Skip to content

Refactor AndersAgent implementation and update lab documentation#3

Open
oaviles wants to merge 14 commits intowarnov:masterfrom
oaviles:master
Open

Refactor AndersAgent implementation and update lab documentation#3
oaviles wants to merge 14 commits intowarnov:masterfrom
oaviles:master

Conversation

@oaviles
Copy link
Copy Markdown

@oaviles oaviles commented Apr 7, 2026

This pull request removes the old ai-foundry implementation and introduces a new, modernized Microsoft Foundry agent for the AndersAgent project. The new implementation adopts the latest Azure AI SDKs, restructures configuration, and simplifies the agent's setup and usage. The most significant changes are summarized below.

Migration to Microsoft Foundry SDK and Modernization:

  • Replaced the previous ai-foundry implementation with a new ms-foundry version using the latest Azure.AI.Projects and Microsoft.Agents.AI SDKs, enabling a streamlined and future-proof agent experience. (Program.cs, AndersAgent.csproj) [1] [2] [3]

Configuration and Project Structure Updates:

  • Updated configuration files to use a new nested Foundry section, reflecting the new SDK's requirements and improving clarity and maintainability. (appsettings.json) [1] [2]
  • Removed obsolete project files and configuration from the old ai-foundry implementation. (AndersAgent.csproj, appsettings.json) [1] [2]

Agent Logic and Features:

  • Introduced a sample tool (GetWeather) to demonstrate function calling with the new SDK, and restructured agent creation and chat interaction to use modern async streaming APIs for multi-turn conversations. (Program.cs)
  • Simplified agent deployment and verification via command-line arguments, and ensured agents are cleaned up after runs for resource management. (Program.cs)

These changes collectively modernize the agent implementation, make configuration more robust, and align the project with current Azure AI best practices.

oaviles and others added 14 commits April 7, 2026 18:55
- Replaced the existing agent code with a new implementation using the Azure.AI.Projects and Microsoft.Agents.AI namespaces.
- Updated appsettings.json to reflect new configuration structure.
- Removed obsolete nuget.config file.
- Simplified the lab documentation, focusing on the new agent implementation and its features.
- Added deployment scripts (deploy-foundry-agent.ps1 and deploy-foundry-agent.sh) for easier agent deployment.
- Created a new project file (ms_foundry_agent.csproj) and solution file (ms_foundry_agent_v2.sln) for the updated agent.
- Introduced a simple weather function as a demonstration tool for the agent.
…entation

- Introduced Lab 4 documentation detailing the setup and functionality of the Anders Executor Agent, including OpenAPI integration and agent deployment steps.
- Added Lab 5 documentation for the Julie Planner Agent, outlining its architecture, workflow orchestration, and interaction with SQL and marketing agents.
- Included detailed instructions for configuring environment variables, validating permissions, and executing end-to-end workflows for both agents.
…, orchestration, and publishing

- Created lab06 for setting up the Microsoft Copilot Studio environment.
- Added lab07 for building the Charlie agent focused on product analysis.
- Introduced lab08 for creating Ric as a child agent of Bill for email notifications.
- Developed lab09 for orchestrating agents Mark, Anders, and Charlie under Bill.
- Implemented lab10 for publishing the Bill agent and testing through Microsoft 365.
- Removed outdated lab10-ric-child-agent.md file.
…gente planificador en la documentación de Foundry
…y-script

fix: add execute permission to deploy-foundry-agent.sh
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants