (a custom function defined with OpenAPI 3.0 specification to connect an Agent to external OpenAPI-based APIs securely)
+* **Model Context Protocol tools** (a custom service connected via Model Context Protocol through an existing remote MCP server to an Agent). 
 
 #### Orchestrating multi-agent systems
 
@@ -140,6 +141,7 @@ We encourage customers to use Azure AI Agent Service in their innovative solutio
 - **Avoid high-stakes scenarios that could lead to harm**. The model used in an agent may reflect certain societal views, biases, and other undesirable content present in the training data or the examples provided in the prompt. As a result, we caution against using agents in high-stakes scenarios where unfair, unreliable, or offensive behavior might be extremely costly or lead to harm. 
 - **Carefully consider use cases in high stakes domains or industry where Agent actions are irreversible or highly consequential**. Such industries include but are not limited to healthcare, medicine, finance, or legal domains. For example: the ability to make financial transactions or give financial advice, the ability to directly interact with outside services, the ability to administer medicine or give health-related advice, the ability to share sensitive information publicly, or the ability to grant access to critical systems.  
 - **Legal and regulatory considerations**. Microsoft takes safety and compliance with legal and regulatory obligations seriously. We always strive to abide by applicable laws, regulations, and standards in developing and deploying AI technologies, including the Microsoft Responsible AI Standard. It is your organization’s responsibility to evaluate safety implications and potential specific legal and regulatory obligations when using any AI services and solutions, including agents, Agents Catalog, and underlying Agent code samples. Certain uses and offerings may be subject to legal and regulatory requirements, may require licenses, or may not be suitable for all industries, scenarios, or use cases. Additionally, agents, Agents Catalog, and underlying Agent code samples may not be used in ways prohibited by applicable laws, regulations, terms of service, or relevant codes of conduct. 
+- Microsoft did not create, test, or verify any third-party systems, APIs, servers and services you may decide to connect to. When you connect to a third-party (non-Microsoft) system, API, server, or service, some data will be shared with that service, and your application may receive data in return. Microsoft has no responsibility to you or others in relation to your use of any remote systems, APIs, servers and services. Your use of these services is governed by your agreement with the provider. You are responsible for any usage and associated costs.
 
 ## Limitations
 
@@ -171,7 +173,7 @@ We encourage customers to use Azure AI Agent Service in their innovative solutio
 ## Evaluating and integrating Azure AI Agent Service for your use
 
 * **Map Agent risks and impacts.** Before developing or deploying your agentic application, carefully consider the impact of the intended actions and the consequences of actions or tool use not working as intended – such as generating or taking action on inaccurate information or causing biased or unfair outcomes – at different stages.
-* **Ensure adequate human oversight and control.**  Consider including controls to help users verify, review and/or approve actions in a timely manner, which may include reviewing planned tasks or calls to external data sources, for example, as appropriate for your system. Consider including controls for adequate user remediation of system failures, particularly in high-risk scenarios and use cases.
+* **Ensure adequate human oversight and control.** Consider including controls to help users verify, review and/or approve actions in a timely manner, which may include reviewing planned tasks or calls to external data sources, for example, as appropriate for your system. Consider including controls for adequate user remediation of system failures, particularly in high-risk scenarios and use cases. As an example, the MCP tool allows you to pass custom headers such as authentication keys or schema as may be needed by a remote MCP server. In cases such as this, we recommend you review all data being shared with remote servers and optionally logging it for auditing purposes. Be cognizant of third party practices for retention and location of data.
 * **Clearly define actions and associated requirements.** Clearly defining which actions are allowed (action boundaries), prohibited, or need explicit authorization may help your agentic system operate as expected and with the appropriate level of human oversight.
 * **Clearly define intended operating environments.** Clearly define the intended operating environments (domain boundaries) where your agent is designed to perform effectively.
 * **Ensure appropriate intelligibility in decision making.** Providing information to users before, during, and after actions are taken and/or tools are called may help them understand action justification or why certain actions were taken or the application is behaving a certain way, where to intervene, and how to troubleshoot issues.