Can I use these patterns for browser automation?

Absolutely. The skill covers the architectural patterns necessary to integrate browser-based tools and screenshot capabilities into an agent's observation loop.

Does this skill provide code for tool calling?

Yes, it includes detailed Python implementation examples for tool schemas, file operations, and terminal execution handlers.

How does this skill handle agent safety?

It provides structured patterns for multi-level permission systems (Auto, Ask Once, Ask Each) and sandboxed execution environments to prevent unauthorized system changes.

What are autonomous agent patterns?

They are standardized architectural designs and best practices for creating AI systems that can reason, use tools, and complete complex coding tasks independently.

Autonomous Agent Patterns

Name: Autonomous Agent Patterns
Author: claudiodearaujo

byclaudiodearaujo

0•

数据科学与机器学习

Provides architectural patterns and implementation guides for building robust, secure autonomous AI coding agents.

This skill serves as a comprehensive blueprint for developers building autonomous coding assistants and AI agents. It details the 'Think-Decide-Act-Observe' loop, offering concrete implementation patterns for multi-model architectures, JSON tool schemas, and precise file-editing logic. Beyond core logic, the skill prioritizes security and reliability by providing frameworks for multi-level permission systems, human-in-the-loop approval UIs, and sandboxed execution environments, ensuring that AI agents operate within safe and predictable boundaries.

主要功能

01Sandboxed environment implementation for safe autonomous actions

020 GitHub stars

03Scalable tool design patterns with JSON Schema integration

04Precise file editing and terminal execution strategies

05Standardized Think-Decide-Act-Observe agent loop architecture

06Multi-level permission and human-in-the-loop approval systems

使用场景

01Designing secure tool-calling APIs for AI-powered DevOps utilities

02Building a custom autonomous coding assistant for internal software repositories

03Implementing browser-based automation workflows with strict permission gates

What are Skills?·How to Install

Install with 🐟 Skill.Fish

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

For use in Claude.ai and ChatGPT

Download Skill

主要功能

01Sandboxed environment implementation for safe autonomous actions

020 GitHub stars

03Scalable tool design patterns with JSON Schema integration

04Precise file editing and terminal execution strategies

05Standardized Think-Decide-Act-Observe agent loop architecture

06Multi-level permission and human-in-the-loop approval systems

使用场景

01Designing secure tool-calling APIs for AI-powered DevOps utilities

02Building a custom autonomous coding assistant for internal software repositories

03Implementing browser-based automation workflows with strict permission gates

What are Skills?·How to Install

Install with 🐟 Skill.Fish

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

For use in Claude.ai and ChatGPT

Download Skill