What information can I track during an agent handoff?

You can log the source and destination agents, the specific reason for the handoff, the task summary, and the current state of the workflow.

What frameworks does this skill support?

This skill includes implementation patterns and guidance for popular multi-agent frameworks such as LangGraph, CrewAI, and AutoGen.

Can this skill help visualize parallel agent execution?

Yes, it supports various coordination patterns including Parallel (fan-out/fan-in), Hub-and-Spoke (Supervisor), and Sequential Pipeline architectures.

Why is context propagation important in multi-agent systems?

Context propagation ensures that child agents are correctly linked to their parent spans, preventing orphaned traces and providing a clear, navigable hierarchy of the entire workflow.

Does this skill work with custom agent frameworks?

Absolutely. While it provides specific examples for LangGraph and CrewAI, the core principles and span attributes can be applied to any custom multi-agent architecture.

Multi-Agent Coordination Instrumentation

Name: Multi-Agent Coordination Instrumentation
Author: nexus-labs-automation

bynexus-labs-automation

0•

分析と監視

Instruments complex multi-agent workflows to provide deep visibility into agent handoffs, hierarchies, and coordination patterns.

The Multi-Agent Coordination skill provides a standardized framework for tracing and observing complex AI agent systems within Claude Code. It enables developers to map out intricate parent-child relationships, track task handoffs between supervisor and worker agents, and visualize the complete flow of work across distributed agent architectures. By implementing consistent span attributes and context propagation patterns, it ensures that every step—from initial planning to final synthesis—is transparent and measurable across popular frameworks like LangGraph, CrewAI, and AutoGen, making it essential for debugging and optimizing autonomous agent teams.

主な機能

01Standardized span attributes for agent identity, framework, and workflow state

02Context propagation management to eliminate orphaned traces in complex runs

03Hierarchical trace mapping for parent-child agent relationships

04Detailed handoff tracking with reason-based logging and task summaries

05Pre-built instrumentation patterns for LangGraph, CrewAI, and AutoGen

060 GitHub stars

ユースケース

01Debugging complex supervisor-worker architectures and task delegation

02Auditing multi-agent system decision-making paths for compliance and safety

03Optimizing agent handoff logic to reduce workflow latency and errors

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nexus-labs-automation/agent-observability multi-agent-coordination

For use in Claude.ai and ChatGPT

主な機能

01Standardized span attributes for agent identity, framework, and workflow state

02Context propagation management to eliminate orphaned traces in complex runs

03Hierarchical trace mapping for parent-child agent relationships

04Detailed handoff tracking with reason-based logging and task summaries

05Pre-built instrumentation patterns for LangGraph, CrewAI, and AutoGen

060 GitHub stars

ユースケース

01Debugging complex supervisor-worker architectures and task delegation

02Auditing multi-agent system decision-making paths for compliance and safety

03Optimizing agent handoff logic to reduce workflow latency and errors

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nexus-labs-automation/agent-observability multi-agent-coordination

For use in Claude.ai and ChatGPT