The market has answers. They don't work for autonomous agents.
The Illusion of Securing Agents
AI Gateways add latency,
not protection.
Autonomous agents operate across MCP calls, local file reads, and tool calls. Gateways cannot see any of that, and add latency for the traffic they do intercept.
Sandboxes break the moment agents evolve.
Every new tool or integration breaks the sandbox configuration and forces a full reconfiguration. Sandboxes are not supported for most agents in your environment.
LLM Guardrails are suggestions, not enforcement.
A recent study tested every major defense for coding agents: over 80% bypass rate. Guardrails watch prompts and outputs, not tool calls, credential access, or lateral movemen
Identity Security, PAM, and NHI were not built for agent runtime.
They are static permission frameworks: they define what agents are allowed to do, not what they are actually doing. Rebranding a permissions model as agent security does not give you runtime visibility.
EDR cannot see the agentic layer.
It watches processes and files, not what multiple agents and sessions are doing inside them. Tool calls, credential access, and agent-to-agent attacks are all invisible to it.
AI Governance platforms are auditors, not defenders.
Overnance tools rely on static integrations and periodic scans. By the time they surface a risk, the damage is already done.
The Tenet Way
AI agents are the new layer between humans and systems. Autonomous, privileged, and already attacking each other in production. Security teams have no defense built for how agents run. Until now.
01
No proxies.
No gateways. No latency
Tenet deploys where agents run, not in front of them. A lightweight sensor sits directly inside the agent runtime, across endpoint and cloud. No code changes.
Deployed in hours.
Deployed in hours.
02
Agent-Side Simulation (patent pending)
Security teams are drowning in false positives and unexplained alerts. Before any action executes, Tenet’s models simulate where it leads, block the path if it’s dangerous, and produce a full trace that proves why. A receipt for every action.
03
Threat intelligence built on production reality
Trained on millions of live agentic sessions across 1,000+ unique agent profiles. 100+ detectors, updated continuously from Tenet’s Threat Lab
and global Honeypot Network.
04
Sees the full agentic layer
True agentic sessions reconstructed across OS, network, LLM calls, tool calls, and intent in real time. The complete picture of what every agent is actually doing, including the sessions every other tool
treats as a black box.
treats as a black box.
05
Built by the team that confirmed the threat in production
Barak and Nevo led Cisco’s AI Defense research team before founding Tenet. Barak is an Israel Defense Prize recipient. Both have spoken at industry conferences including DEFCON. They did not pivot into AI security. They built it from scratch.
Battle-Tested:
From 0 to 5,000,000+ Agentic Sessions
$1Bn ARR enterprise with nearly 4,000 employees,
started with 2 autonomous agents in production. With Tenet, they expanded to 20+ agents serving over 24 million users. Tenet blocked real attacks along the way, including agents manipulated to leak internal data, before any customer was affected.
started with 2 autonomous agents in production. With Tenet, they expanded to 20+ agents serving over 24 million users. Tenet blocked real attacks along the way, including agents manipulated to leak internal data, before any customer was affected.
“Tenet gave us the confidence to go from 2 AI agents to 20 in six months – serving 24 million users. They stopped real attacks along the way. Without that, we’d still be stuck at 2.”
Head of Information Security @ $1Bn+ U.S. Legal Institution
Agents
2- 0
Users protected
0 M+
Attacks blocked
0 +
Agentic sessions
~ 0 M
Sessions analyzed
0 M+
Featured by the industry
Podcasts & Webinars
2 min read
Barak Sternberg Joins CSA's AI Security Innovation Panel: When Agents Find Zero-Days Nobody Asked For
"Give an agent a CTF and it'll find a zero-day path to escape the container. They manipulate their way around in ways nobody scoped for." Barak Sternberg at the CSA Agentic AI Panel.
Podcasts & Webinars
2 min read
Barak Sternberg Joins the Execweb Podcast to Talk AI Agent Security
"Agents are doing exactly what they're allowed to do.. just in a way they shouldn't." That line captures the core challenge of securing AI agents at runtime, and it's where the conversation with Barak Sternberg begins.
In the Press
External
June 17, 2026
3 min read
The attack that hijacked Claude Code came through Sentry. Datadog, PagerDuty, and Jira have the same exposure.
A single fake error report hijacked Claude Code in controlled testing — the agent ran the attacker's code with the developer's full privileges, and not one alert fired. EDR, WAF, IAM, and the firewall all missed it completely.
External
June 17, 2026
3 min read
Tenet Security, a New York City-based cybersecurity startup that builds tools to secure autonomous AI agents and agentic workflows against data exfiltration and misuse, raised $6 million in seed funding. The Westly Group and MizMaa Ventures led the round.
Tenet Security, a New York City-based cybersecurity startup that builds tools to secure autonomous AI agents and agentic workflows against data exfiltration and misuse, raised $6 million in seed funding. The Westly Group and MizMaa Ventures led the round.
External
June 17, 2026
3 min read
Ex-Cisco researchers launch Tenet Security to lock down rogue AI agents
Former Cisco artificial intelligence security researchers have launched a new company to tackle a problem that barely existed a year ago: securing the autonomous AI agents enterprises are handing the keys to their most critical systems.
External
June 17, 2026
3 min read
Ex-Cisco researchers launch Tenet Security to lock down rogue AI agents
Former Cisco artificial intelligence security researchers have launched a new company to tackle a problem that barely existed a year ago: securing the autonomous AI agents enterprises are handing the keys to their most critical systems.
External
June 12, 2026
3 min read
Agentjacking: a fake bug report can hijack your AI coding agent
Researchers turned a public Sentry credential into remote code execution on developer machines. The agents ran the attacker's code even when told to ignore it, and EDR, firewalls and prompts all missed it.
Recognized By The Industry
Announcments
2 min read
Tenet Security Named a Hot Company in GenAI Agentic Application Security by Cyber Defense Magazine
Tenet Security has been named a Hot Company in GenAI Agentic Application Security at the InfoSec Awards by Cyber Defense Magazine, recognizing our platform's approach to securing autonomous AI agents at runtime — a gap traditional security tools are not built to address.
Announcments
2 min read
Tenet Security Wins Two Gold Awards at the 2026 Cybersecurity Excellence Awards
Tenet Security wins two Gold Awards at the Cybersecurity Excellence Awards in the Most Innovative Cybersecurity Company and Agentic AI Security categories, recognizing our work building the industry's first Runtime Defense Platform for autonomous AI agents.