How does it handle agent memory?

The skill focuses on hierarchical memory systems, teaching you how to manage short-term working memory, episodic memory for past interactions, and long-term semantic memory using RAG to maintain context without bloating the context window.

What agent patterns does this skill support?

It supports a variety of architectural patterns including ReAct (Reason-Act) loops for step-by-step execution, Plan-and-Execute for complex task decomposition, and the Supervisor pattern for coordinating multiple specialist agents.

When should I use multi-agent systems over a single agent?

The AI Agents Architect provides logic to justify multi-agent overhead, recommending them only when tasks require truly independent specialized skills that would overwhelm a single agent's tool set or reasoning capacity.

Does it provide guidance on preventing runaway API costs?

Yes, a core focus of the skill is setting strict iteration limits, using circuit breakers for tool failures, and implementing selective memory to ensure agents don't enter infinite loops or waste tokens.

AI Agents Architect

Name: AI Agents Architect
Author: sickn33

bysickn33

•

31,722

•

データサイエンスとML

Designs and builds robust autonomous AI agents using advanced orchestration, tool management, and hierarchical memory systems.

The AI Agents Architect skill transforms Claude into an expert systems designer for autonomous AI. It provides a comprehensive framework for implementing sophisticated agentic patterns like ReAct and Plan-and-Execute while focusing heavily on system reliability and graceful degradation. By mastering tool registries, multi-level memory architectures, and multi-agent orchestration, this skill helps developers build agents that are powerful yet controllable, ensuring they fail loudly when necessary and operate efficiently within token and cost constraints.

主な機能

01Observability, tracing, and checkpoint recovery strategies

0231,722 GitHub stars

03Hierarchical memory management including episodic and semantic systems

04Comprehensive tool registry and function calling optimization

05Multi-agent orchestration and supervisor pattern implementation

06Advanced agent loop design (ReAct, Plan-and-Execute, and more)

ユースケース

01Building autonomous research agents that navigate complex tool ecosystems

02Implementing long-running AI assistants with persistent state and RAG-based memory

03Developing multi-agent developer workflows for specialized coding tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sickn33/antigravity-awesome-skills ai-agents-architect

For use in Claude.ai and ChatGPT

主な機能

01Observability, tracing, and checkpoint recovery strategies

0231,722 GitHub stars

03Hierarchical memory management including episodic and semantic systems

04Comprehensive tool registry and function calling optimization

05Multi-agent orchestration and supervisor pattern implementation

06Advanced agent loop design (ReAct, Plan-and-Execute, and more)

ユースケース

01Building autonomous research agents that navigate complex tool ecosystems

02Implementing long-running AI assistants with persistent state and RAG-based memory

03Developing multi-agent developer workflows for specialized coding tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sickn33/antigravity-awesome-skills ai-agents-architect

For use in Claude.ai and ChatGPT