Mrs. Hema Prabha G, Madhusree M, Neethish S, Sanjanaa S
Abstract
This project focuses on developing an intelligent and autonomous system capable of accurately reviewing source code and identifying bugs, code smells, security vulnerabilities, and performance inefficiencies. By leveraging advanced technologies such as Agentic AI, Large Language Models (LLMs), and rule-based analysis frameworks, the system ensures high accuracy and consistent performance in detecting code-level issues. The AgentCode Inspector allows developers to submit code through a web interface, and the AI agent autonomously performs multi-step code analysis and delivers human-like review comments and optimized code fixes. Key features include syntax checking, complexity analysis, security scanning, and intelligent suggestion generation for seamless software quality improvement. The system is supported by a robust agent framework that enables continuous reasoning and iterative evaluation, ensuring reliable and efficient performance in automating the software code review process.
The AgentCode Inspector effectively provides a reliable and autonomous solution for reviewing source code and delivering intelligent, actionable feedback. By leveraging Agentic AI frameworks such as LangChain, CrewAI, and AutoGen along with advanced LLM-based reasoning, the system achieves high consistency and coverage in identifying bugs, security vulnerabilities, code smells, and performance issues across diverse codebases. Overall, this project contributes to the field of AI-assisted software engineering by offering an accessible, efficient, and scalable platform that reduces manual review effort and improves code quality. It also lays the foundation for future improvements, such as supporting additional programming languages, integrating with CI/CD pipelines, and extending agent capabilities for real-time collaborative code review.
References
I. Sharma & D. Rattan, “Code Quality Generated by AI Tools: A Review,” IOSR Journal of Computer Engineering (IOSR-JCE), vol. 27, no. 3, pp. 55–68, 2025.
G. Viswanathan, “AI Agentic Scriptless Automation in Software Testing,” International Journal of Computer Trends and Technology (IJCTT), vol. 72, no. 9, pp. 120–125, 2024.
S. K. Jangam & N. Karri, “AI Tools for Automating Code Reviews, Providing Contextual Feedback, and Improving the Efficiency of the Review Process,” Advances in Interdisciplinary Journal of Computer Science and Technology (AIJCST), vol. 1, no. 1, pp. 36–42, 2025.
Official LangGraph Documentation — https://langchain-ai.github.io/langgraph/
Official AutoGen Documentation — https://microsoft.github.io/autogen/
FastAPI Documentation — https://fastapi.tiangolo.com/
How to Cite This Paper
Mrs. Hema Prabha G, Madhusree M, Neethish S, Sanjanaa S (2026). AGENTCODE INSPECTOR: AN AGENTIC AI-BASED AUTONOMOUS CODE REVIEW SYSTEM. International Journal of Computer Techniques, 13(2). ISSN: 2394-2231.
