Tenet Security launched publicly last week with $6 million in seed funding to solve a problem most enterprises didn't know they had: autonomous AI agents are operating inside corporate systems with almost no real-time oversight. The round was led by The Westly Group and MizMaa Ventures. As organizations grant these agents access to code repositories, databases, and other critical applications, traditional security tools are failing to track what the agents actually do at runtime.
Agent-side simulation predicts actions before they execute
Tenet's platform centers on a patent-pending technology called Agent-side simulation. The system predicts an agent's likely next actions before they hit production environments. When it detects risky behavior, it can block execution and generate a full trace of the decision that led to the intervention. "EDR watches processes. IAM watches users. Guardrails watch prompts and outputs. None of them see what an agent actually does at runtime," the company said, describing what its founders call a fundamental blind spot in existing security stacks.
The startup was founded by Barak Sternberg and Nevo Poran, both former researchers in Cisco's AI Defense unit. Sternberg said that "AI agents may be the biggest productivity unlock enterprises have seen in decades… but they also create an entirely new security layer that requires a fundamentally different approach to protection."
The "Agentjacking" threat class
Tenet Threat Labs identified a new attack vector it calls Agentjacking. Malicious instructions hidden in emails, documents, or databases manipulate agents into executing attacker-controlled actions while staying within their authorized permissions. "Attackers can manipulate agents to access sensitive data, abuse privileges, or take actions on their behalf in ways traditional security tools were never designed to detect," said CTO Nevo Poran.
Early deployments already show the scope of the problem. One large enterprise grew from two to more than twenty AI agent deployments in six months. During that period, Tenet detected and blocked multiple attacks, including a critical XSS attempt. In another case, a runaway agent generated significant unnecessary token costs before the platform caught it. The field of AI for Product Development is accelerating, and security practices have not kept pace.
Why this matters for product development teams
Product teams are shipping AI agents into production faster than security teams can assess the risk. When a single agent can write code, query databases, and trigger downstream workflows, a compromised or misdirected agent becomes a direct threat to the product pipeline. The rise of AI Agents & Automation means that product developers need to treat agent behavior as a runtime security concern, not just a model training or prompt engineering problem. Tenet's approach signals that security tooling is shifting from static guardrails to real-time behavioral prediction - and product teams will be expected to integrate that layer into their deployment workflows.
Your membership also unlocks:
