Unite.Ai Cato Networks Launches AgentSense to Combat AI Agent Manipulation
Article Content
- •Cato Networks launched AgentSense to prevent AI agent manipulation.
- •The solution targets indirect prompt injection risks in AI workflows.
- •Cato offers $30,000 for researchers who can successfully exploit AgentSense.
Cato Networks has introduced AgentSense, a security solution aimed at preventing harmful actions by AI agents before they execute. The technology addresses indirect prompt injection, where untrusted content can manipulate an agent's behavior despite it having valid credentials. To validate AgentSense's effectiveness, Cato is offering a $30,000 challenge to researchers who can successfully redirect an agent's workflow. The challenge invites participants to embed malicious instructions in content accessed by the agent, with the aim of diverting it from its intended task. Cato emphasizes that traditional access controls are insufficient, as they do not monitor how agents use their permissions during execution. The company aims for AgentSense to enhance the security of AI agents, allowing organizations to trust them with critical business processes. The technology is integrated into Cato's platform at no additional cost for existing users.
Ask AI about this cluster
Answers cite the sources they use
Timeline
More articles in this cluster (6)
Following this threat?
Track Cato Networks in your own feed — you're alerted when they show up in new reporting, leak sites or exploitation.
Free account · no card needed
Common questions
What is AgentSense?
How can researchers participate in the challenge?
What risks does AgentSense address?
Continue Reading
Lazarus Group Targets 30 South Korean Servers Using PHP Vulnerabilities North Korea-linked Lazarus Group has targeted 30 servers in South Korea as part of a larger cyber operation involving 581 servers across 10 countries. The attacks exploit two critical remote code execution vulnerabilities in PHP (CVE-2026-109618 and CVE-2026-109619), allowing unauthorized server command execution. The…