๐ง AI-Powered Code Review ToolsMay 21, 2026โ
Tests passing
Multi-Agent Code Refactor
A Python library that utilizes multiple AI agents to collaboratively review and refactor code. Each agent is specialized in a specific aspect, such as optimization, readability, or security. The agents work together to propose and implement comprehensive improvements to the input code.
What It Does
- Supports multiple agents with different roles (e.g., style, performance, security).
- Accepts Python code as a string or from a file.
- Outputs refactored code and a detailed change report.
- Optionally saves the refactored code to a file.
Installation
Install the required dependencies:
pip install pytestUsage
Run the tool from the command line:
python multi_agent_code_refactor.py "path_or_code" --agents style performance security --output_file output.pypath_or_code: Path to the Python script or Python code as a string.--agents: List of agent roles (default:style,performance,security).--output_file: (Optional) Path to save the refactored code.
Source Code
import argparse
import os
from typing import List, Union
class CodeRefactorAgent:
def __init__(self, role: str):
self.role = role
def review_code(self, code: str) -> str:
"""
Mock review and refactor method for the given code based on the agent's role.
"""
# Simulate a response for testing purposes
return f"# Refactored by {self.role} agent\n{code}"
def refactor_code(
input_code: Union[str, os.PathLike], agents: List[str], output_file: Union[str, os.PathLike, None] = None
) -> str:
"""
Refactor Python code using multiple AI agents with different expertise.
Args:
input_code (str or os.PathLike): Python code as a string or a file path to a Python script.
agents (List[str]): List of agent roles (e.g., ['style', 'performance', 'security']).
output_file (str or os.PathLike, optional): File path to save the refactored code. Defaults to None.
Returns:
str: Refactored Python code.
"""
# Load the input code
if os.path.isfile(input_code):
with open(input_code, "r") as f:
code = f.read()
else:
code = input_code
# Initialize agents
agent_objects = [CodeRefactorAgent(role) for role in agents]
# Each agent reviews and refactors the code
reports = []
for agent in agent_objects:
response = agent.review_code(code)
reports.append(f"Agent ({agent.role}):\n{response}\n")
# Assuming the response contains the refactored code
code = response
# Save the refactored code if output_file is provided
if output_file:
with open(output_file, "w") as f:
f.write(code)
# Return the refactored code and the reports
return code, "\n".join(reports)
def main():
parser = argparse.ArgumentParser(description="Multi-Agent Code Refactor Tool")
parser.add_argument("input_code", help="Path to the Python script or Python code as a string.")
parser.add_argument(
"--agents",
nargs="+",
default=["style", "performance", "security"],
help="List of agent roles (e.g., style, performance, security).",
)
parser.add_argument(
"--output_file", help="Path to save the refactored code.", default=None
)
args = parser.parse_args()
try:
refactored_code, report = refactor_code(args.input_code, args.agents, args.output_file)
print("Refactored Code:")
print(refactored_code)
print("\nChange Report:")
print(report)
except Exception as e:
print(f"Error: {e}")
if __name__ == "__main__":
main()Community
Downloads
ยทยทยท
Rate this tool
No ratings yet โ be the first!
Details
- Tool Name
- multi_agent_code_refactor
- Category
- AI-Powered Code Review Tools
- Generated
- May 21, 2026
- Tests
- Passing โ
- Fix Loops
- 2
Quick Install
Clone just this tool:
git clone --depth 1 --filter=blob:none --sparse \ https://github.com/ptulin/autoaiforge.git cd autoaiforge git sparse-checkout set generated_tools/2026-05-21/multi_agent_code_refactor cd generated_tools/2026-05-21/multi_agent_code_refactor pip install -r requirements.txt 2>/dev/null || true python multi_agent_code_refactor.py