Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
I see some similar issues here. I want to specific it in detail and wish crewai team or other experts can solve the problem. I create a agent which play guessing number with human. The agent provide a number and human provide feedback. I want the task cannot stop until the agent reach the right number. But the task always run once and I can not solve it.
player = Agent(
role="Game Player",
goal=(
"guess the right number from 1 to 100 which just human know"
"and summarize a strategy that can help yourself to reach the right number as soon as possible"
),
verbose=True,
backstory=(
"You play a guessing number game with human."
"Human know the number and You don't know the right number."
"When you provide a number, human will give you the feedback."
"You should try your best to guess the number as soon as possible."
"Finally you summarize the strategy to reach the right number."
)
)
guess = Task(
description="Guess the right number. This task cannot stop until player reach the right number",
expected_output="summary a strategy regarding how to guess number",
human_input=True,
output_file="strategy.md",
agent=player,
)
crew = Crew(
agents=[player],
tasks=[guess],
memory=True,
verbose=True
)
result = crew.kickoff()
just in terms of this issue, using from langchain.agents import load_tools human_tools = load_tools(["human"]), not the human_input=True can solve the problem
I see some similar issues here. I want to specific it in detail and wish crewai team or other experts can solve the problem. I create a agent which play guessing number with human. The agent provide a number and human provide feedback. I want the task cannot stop until the agent reach the right number. But the task always run once and I can not solve it.