A 23-year-old Nigerian tech developer based in the United States, Tony Kabilan Okeke, has led a team of innovators to develop “Xploit,” an automated cybersecurity testing platform designed to identify vulnerabilities in Artificial Intelligence (AI) agents before they are exploited by hackers.
The project, which has already gained recognition in the United States startup ecosystem, was developed by a five-man team comprising alumni and students of Drexel University. The team includes Tony Okeke, Kamdi Okeke, Kiitan Fawole, Dalu Okonkwo and Michael Moemeke.
The innovation was first conceived during the “Start-Up In a Weekend” Hackathon organised by The Foundry and Velric in Philadelphia between November 21 and 23, 2025.
Speaking on the breakthrough, Tony Okeke explained that the increasing deployment of AI agents across businesses globally has created major security concerns requiring urgent solutions.
“As more businesses deploy AI agents that can take actions and use tools on behalf of customers, these systems become potential security risks,” he said.
“Unlike simple AI assistants, agents have access to tools and can perform real actions — meaning a security vulnerability isn’t just a PR problem, it could have serious real-world consequences.”
According to him, the team developed Xploit to function like an automated digital attacker capable of testing AI systems for weaknesses before cybercriminals exploit them.
Tony said he designed the architecture of the system and created the initial user interface prototype, which visually displays how the automated attacker thinks, strategises and launches attacks against target AI systems in real time.
The project, built under intense hackathon conditions, involved strategic collaboration among team members, with some creating sample AI agents for testing while others handled infrastructure, interface development and backend integration.
“The attacking system works like a strategic game player,” Tony explained.
“It chooses an attack strategy, creates a detailed plan, executes it step-by-step, analyses responses from the target AI system, and decides whether to continue or switch approaches.”
The team completed the project before the November 23 deadline and later pitched the innovation before judges, highlighting the growing global demand for AI security tools, especially among small and medium-scale businesses unable to afford dedicated cybersecurity teams.
The developers argued that Xploit addresses a rapidly expanding market for continuous automated red-teaming — an AI security testing model projected to grow from about $495 million in 2024 to nearly $4.9 billion by 2032.
Their presentation reportedly impressed judges so much that organisers demanded access to the team’s codebase and development history to verify the product was genuinely built during the hackathon weekend.
The team eventually won the “New Project Track” category and received a $1,500 cash prize.
Reflecting on the achievement, team member Kamdi Okeke said the feat was even more remarkable because the product was developed in less than a full day of actual coding time.
“What made the achievement particularly remarkable wasn’t just that we built it over a weekend,” Kamdi said.
“It was that, competing among over 100 of Philadelphia’s most driven creators, we transformed an abstract idea into a polished working prototype in less than a day through focused collaboration and strategic planning.”
The team later participated in another major innovation programme — the Venture Building Weekend hosted by United Effects Ventures between March 12 and 14, 2026.
At the event, Xploit reportedly emerged as the top project among 15 competing teams after undergoing rigorous mentorship and market validation sessions with investors and startup operators.
Tony described the experience as a major turning point for the startup.
“The mentorship and feedback we received from industry operators helped sharpen how we think about the problem and where our approach fits in the market,” he said.
Industry mentors at the programme reportedly validated both the team’s identification of the growing security risks posed by autonomous AI systems and their continuous red-teaming approach as a viable long-term cybersecurity solution for AI-powered products.
