Weak-Link Optimization for Multi-Agent Reasoning and Collaboration

Develop an automated debugging tool that identifies 'weak-link' agents in a multi-agent swarm by tracking cascading error rates. Create a library that wraps agent steps in a verification layer to prevent error amplification.

Suggested repo: agent-audit

"Identify the failing agent in your chain before the whole system crashes."

Estimated effort: 80h