Module assessment

1.

In a hub-and-spoke orchestration, a spoke agent (market data feed) starts failing consistently with timeout errors. Which approach best maintains overall system reliability?

Remove the failing spoke agent from the hub's tool registry and redeploy the hub agent.

Implement a circuit-breaker on the hub that stops delegating to the failing spoke after a threshold of consecutive failures, and return a graceful fallback response for that spoke's domain.

Catch the timeout exception in the hub and retry the spoke three times before proceeding with no result for that domain.

2.

You're implementing parallel agent execution using asyncio.gather for four independent research agents. One agent is optional context enrichment, while three are required for the final report. Which approach correctly handles this requirement?

Use asyncio.gather with return_exceptions=True, then check results—treat any exception in the three required agents as a fatal failure, but allow the optional agent to fail gracefully.

Run the optional agent separately with a try/except, then use asyncio.gather for the three required agents.

Wrap all four agents in asyncio.gather with return_exceptions=False and catch the outer exception if any agent fails.

3.

Your team needs to build a multi-agent code review system that must maintain strict, auditable control flow—different agents trigger based on specific code change types (security-sensitive files, test coverage issues, style violations). Which orchestration framework best fits this requirement?

AutoGen, because its conversation-based model allows agents to discuss and critique code collaboratively.

CrewAI, because its role-based crew model assigns reviewers as specialized agents with defined responsibilities.

LangGraph, because its graph-based state machine allows you to define conditional edges that route to reviewer agents based on explicit code change analysis conditions.

Feedback