SMART INDIA HACKATHON 2025
URLGuard
• Problem Statement ID – 25229
• Problem Statement Title- Identification of URL based attacks from IP data
• Theme- Blockchain & Cybersecurity
• PS Category- Software
• Team ID-
• Team Name (Registered on portal): BeeHive
URLGuard
Project Idea: URLGuard A full-stack system that ingests HTTP traffic (PCAP or live), extracts URL features, and detects 11 types of URL-based
attacks in real time. It classifies and scores threats, validates them in a sandbox, and delivers results via a dashboard and REST API with export,
filtering, and security controls.
Proposed Solution Address to the problems Innovation & Uniqueness
@SIH Idea submission- Template 2
TECHNICAL APPROACH
Methodology Technologies used:
● Programming Language: Pythons it is widely supported for
cybersecurity, automation, and data analytics.
● Web Frameworks: Flask or Django for web GUI/dashboard and
REST API development. Django adds built-in security for common
vulnerabilities.
● Machine Learning/Data Analysis: Pandas, Scikit-learn,
TensorFlow for dataset handling, detection models, and anomaly
analysis.
● Network/Packet Analysis: Scapy and PyShark to parse and analyze
PCAP files or network streams.
● Visualization: Matplotlib or Plotly for visual analytics in
dashboards.
● Security Tool Integration: Automate tools like SQLmap using
Python scripts and interact with APIs for others (Burp Suite,
XSStrike).
● Hardware: Standard modern PC or server (8-core CPU, 16-32GB
RAM, SSD storage) is usually sufficient for our lab and mid-scale
deployments.
@SIH Idea submission- Template 3
FEASIBILITY AND VIABILITY
@SIH Idea submission- Template 4
IMPACT AND BENEFITS
Impact Benefits
● Supports compliance requirements ● Enhances research and algorithm development
● Reduces manual security workload ● Facilitates hands-on security training
● Minimizes downtime ● Scales across environments
● Enables forensic investigations ● Enables trend analytics
● Improves network performance ● Allows easy data export and reporting
@SIH Idea submission- Template 5
RESEARCH AND REFERENCES
A survey of malicious URL detection techniques, datasets … (Y. Tian et al., 2025) — a comprehensive review of malicious URL detection,
covering rule-based to advanced ML & deep learning methods. ScienceDirect
Phishing Detection using Machine Learning based URL Analysis: A Survey (Arathi Krishna V et al., IJERT, 2021) — focuses on classical ML
techniques for phishing detection using URL features. IJERT
Phishing URL detection with neural networks: an empirical study (H. Ghalechyan et al., 2024) — explores deterministic & probabilistic neural
networks for URL classification. Nature
Malicious URL Detection with Explainable Machine Learning (2025) — introduces explainability into malicious URL detection models.
ACM Digital Library
Detecting Phishing URLs Based on a Deep Learning Approach (MDPI) — uses 1D CNN for robust phishing URL detection; evaluates on multiple
datasets. MDPI
Applications of deep learning for phishing detection: a systematic review — surveys ML/DL techniques applied to phishing detection.
PubMed Central
Modeling Hybrid Feature-Based Phishing Websites (SD Guptta et al., 2022) — uses URL + hyperlink features, hybrid ML methods.
PubMed Central
Machine Learning-Driven Detection of Cross-Site Scripting Attacks (Alhamyani & Alshammari, 2024) — specific work tackling XSS detection
using ML. MDPI
A comprehensive literature review on phishing URL identification (2024) — looks at ~41 studies from 2019-2024, especially deep learning
applied to phishing URL detection. Taylor & Francis Online
PhishMatch: A Layered Approach for Effective Detection of Phishing URLs (Harshal Tupsamudre et al.) — combines lightweight checking,
whitelist + ML for efficient client-side phishing detection.arXiv
@SIH Idea submission- Template 6
