Black space background with purple-pink planets and small white stars scattered around.

Onyx Blog

Insights on AI security, governance, and the evolving risk landscape

Onyx’s $113M Series B: Keeping Humans in Control as AI Becomes Smarter

Onyx CEO Maxim Bar Kogan shares why the company raised $113 million to build the control plane enterprises need as AI agents take on more consequential work.

Wake-Up Call: What the OpenAI-Hugging Face Incident Should Teach Every Security Team

An OpenAI offensive agent escaped its sandbox into Hugging Face – the latest signal that offensive AI capability is outpacing the defenses around it.

Knitted-yarn illustration of a checklist on a clipboard with a magnifying glass over a question mark, surrounded by user, code, database, and padlock icons, on a pink grid background.

01 Jul 2026

⸱

Cheng Ren

The 5 Questions I would ask any AI Security Vendor

Five questions that separate real AI agent control from detection, before you sign.

Knitted-yarn illustration of curly brackets framing a tangled cord, evoking a code snippet, on a green background.

17 Jun 2026

⸱

Patrick Burke

An MCP Server Walks Into a Bar. It Orders on Your Tab

The two MCP server risks enterprises face, and the controls that catch both early.

04 Jun 2026

⸱

Mark Loewenstein

What to Make of the AI Executive Order

What the June 2026 AI executive order covers, and what it means for your AI agents.

A "Stale" risk-assessment snapshot card over a row of connected-app icons, contrasting a point-in-time scan with a live environment.

03 Jun 2026

⸱

Evin Safdia

Snapshot vs Video: Why Point-in-Time AI Risk Reports Miss Real Threats

Why quarterly AI risk reports go stale, and what continuous posture replaces them.

Knitted-yarn illustration of an AI chat bubble with a robot icon, surrounded by a tangle of yarn threads, a padlock, a warning triangle, a gear, and a document.

20 May 2026

⸱

Stephen Fitzgerald

Tales from the Runtime Layer: Why Authorized Agents Break Your Security and Your Infrastructure

Two runtime cases where authorized AI agents broke security using legitimate access.

Knitted-yarn illustration of an archway strung with four horizontal rails, each holding a colored padlock, set against a soft blue sky with woolly clouds.

06 May 2026

⸱

Christopher DiClerico

Four Guardrails: The Customer Success Pattern for Safe AI Agent Adoption

Four controls to govern AI agents already running, each deployable in a week.

Knitted-yarn illustration of a stack of file folders wrapped in tangled colored cords, tagged with a round "24h" label, on a green starburst background.

15 Apr 2026

⸱

Christopher DiClerico

Thousands of Agents: What Onyx Discovers in 24 Hours

Why the first agent inventory scan finds far more than enterprises expect to see.

24 Mar 2026

⸱

Yuval Neuhaus

Onyx Named a Representative Vendor in Gartner’s 2026 Market Guide for Guardian Agents

Onyx is a Representative Vendor in Gartner's 2026 Market Guide for Guardian Agents.

12 Mar 2026

⸱

Maxim Bar Kogan

Introducing Onyx Security: The Secure AI Control Plane for Enterprises

A single platform to oversee autonomous AI agents that’s already trusted by Fortune 500 companies