YesWeHack’s Penetration Testing utilizes on-demand autonomous AI agents

YesWeHack Unveils Pentest Agentique: AI-Driven Security Testing at Your Fingertips

On June 25, 2026, French cybersecurity platform YesWeHack announced the launch of its Pentest Agentique, a service leveraging autonomous artificial intelligence agents that can be activated on-demand to identify vulnerabilities within an organization and deliver results on the same day. This innovative offering targets web and mobile applications, APIs, and other internet-exposed assets, utilizing black box, gray box, and white box testing methodologies. Results are integrated within the platform alongside the company’s other security services.

This development signifies a pivotal shift in the offensive security landscape. Traditionally, penetration testing has been performed as a one-time consulting service by human experts. However, with the rise of automated solutions, it is now evolving into a continuous function that can adapt quickly to emerging threats. This change is crucial as attackers increasingly leverage AI, shortening the window of opportunity between vulnerability disclosure and exploitation. Offensive security is now following the same trajectory as defensive measures, moving toward systems that operate at machine speed.

Testing Scope and Limitations

Pentest Agentique supports black box, gray box, and white box testing for web and mobile applications, APIs, and other online assets. It addresses high-impact vulnerabilities highlighted in the OWASP Top 10 as well as a range of potential attack vectors. YesWeHack emphasizes that its agents utilize the most advanced models available for offensive testing, including open-weight models. This flexibility allows organizations to employ models developed or hosted within the European Union or Asia-Pacific, depending on their specific needs.

The agents operate within a predefined set of rules established by YesWeHack to safeguard the confidentiality, integrity, and availability of the tested systems. This framework not only defines the boundaries of the offering but also highlights a parallel service: YesWeHack’s Continuous Pentests, designed to identify complex vulnerabilities that may elude fully automated approaches. While agent-based screening prioritizes speed and volume, nuanced business logic and intricate exploitation scenarios remain under human expertise.

An Integral Component of a Comprehensive Security Strategy

Pentest Agentique is fully integrated into the YesWeHack platform, where its findings can be managed alongside other security services, including bug bounty programs, Continuous Pentests, vulnerability disclosure policies, and checkpoints that identify actively exploited CVEs. Security teams can also access YesWeHack’s triage team, available around the clock to validate, reproduce, and enhance reports, thus ensuring a reduction in false positives by reintroducing a human element within the automated process.

Guillaume Vassault-Houlière, CEO and co-founder of YesWeHack, frames this offering as part of a broader strategy. “Pentest Agentique is quicker and simpler to deploy than traditional human-led pentests, offering wider coverage, scalability, and reduced costs,” he stated. He further emphasized the importance of maintaining a diversified offensive security strategy, highlighting that bug bounty programs and community expertise remain essential components of a proactive approach. In this context, automation is seen as a complement to human security services rather than a replacement.

Implications for CISOs in the Age of Continuous Screening

The introduction of an automated and persistent penetration testing service necessitates a reevaluation of priorities for Chief Information Security Officers (CISOs). One primary consideration is how vulnerabilities are prioritized: the agent determines their actual exploitability, requiring a level of trust in machinery to make risk assessments. This delegation introduces a broadened scope, as CISOs often take personal responsibility for decisions impacted by these assessments. Additionally, continuous screening from a vendor cannot substitute for independent audits, a requirement that remains in line with various compliance frameworks.

The economic model is shifting as well, transitioning from a one-off, billable engagement toward a permanent security function. For teams overwhelmed by an expanding attack surface, the advantages lie in the speed and frequency of testing. However, this also engenders increased dependence on a platform and its agents. YesWeHack’s flexibility in model selection, including those with open weights, mitigates some of this concern without eliminating it entirely. The broader question raised by this offering goes beyond the product itself: as defense mechanisms automate vulnerability detection at the same pace attackers are enhancing their exploitation techniques, the competitive edge will hinge on the swift remediation of threats, where human oversight remains crucial.

Scroll to Top