sandbox-crewai

🤖 A sandbox for experimenting with CrewAI multi-agent AI systems that research topics and generate comprehensive reports 🔬

CrewAI Docs · Agents Guide · Tasks Guide

Overview

This project provides a ready-to-use sandbox for exploring the CrewAI framework. It includes a sample crew implementation called latest_ai_development that demonstrates a two-agent research and reporting workflow:

• Researcher Agent - Discovers cutting-edge developments on any topic
• Reporting Analyst Agent - Transforms research into detailed, structured reports

Features

• Decorator-based architecture - Clean @agent, @task, and @crew decorators for defining crews
• YAML configuration - Externalized agent and task definitions for easy customization
• Template variables - Dynamic {topic} interpolation across all configurations
• Sequential processing - Tasks execute in order with output chaining
• Markdown output - Reports generated as formatted report.md files

Quick Start

# Clone and setup
git clone https://github.com/tsilva/sandbox-crewai.git
cd sandbox-crewai
conda env create -f environment.yml
conda activate crewai-sandbox

# Configure API key
cp .env.example .env
# Edit .env and add your OPENAI_API_KEY

# Run the crew
python main.py

Installation

Prerequisites

• Conda or Mamba
• OpenAI API key

Environment Setup

# Create the conda environment
conda env create -f environment.yml

# Activate it
conda activate crewai-sandbox

Configuration

Create a .env file in the project root:

OPENAI_API_KEY=your-api-key-here

Usage

Run the Crew

# From project root
python main.py

This runs the default topic "AI Agents". To research a different topic, edit main.py:

inputs = {
    'topic': 'Your Topic Here'
}

Alternative Entry Points

# Using the crew's CLI entry point
cd latest_ai_development
run_crew

# Train the crew
train <n_iterations> <filename>

# Test the crew
test <n_iterations> <openai_model_name>

# Replay a specific task
replay <task_id>

Architecture

sandbox-crewai/
├── main.py                          # Main entry point
├── environment.yml                  # Conda environment definition
├── .env.example                     # Environment template
└── latest_ai_development/
    └── src/latest_ai_development/
        ├── crew.py                  # Crew definition with decorators
        ├── main.py                  # CLI entry points
        ├── config/
        │   ├── agents.yaml          # Agent definitions
        │   └── tasks.yaml           # Task definitions
        └── tools/
            └── custom_tool.py       # Custom tools (extensible)

Crew Flow

graph LR
    A[Input Topic] --> B[Researcher Agent]
    B --> C[10 Bullet Points]
    C --> D[Reporting Analyst]
    D --> E[report.md]

Loading

Agent Configuration

Agents are defined in config/agents.yaml:

Agent	Role	Output
Researcher	{topic} Senior Data Researcher	10 bullet points of findings
Reporting Analyst	{topic} Reporting Analyst	Full markdown report

Task Configuration

Tasks are defined in config/tasks.yaml:

Task	Description	Agent
research_task	Conduct thorough research on {topic}	Researcher
reporting_task	Expand findings into detailed report sections	Reporting Analyst

Customization

Adding New Agents

1. Add agent definition to config/agents.yaml:

new_agent:
  role: >
    {topic} Expert Role
  goal: >
    Your agent's goal
  backstory: >
    Agent background story

2. Register in crew.py:

@agent
def new_agent(self) -> Agent:
    return Agent(
        config=self.agents_config['new_agent'],
        verbose=True
    )

Adding New Tasks

1. Add task definition to config/tasks.yaml:

new_task:
  description: >
    Task description with {topic} variable
  expected_output: >
    What the task should produce
  agent: new_agent

2. Register in crew.py:

@task
def new_task(self) -> Task:
    return Task(
        config=self.tasks_config['new_task'],
    )

Adding Custom Tools

Create tools in tools/custom_tool.py and attach to agents:

from crewai.tools import tool

@tool
def my_custom_tool(input: str) -> str:
    """Tool description for the agent."""
    return f"Processed: {input}"

Dependencies

Package	Purpose
crewai	Multi-agent orchestration framework
crewai[tools]	Built-in tools for agents
crewai[embeddings]	Embedding support
python-dotenv	Environment variable management

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is available under the MIT License.

Resources

• CrewAI Documentation
• CrewAI GitHub
• Agents Concept
• Tasks Concept
• Crews Concept