Does this support timeouts for individual tasks?

Yes, the skill includes patterns for wrapping parallel tasks in asyncio.wait_for to prevent slow branches from blocking the overall execution.

How is state managed when multiple nodes write at once?

It utilizes LangGraph’s state reducers, specifically the Annotated[list, add] pattern, to safely accumulate results from parallel branches into a single state key.

How does this skill handle errors in parallel branches?

It implements error isolation patterns using asyncio.gather with return_exceptions=True, ensuring that one failing branch doesn't crash the entire workflow.

What is the benefit of using the Send API in LangGraph?

The Send API allows for dynamic fan-out, enabling you to launch multiple parallel branches of a workflow based on runtime data rather than a fixed graph structure.

Can I use this for multi-agent systems?

Yes, it is specifically designed to run multiple specialized agents (like security, tech, and quality agents) concurrently to provide faster, comprehensive analysis.

LangGraph Parallel Execution

Name: LangGraph Parallel Execution
Author: yonatangross

byyonatangross

•

データサイエンスとML

Implements high-performance parallel execution patterns for LangGraph workflows using fan-out/fan-in and map-reduce strategies.

This skill provides production-ready patterns for executing independent nodes concurrently within LangGraph, significantly reducing latency in AI workflows. It empowers developers to implement the Send API for dynamic fan-out, manage parallel agent analysis, and utilize map-reduce patterns for large-scale tasks. With built-in guidance on error isolation, per-branch timeouts, and state accumulation using annotated reducers, it ensures that complex multi-agent systems remain performant, resilient, and scalable.

主な機能

01Production-ready map-reduce patterns for concurrent task processing

0269 GitHub stars

03Dynamic fan-out/fan-in using the LangGraph Send API

04Parallel multi-agent orchestration for comprehensive analysis

05Robust error isolation and individual branch timeout management

06State accumulation patterns with Annotated list reducers

ユースケース

01Processing large datasets through parallel map-reduce workflows

02Running multiple specialized agents concurrently to analyze a single input

03Reducing latency in AI pipelines by executing independent steps simultaneously

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yonatangross/orchestkit langgraph-parallel

For use in Claude.ai and ChatGPT

Download Skill