How does this skill improve agent reliability?

It provides guardrails, suggests self-correction reflection patterns, and emphasizes the principle that autonomy should be added incrementally only after reliability is proven.

What is the primary focus of the Autonomous Agents skill?

The skill focuses on making AI agents reliable for production use by implementing specific architectural patterns like ReAct and Plan-Execute while managing compounding error rates in multi-step tasks.

Which agent patterns are supported?

This skill covers the ReAct loop (Reasoning + Action), the Plan-Execute pattern for separating strategy from implementation, and Reflection patterns for self-evaluation.

What are the common pitfalls this skill helps avoid?

It specifically addresses anti-patterns such as unbounded autonomy, trusting unverified agent outputs, and the lack of hard cost or step limits.

Autonomous Agent Engineering

Name: Autonomous Agent Engineering
Author: claudiodearaujo

byclaudiodearaujo

データサイエンスとML

Provides architectural patterns and reliability guardrails for building production-grade autonomous AI systems.

概要

This skill equips Claude with the expertise of an agent architect to design, implement, and refine autonomous AI systems using industry-standard patterns like ReAct and Plan-Execute. It prioritizes system reliability by addressing the compounding error rates inherent in multi-step processes, offering specific guidance on goal decomposition, self-reflection, and the implementation of essential guardrails. By focusing on earning autonomy through proven reliability rather than unbounded capability, this skill helps developers move from fragile AI demos to robust production agents.

主な機能

Enforces production reliability through error rate reduction
0 GitHub stars
Implements ReAct and Plan-Execute agent loops
Identifies and avoids common agent anti-patterns
Establishes operational guardrails and cost limits
Provides strategies for goal decomposition and self-correction

ユースケース

Implementing structured reflection patterns for iterative agent improvement
Designing complex multi-step AI workflows that require independent decision-making
Optimizing existing agentic systems to reduce failure rates and improve consistency

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add claudiodearaujo/sistema-de-narra-o-de-livro-front autonomous-agents

For use in Claude.ai and ChatGPT

Download Skill

GitHub

概要