Does this skill facilitate parallel execution?

Yes, patterns like Swarm and MapReduce utilize parallel execution and fan-out primitives to handle tasks across multiple agents simultaneously.

What patterns does Agentica Infrastructure support?

It supports 11 patterns including Swarm, Pipeline, Hierarchical, Jury, Generator-Critic, CircuitBreaker, Adversarial, ChainOfResponsibility, MapReduce, Blackboard, and EventDriven.

How is agent state managed in this infrastructure?

State is managed through a CoordinationDB for SQLite tracking, a BlackboardCache for hot-tier communication, and a MemoryService for core and archival storage.

Can I use this skill to debug agent handoffs?

Yes, the skill provides exact signatures for HandoffAtoms and tracking databases specifically designed for orphan detection and agent tracking.

Agentica Multi-Agent Infrastructure

Name: Agentica Multi-Agent Infrastructure
Author: scooter-lacroix

byscooter-lacroix

数据科学与机器学习

Orchestrates complex multi-agent AI workflows using standardized coordination patterns and robust infrastructure primitives.

关于

Agentica Multi-Agent Infrastructure is a comprehensive reference guide designed to help developers build, debug, and optimize sophisticated AI agent ecosystems. It provides a blueprint for eleven distinct coordination patterns—such as Swarms, Pipelines, and Juries—enabling structured communication and task delegation. By offering detailed specifications for SQLite-based tracking, shared blackboard caches, and universal handoff atoms, this skill ensures that multi-agent systems remain observable, stateful, and resilient during complex problem-solving tasks.

主要功能

0 GitHub stars
Universal HandoffAtom format for structured data exchange between agents
Fine-grained consensus and aggregation primitives for agent voting
SQLite-based CoordinationDB for persistent session and agent tracking
Library of 11 coordination patterns including Swarm, Pipeline, and MapReduce
Blackboard and MemoryService for shared state and long-term archival

使用场景

Creating an iterative document refinement loop using the Generator-Critic pattern
Building a multi-expert code analysis system using parallel Swarm perspectives
Implementing a sequential data processing pipeline with automated stage transitions

关于

主要功能

0 GitHub stars
Universal HandoffAtom format for structured data exchange between agents
Fine-grained consensus and aggregation primitives for agent voting
SQLite-based CoordinationDB for persistent session and agent tracking
Library of 11 coordination patterns including Swarm, Pipeline, and MapReduce
Blackboard and MemoryService for shared state and long-term archival

使用场景

Creating an iterative document refinement loop using the Generator-Critic pattern
Building a multi-expert code analysis system using parallel Swarm perspectives
Implementing a sequential data processing pipeline with automated stage transitions