Python: New Feature: pre-built ReAct agents?

[smartnose]

Basically, can we create a pre-built agent for Semantic Kernel using the ReAct pattern like the one provided in LangChain?

https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/agents/react/agent.py

This pattern is very useful for real world robust applications, esp. when I need the agent to be able to recover from policy errors or other transient errors gracefully, reason, replan and retry.

For a toy example illustrating why ReAct is important, see this repo I added to illustrate the issue.

https://github.com/smartnose/policy-enforcer/tree/main/policy_enforcer

name: Feature request
about: Suggest an idea for this project

---

[markwallace-microsoft]

Adding this to the current sprint so I can better understand the ask and make a plan

[smartnose]

Thank you for consideration, Mark. I saw you labelled it under .Net. My specific use case is more on the python front (a data point for your prioritization)