CrewAI Skill
Alpha Software Notice
This skill is in alpha stage and under active development. APIs, features, and functionality may change without notice. Use with caution in production environments and expect potential breaking changes in future releases.
Simplified CrewAI integration for multi-agent orchestration. Execute pre-configured crews of specialized AI agents with collaborative workflows.
Features
- Pre-configured Crews: Set up crews during skill initialization
- Single Tool Execution: One simple tool to run configured crews
- Multi-Agent Orchestration: Specialized AI agents with defined roles
- Process Management: Sequential and hierarchical workflow execution
- Input-driven Execution: Provide inputs to drive crew workflows
- No Dependencies: Simplified architecture with no external skill dependencies
Quick Setup
Prerequisites
Install CrewAI:
Option 1: Using CrewAI Crew Object (Recommended)
from webagents.agents import BaseAgent
from webagents.agents.skills.ecosystem.crewai import CrewAISkill
from crewai import Agent, Task, Crew, Process
# Create CrewAI agents and tasks using native API
researcher = Agent(
role='Senior Researcher',
goal='Uncover groundbreaking technologies in AI for year 2024',
backstory='Driven by curiosity, you explore and share the latest innovations.',
verbose=True
)
research_task = Task(
description='Identify the next big trend in {topic} with pros and cons.',
expected_output='A 3-paragraph report on emerging {topic} technologies.',
agent=researcher
)
# Form the crew
crew = Crew(
agents=[researcher],
tasks=[research_task],
process=Process.sequential,
verbose=True
)
# Create agent with CrewAI crew object
agent = BaseAgent(
name="research-agent",
model="openai/gpt-4o",
skills={
"crewai": CrewAISkill(crew) # Pass Crew object directly
}
)
Option 2: Using Configuration Dictionary
from webagents.agents import BaseAgent
from webagents.agents.skills.ecosystem.crewai import CrewAISkill
# Define your crew configuration
crew_config = {
'agents': [
{
'role': 'Senior Researcher',
'goal': 'Uncover groundbreaking technologies in AI for year 2024',
'backstory': 'Driven by curiosity, you explore and share the latest innovations.',
'verbose': True
}
],
'tasks': [
{
'description': 'Identify the next big trend in {topic} with pros and cons.',
'expected_output': 'A 3-paragraph report on emerging {topic} technologies.',
'agent_index': 0
}
],
'process': 'sequential',
'verbose': True
}
# Create agent with configured CrewAI skill
agent = BaseAgent(
name="research-agent",
model="openai/gpt-4o",
skills={
"crewai": CrewAISkill(crew_config)
}
)
Usage
Core Tool
Tool: crewai_run(inputs)
Execute the configured CrewAI crew with the provided inputs.
# Example usage via LLM
messages = [{
'role': 'user',
'content': 'Research AI Agents and their applications'
}]
response = await agent.run(messages=messages)
The LLM will automatically call crewai_run({'topic': 'AI Agents'}) based on the crew configuration and user request.
Use Case Example: Research Team
This example shows a complete research crew that analyzes topics and provides insights:
Using CrewAI Objects (Recommended)
from crewai import Agent, Task, Crew, Process
# Create specialized agents
researcher = Agent(
role='Senior Researcher',
goal='Conduct comprehensive research on specified topics',
backstory='You are an experienced researcher with access to latest information and analytical skills.'
)
analyst = Agent(
role='Data Analyst',
goal='Analyze research data and extract key insights',
backstory='You specialize in data analysis and pattern recognition.'
)
# Create sequential tasks
research_task = Task(
description='Research the current state of {topic} technology and market trends',
expected_output='Detailed research report with key findings',
agent=researcher
)
analysis_task = Task(
description='Analyze the research data and identify the top 3 opportunities in {topic}',
expected_output='Analysis report with ranked opportunities and recommendations',
agent=analyst
)
# Form the crew
research_crew = Crew(
agents=[researcher, analyst],
tasks=[research_task, analysis_task],
process=Process.sequential
)
# Initialize skill with crew
agent = BaseAgent(
name="research-agent",
model="openai/gpt-4o",
skills={
"crewai": CrewAISkill(research_crew)
}
)
Using Configuration Dictionary
research_crew_config = {
'agents': [
{
'role': 'Senior Researcher',
'goal': 'Conduct comprehensive research on specified topics',
'backstory': 'You are an experienced researcher with access to latest information and analytical skills.'
},
{
'role': 'Data Analyst',
'goal': 'Analyze research data and extract key insights',
'backstory': 'You specialize in data analysis and pattern recognition.'
}
],
'tasks': [
{
'description': 'Research the current state of {topic} technology and market trends',
'expected_output': 'Detailed research report with key findings',
'agent_index': 0
},
{
'description': 'Analyze the research data and identify the top 3 opportunities in {topic}',
'expected_output': 'Analysis report with ranked opportunities and recommendations',
'agent_index': 1
}
],
'process': 'sequential'
}
# Initialize skill with configuration
agent = BaseAgent(
name="research-agent",
model="openai/gpt-4o",
skills={
"crewai": CrewAISkill(research_crew_config)
}
)
Example Interaction
# User asks for research
messages = [{
'role': 'user',
'content': 'Research the latest developments in quantum computing and identify business opportunities'
}]
# The LLM automatically:
# 1. Recognizes this as a research request
# 2. Calls crewai_run({'topic': 'quantum computing'})
# 3. The researcher agent researches quantum computing trends
# 4. The analyst agent identifies top 3 business opportunities
# 5. Returns comprehensive analysis
response = await agent.run(messages=messages)
print(response)
Configuration Reference
crew_config = {
'agents': [
{
'role': 'Agent Role/Title', # Required
'goal': 'Agent primary objective', # Required
'backstory': 'Agent background', # Required
'verbose': True, # Optional
'allow_delegation': False # Optional
}
],
'tasks': [
{
'description': 'Task with {input_variable} placeholders', # Required
'expected_output': 'Expected deliverable format', # Optional
'agent_index': 0 # Optional, defaults to 0
}
],
'process': 'sequential', # 'sequential' or 'hierarchical'
'verbose': True # Optional, defaults to True
}
Troubleshooting
"No CrewAI crew configured" - Initialize the skill with a valid crew configuration or Crew object
"Inputs are required to run the crew" - Ensure your tasks use input variables like {topic} that the LLM can populate
"CrewAI execution failed" - Check CrewAI installation: pip install crewai