What is the core agent loop provided in this skill?

The core agent loop follows a Think-Decide-Act-Observe pattern where the AI reasons through a task, plans a tool call, executes the action, and incorporates the result back into its context.

Does it support using different LLMs for different tasks?

Yes, it includes a Multi-Model Architecture pattern that demonstrates how to route planning, reasoning, and code generation to the most appropriate models.

How does this skill address agent security?

It provides implementation patterns for sandboxing execution, validating paths, and a multi-tiered permission system that distinguishes between low-risk and dangerous commands.

What are the essential tools covered for coding agents?

The skill outlines schemas for file operations (read, write, edit), terminal interaction, browser automation, and context-gathering tools like web search and user prompts.

Autonomous Agent Patterns

Name: Autonomous Agent Patterns
Author: claudiodearaujo

byclaudiodearaujo

•

생산성 및 워크플로

Provides architectural design patterns and implementation strategies for building secure, autonomous AI coding agents.

This skill serves as a comprehensive reference for developers building autonomous software agents. It covers critical architectural components including the Think-Decide-Act-Observe loop, multi-model orchestration, and standardized tool schemas for file manipulation and terminal access. By emphasizing security through granular permission levels, sandboxing, and human-in-the-loop approval workflows, it enables the creation of agents that can reliably perform complex coding tasks while maintaining strict safety standards.

주요 기능

011 GitHub stars

02Standardized agent loop implementation (Think-Decide-Act-Observe)

03Comprehensive tool schemas for file operations and terminal execution

04Multi-model architecture for specialized task orchestration

05Granular permission systems with auto-approval and manual check gates

06Sandboxed execution patterns for secure command processing

사용 사례

01Designing human-in-the-loop workflows for sensitive system operations

02Building a self-healing autonomous coding assistant

03Implementing secure tool-calling interfaces for LLM applications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add claudiodearaujo/sistema-de-narra-o-de-livro autonomous-agent-patterns

For use in Claude.ai and ChatGPT

Download Skill