Vegas is heating up (literally), and day 1 at #BlackHat is officially in the books! The house was packed during our morning 'off the record' breakfast, with bold conversations, controversies (the good kind), and the no-nonsense security talk you only get behind closed doors. In case you missed it, don't worry, we're just getting started 😎 Stop by our booth (#1361) to chat with our experts, catch live demos, and learn why leading enterprises trust Astrix to keep their AI-driven workflows breach-free. Did we mention we have Transformers LEGOs?
Astrix Security
Computer and Network Security
Securing Agentic & Non-Human Identities
About us
Modern businesses run on millions of machine‑to‑machine connections: Autonomous AI Agents orchestrating work, CI/CD service accounts, SaaS OAuth apps, and API keys that keep data flowing. These Agentic and Non-Human identities outnumber humans ~45 to one, inherit broad, static privileges, sit outside IAM review cycles, and often persist long after their creators leave. Attackers exploit the gap: high-profile attacks against the US Treasury, Okta, and Microsoft (Midnight Blizzard) demonstrate how a single leaked token or hijacked agent can grant root‑level access across cloud and SaaS. While existing IAM solutions focus solely on human identities, Astrix provides security teams with continuous discovery of AI agents and NHIs, risk‑based prioritization, real‑time threat detection, and automated remediation. Fortune 1,000 enterprises such as Workday, NetApp, Priceline, Hubspot, Figma and Agoda already rely on Astrix to securely unleash AI-driven automation.
- Website
-
https://www.astrix.security/
External link for Astrix Security
- Industry
- Computer and Network Security
- Company size
- 51-200 employees
- Headquarters
- New York
- Type
- Privately Held
- Founded
- 2021
Products
Locations
-
Primary
Get directions
New York, US
-
Get directions
Manhattan Ave
New York, NY, US
Employees at Astrix Security
-
Dimitri Sirota
BigID - Know Your Data | Control Your Data
-
Rodney Toy
Strategic Advisor; GTM /Sales / Revenue Operations Executive; Executive Board Member; Contemporary Artist
-
Jonathan Sander
Philosopher Turned Technologist
-
Shira Kaplan
Cyber-Security Entrepreneur & Investor | Board Advisor | World Economic Forum Young Global Leader (2017) | Bilanz 100 Digital Shapers "Hall of Fame"
Updates
-
Abused OAuth app ➡️ voice phishing ➡️ one-click access. This is what happened in a recent attack that fooled even the great Google. No fancy exploit. Just clever use of OAuth and a bit of voice phishing. Astrix researcher Roy Aviram breaks down this slick phishing campaign and what it means for OAuth security moving forward.
A recent phishing campaign used voice phishing to trick users into installing a malicious Salesforce OAuth app. The attackers didn’t need passwords or MFA, just a single click of a button. The crazy part? Anyone can fall for it, even Google. In this attack, the attackers impersonated IT support via voice calls and convinced employees to visit Salesforce’s connected‑app setup page and approve a malicious app. The app was often branded to look like Salesforce’s Data Loader, a well-known and trusted tool. Once approved, it gave attackers full access to the Salesforce environment, which they used to exfiltrate data and eventually pivot into services like Okta and Microsoft 365. In their latest update, Google confirmed that one of their Salesforce instances was also affected by the campaign. According to the company, the data accessed by the attackers was limited to basic and mostly publicly available business information, such as business names and contact details. At Astrix Security, we’ve seen firsthand how widespread and risky OAuth abuse has become, and this incident is just one example. With one consent screen, attackers can gain persistent access, bypassing both MFA and traditional phishing protections. While these attacks are becoming more common, most organizations still lack visibility into which apps have access to their environments. If Google can be targeted this way, anyone can. It’s time for security teams to review which apps have access, understand what scopes are granted, and monitor for suspicious behavior. OAuth applications are part of the organization's security perimeter and should be treated as such. Read the full report by Google here: https://lnkd.in/epdZaw6v
-
-
We're back with part 2 of our MCP-focused blog series! In this chapter, Jonathan Sander flips the script, revealing how MCP's biggest weakness (hello, abstraction layer) can actually become your strongest security asset. Curious how MCP can level up your AI security game? Read the full article here: https://lnkd.in/dWTwPdiP 👀 If you want to be part of the change and help your peers see into the future of MCP, please take our short survey: https://lnkd.in/d4d6ngkq
-
-
Losing sleep over managing your AI agents? We've got your back 🤖 We help industry-leading brands to tackle this challenge head-on, scaling AI governance effortlessly, securely, and at enterprise speed. Our AI security platform simplifies complex AI agent governance, giving your security teams peace of mind. Interested to learn how our customers are doing it? Read more here: https://lnkd.in/dCmH-bkx Also, if you're in #BlackHat, we're doing some exciting stuff there. Come say hi at booth #1361 🖖🏼
-
-
How does Astrix solve the Shadow AI problem? 🥷🏼 The instant you let a humble AI assistant fetch data or trigger actions on your behalf, it evolves into a full-fledged AI agent, spinning up a web of non-human identities that rarely show up in your standard IAM dashboards. When your organization connects to AI platforms like ChatGPT, Copilot, and Claude, fingerprinting capabilities become essential to pinpoint which AI tools are accessing your systems and enforce strict governance with on-demand detection playbooks. This is where we come in: Astrix streamlines practical detection methods, introducing simple but effective ways to bring AI agents into governance without breaking anything, without slowing down innovation. Heading to #BlackHat? Stop by our booth (#1361) to learn more.
-
-
We're bringing the 🔥 to Vegas! Astrix is heading to Black Hat USA 2025 to help security teams rethink how they manage the fastest-growing threat surface in the enterprise: 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀, 𝗮𝗻𝗱 𝘁𝗵𝗲 𝗡𝗛𝗜𝘀 𝘁𝗵𝗮𝘁 𝗽𝘂𝗹𝗹 𝘁𝗵𝗲𝗶𝗿 𝘀𝘁𝗿𝗶𝗻𝗴𝘀. Swing by our booth to discover how we: ✔️ Automatically discover every AI agent and NHI in your environment ✔️ Lock down risky permissions and secrets ✔️ Detect and remediate real-time threats before they become breaches Get a personalized demo, ask our experts anything, and walk away with some exclusive Astrix swag 👉🏼 https://lnkd.in/dpQWccuj 👀 And if you’re up for something a little more "off the record", we’re hosting an exclusive invite-only breakfast with top security leaders. No slides, no filters, just real conversation and insights over great coffee. Secure your seat before it fills up 👉🏼 https://lnkd.in/d8WXJKPt
-
-
Let's talk MCP. We're kicking off a fresh, no-nonsense, three-part series diving into MCP (Model Context Protocol) and how it shapes AI security today. In Part 1, our field CTO Jonathan Sander tackles the elephant in the room: why MCP itself might be your biggest security headache (for now). But don’t worry, there's good news ahead! 👉 Part 2: MCP becomes the hero. 👉 Part 3: MCP shapes the future. Grab your coffee, settle in, and let's unravel MCP together. Jonathan’s insights are just a click away: https://lnkd.in/dR-gdhW5
-
-
🚀 We're AMP'd to announce that Astrix is an official launch partner for the Torq AMP program! This collaboration is all about speed and efficiency - faster detection and response while taking security automation to the next level. By integrating Astrix Security’s AI and NHI protection with Torq’s powerful automation capabilities, we're streamlining security workflows and redefining what rapid, effective threat response looks like.
-
-
#BlackHat is just around the corner, and we’re bringing our A(I) game! This year, Astrix is hitting Vegas to help you conquer the chaos of AI agents and the non-human identities (NHI) that pull the strings 🤖 Come chat with our experts, catch live demos, new showcases, and find out why leading enterprises trust Astrix to keep their AI-driven workflows breach-free. Ready to master your AI and NHI security playbook? More details in the first comment ⤵️
-
-
From buried risk to instant insights💡 Want to hear how Xerox moved from weeks of manual effort to a game-changing NHI visibility? In our latest Ask-Me-Anything session, we sat down with Mike Hildebrandt to talk about how his team took control of their Non-Human Identities. If you want to hear what real-world NHI remediation actually looks like, and how to seamlessly implement it within your organization, this one is for you. Link in the first comment below 👇🏼