Cato Networks Launches $30,000 AI Agent Hacking Challenge to Test New Defense
Cato Networks launched a $30,000 challenge for the first three hackers to compromise its new AgentSense AI security solution, highlighting the booming bug bounty market.

Israeli cybersecurity firm Cato Networks launched a public challenge offering $10,000 each to the first three participants who successfully hijack an autonomous AI agent and force it to deviate from its assigned task. Participants are invited to attack an AI workflow protected by Cato's newly introduced AgentSense security solution, which continuously monitors autonomous AI operations in real time to ensure decisions and outputs align with user intent.
Unlike conventional security tools that rely heavily on identifying known attack patterns, AgentSense analyzes how untrusted content influences agent decisions and actions in real time. This approach aims to defend against emerging vulnerabilities, including zero-day exploits for which no patches yet exist. Participants attempt to inject malicious instructions into the content processed by the agent to derail its operations, while the company utilizes the findings to test and refine its defense mechanisms.
The Industry Context of Bug Bounties
Offering financial rewards for identifying security flaws through public challenges is a well-established practice in the cybersecurity sector, commonly known as Red Teaming or Bug Bounty programs. For instance, AI security firm Gray Swan launched a competition in October 2026 offering $300,000 in prizes funded by major technology corporations including OpenAI, Anthropic, Meta, and Google. Major technology firms also run continuous bug bounty programs, with Microsoft Copilot offering up to $30,000 for critical flaws and OpenAI offering baseline rewards of $30,000 for core vulnerabilities.
Data from the US-Dutch security platform HackerOne indicates that organizations paid a record $89.1 million directly to independent researchers within a single year for vulnerability discoveries. Microsoft distributed approximately $20 million to 562 researchers over the past year, while Google awarded around $17.1 million. Market research from Research and Markets estimates the global AI red teaming services market at approximately $2.26 billion, projected to reach $6.17 billion by 2030 and approach $20 billion by 2035 at an annual growth rate of roughly 30 percent.
Leadership Perspective
Shlomo Kramer, co-founder and CEO of Cato Networks, has frequently warned about the risks associated with the rapid proliferation of artificial intelligence while actively investing in cybersecurity solutions tailored for AI environments. Explaining the rationale behind the new security offering, Kramer emphasized the critical need for governance in modern digital enterprises.
"In an era where organizations entrust AI agents with decisions impacting revenues, customer relations, and business continuity, oversight of agent activity is no less essential than financial control," Kramer stated. "Granting greater autonomy to AI agents requires security mechanisms that prevent harmful actions. Our vision is for AgentSense to provide clients with the confidence to rely on AI agents in managing critical business processes without losing control."
Registration for the challenge is hosted on the company's official website.





