Atomic Agents Framework Guide

Features

• Guides on core Atomic Agents abstractions (schemas, agents, tools, context providers)
• Provides canonical import paths and minimum viable agent examples
• Directs users to specialized skills for schema, agent, tool, and context provider creation
• Details system prompt structure and LLM provider integration
• References files for deeper dives into specific concepts like memory, orchestration, and testing

Use Cases

• Use when writing Python code that imports from `atomic_agents`.
• Use when defining an `AtomicAgent`, `BaseTool`, or `BaseIOSchema`.
• Use when asking about multi-agent orchestration or LLM-provider wiring in an atomic-agents project.
• Use when needing guidance on project structure, testing, or error handling within the framework.

Non-Goals

• This skill does not provide direct execution capabilities.
• It does not replace the need to read the official Atomic Agents documentation for in-depth information.
• It does not offer a GUI or interactive interface; interaction is through Claude's chat.

Workflow

1. User asks about Atomic Agents framework, specific components, or common tasks.
2. Skill identifies relevant sections within SKILL.md or linked reference files.
3. Skill provides concise explanations, code snippets, and directs users to deeper documentation or specialized skills.
4. User follows guidance to understand and implement Atomic Agents components.

Practices

• Framework Guidance
• LLM Application Development
• Agent Design Patterns

Prerequisites

• Claude Code environment