TLDR: Cycode has launched its AI Exploitability Agent, a new tool designed to rapidly prioritize and remediate millions of new code vulnerabilities, particularly those arising from the accelerated pace of AI-generated code. Debuting at Black Hat 2025, this agent aims to help security teams identify and fix exploitable risks 99% faster by providing clear visibility, intelligent prioritization, and automated fixes.
SAN FRANCISCO – July 22, 2025 – Cycode, a leader in AI-native application security, today announced the release of its groundbreaking AI Exploitability Agent. This innovative solution, set to debut at Black Hat 2025, is engineered to empower security teams by prioritizing high-risk, exploitable vulnerabilities and accelerating their remediation by an impressive 99%.
The launch comes at a critical time when artificial intelligence is dramatically speeding up software development. Tools like Cursor are reportedly generating over a billion lines of code daily, inadvertently introducing millions of new security vulnerabilities. Conservative estimates suggest at least one security flaw per 10,000 lines of code, translating to approximately 100,000 new security flaws emerging every day from AI-generated code alone. Furthermore, reports indicate that up to 40% of AI-generated applications contain vulnerabilities, with some tools consistently failing to produce secure code.
Lior Levy, CEO and Co-founder of Cycode, highlighted the escalating challenge facing security professionals. “We were already facing an overwhelming tide of security alerts,” Levy stated. “The AI coding revolution threatens to completely overwhelm traditional approaches. It’s no longer enough to just keep pace; security must take the lead, leveraging automation and AI that provides crystal-clear visibility, intelligent prioritization, and automated fixes.”
Cycode’s new Exploitability Agent is a pivotal component of its AI-native application security platform, fundamentally transforming the often arduous and time-consuming process of triaging alerts and accurately determining true exploitability. The agent is designed to answer three critical questions for security teams:
1. Is a violation exploitable? The agent automates exploitability analysis to determine if attackers can successfully target vulnerabilities in real-world scenarios, distinguishing between theoretical flaws and actual risks.
2. What is the risk? Moving beyond mere severity, Cycode quantifies the relative risk of violations by integrating code-to-runtime context into its risk score calculations. A high-severity vulnerability that is not deployed or exposed, for instance, poses less risk than a medium-severity one with a known exploit in a public-facing application.
3. What is the root cause? While the primary focus of the Exploitability Agent is on the first two points, it works in conjunction with Cycode’s broader suite of AI Security Teammates, which includes a Fix & Remediation Agent capable of analyzing root causes and proposing context-aware code fixes.
Also Read:
- Harness Revolutionizes DevOps with Expanded AI-Powered Pipeline Generation
- UK AI Code Assistant Software Market Set for Explosive Growth, Projected to Exceed $3 Billion by 2035
This new agent expands Cycode’s suite of AI Security Teammates, automatically distilling overwhelming alerts into clear, risk-based priorities. Cycode asserts that while AI is transforming how software is created, their platform is transforming how it is secured, offering a proactive approach to the burgeoning challenges of AI-generated code vulnerabilities.
