Horizon3.ai’s cover photo
Horizon3.ai

Horizon3.ai

Computer and Network Security

San Francisco, CA 26,892 followers

AI-Native Proactive Security

About us

The NodeZero® platform empowers your organization to continuously find, fix, and verify your exploitable attack surface. Reduce your security risk by autonomously finding weaknesses in your network, knowing how to prioritize and fix them, and immediately verifying that your fixes work. NodeZero delivers production-safe autonomous pentests and other key assessment operations that scale across your largest internal, external, cloud, and hybrid cloud environments. No required agents, no code to write, and no consultants to hire. We are a fusion of former U.S. Special Operations cyber operators, startup engineers, and formerly frustrated cybersecurity practitioners. We're committed to helping solve our common security problems: ineffective security tools, false positives resulting in alert fatigue, blind spots, "checkbox” security culture, cybersecurity skills shortage, and the long lead time and expense of hiring outside consultants.

Website
https://www.horizon3.ai
Industry
Computer and Network Security
Company size
201-500 employees
Headquarters
San Francisco, CA
Type
Privately Held
Founded
2019
Specialties
cybersecurity, penetrationtesting, AI, purpleteams, infosec, machinelearning, datasecurity, autonomouspentesting, attack surface management, red teams, offensive security, pentesting, security validation, security verification, and exploitable vulnerabilities

Products

Locations

Employees at Horizon3.ai

Updates

  • NightDragon Founder & CEO Dave DeWalt stopped by our booth at RSAC last week to chat with Horizon3.ai Co-Founder Anthony Pillitiere — watch the video to learn more about our proactive security platform and why it's so critical in the age of AI. ⬇️

  • 🚨 There’s an actively exploited data leak in Citrix NetScaler ADC and NetScaler Gateway. We’ve released a Rapid Response test for it. CVE-2026-3055 (CVSS 9.3) is a SAML IdP memory overread vulnerability. An unauthenticated attacker can send crafted requests to read sensitive data directly from process memory. No auth required. No user interaction required. If the appliance is configured as a SAML Identity Provider, this can expose: • Session tokens • Credentials • Sensitive in-memory data This is a front-door system — NetScaler sits in front of critical apps, VDI, and enterprise access. Compromise here isn’t just data exposure — it can enable follow-on access into downstream systems. Reports indicate this vulnerability is under active exploitation. Affected Systems: • NetScaler ADC (Citrix ADC) • NetScaler Gateway (Citrix Gateway) • Configured as SAML IdP and running vulnerable versions If you’re running NetScaler: ⚠️ Check if SAML IdP is enabled ⚠️ Identify versions prior to the fixed release (14.1-60.58 and earlier) 👉 Run the #NodeZero Rapid Response test to quickly validate internal and external exploitability — and re-run after patching to confirm remediation: https://lnkd.in/gmJ7y5Hx #ProactiveSecurity #AISecurity #cybersecurity #infosec

    • No alternative text description for this image
  • Horizon3.ai reposted this

    My prediction for the rest of the year: AI-accelerated vulnerability research will generate so many new (mostly slop) vulnerabilities, CISO's will be overwhelmed trying to decide what NOT to fix, while at the same time getting crushed with a new CISA KEV hitting the news every day So what? If I were CIO again I'd build muscle memory around three things: 1. Prioritization - Is the issue exploitable? - What is the consequence to the business if the issue is exploited? - Are threat actors actually exploiting the issue? 2. Remediation - Can I quickly apply virtual patches, firewall rule changes, or other mitigations? - Can I prove I'm no longer exploitable? - How can I build more remediation expertise so I'm not single-threaded on a few experts (and automate as much as possible, like rolling credentials)? 3. CISA KEV Rapid Response - How can the IR and patching teams react faster to CISA KEV's, especially KEV's that effect edge gateways like Ivanti, Fortinet, Palo devices, etc ? - How do I ensure my key security controls (EDR, SIEM, etc) are tuned to effectively detect and respond to attacks? - What early warning systems can I put in place to accelerate incident response, like continuously probing for ways to break in from the outside, using decoys to increasing the effort of attack, etc? I also think that it will be difficult for app developers to quickly fix security flaws in custom applications, while at the same time it will be easier and faster for attackers to exploit custom apps. Therefore 3 things will happen: 1. A focus on mitigations like tuning webapp firewalls to prevent exploitation buying time for developers to fix issues will be the most important app sec effort 2. An assume Breach mindset will be incredibly important because there are far too many doors and windows for attackers to break into, so understanding and reducing the the blast radius from each likely initial access point will be paramount 3. Ultimately applications are one of the front doors along with Edge Devices and misconfigured services, but the bulk of the attack really starts with shell on a single host. From there the attacker will live off the land by harvesting credentials, laterally maneuvering, etc. Credentials specifically become the critical enabler to the attack, so locking down Active Directory, getting stricter on password policies, ensuring OS Credential Dumping is enabled on your EDR, etc become key ways to stifle attacks tl;dr: Vulnerability Risk Intelligence based on your exploitable attack surface, and how you respond to CISA KEV's, will be the two most important CISO initiatives in the coming months as AI accelerates vulnerability research #infosec #cybersecurity #ai #vulnerabilitymanagement Horizon3.ai

    • No alternative text description for this image
    • No alternative text description for this image
  • From San Francisco… to Times Square. 🗽 Nasdaq just celebrated Horizon3.ai as one of the 2025 Deloitte Technology Fast 500™ Bay Area winners, and the numbers that got us here are hard to ignore: 🥇 No. 1 fastest-growing company in the Bay Area 🥉 No. 3 fastest-growing company in North America 🚀 19,939% growth in just 3 years Appearing on this list is a reflection of a massive shift in cybersecurity, where organizations are done guessing their business risk. Instead, they’re using Horizon3.ai's Proactive Security Platform to continuously prove what’s exploitable, fix what matters, and stay ahead of attackers at machine speed. This is what the Fast 500 is all about: real momentum, real impact, and companies redefining their industries. 💪 #DeloitteFast500 #Cybersecurity #AI #ProactiveSecurity #NodeZero

    • No alternative text description for this image
  • #RSAC2026: ✅ This week made it clear: despite massive investment, most organizations still can’t prove they’re secure. Risk is still a guess, and AI is giving attackers the edge. Security leaders want proof. Proof of risk, resilience, and what actually matters. With continuous, attacker-driven testing — and #NodeZero, the world’s best AI hacker — teams can finally gain clarity, focus efforts, and prove their security posture. Massive thanks to our team on the ground in San Francisco for an incredible week, and to everyone who stopped by or joined a demo. We loved connecting! 👏 #RSAC26 #RSA2026 #ProactiveSecurity #AISecurity

    • No alternative text description for this image
    • No alternative text description for this image
    • No alternative text description for this image
    • No alternative text description for this image
    • No alternative text description for this image
      +1
  • Thank you to our friends at Code Red Partners and our AI Arms Race panelists, Snehal Antani, Damian Hasse, Evan M., Mark Hillick, and Mrityunjay Gautam, for discussing how AI is reshaping enterprise security. 👏 This was a strong reminder that as AI accelerates exploits and enables a new class of attackers, security teams must continuously test what works, fix what doesn’t, and stay ahead of emerging risks in real time. Brex Moveworks Instacart Robinhood #AISecurity #ProactiveSecurity #RSAC2026 #RSA2026

    • No alternative text description for this image
    • No alternative text description for this image
    • No alternative text description for this image
    • No alternative text description for this image
  • 🚨 A malicious credential stealer was discovered in BerriAI LiteLLM — and if you installed it, you should assume compromise. LiteLLM versions 1.82.7 and 1.82.8 were backdoored and distributed via PyPI, impacting a widely used open-source LLM proxy library (~95M monthly downloads). This is a supply chain attack, not a typical vulnerability. The malicious payload executes automatically (no import required in some cases) and: • Collects SSH keys, environment variables, cloud credentials, API keys, DB passwords • Exfiltrates data to an attacker-controlled domain • Attempts Kubernetes lateral movement and cluster takeover • Installs persistence via systemd backdoor If these versions were installed, your secrets are already gone. This affects: • Developer machines • CI/CD pipelines • Containers and production systems • Any downstream projects pulling LiteLLM as a dependency This campaign has been linked to a broader multi-ecosystem supply chain operation impacting GitHub Actions, Docker Hub, npm, and more. Immediate actions: ⚠️ Remove LiteLLM 1.82.7 and 1.82.8 ⚠️ Purge package caches to prevent reinstall ⚠️ Rotate ALL credentials present on affected systems ⚠️ Audit for persistence and Kubernetes compromise Safe version: 1.82.6 (last known clean) We are tracking this closely. While there is no targeted Rapid Response test for this advisory, this is a full incident response scenario, not a patch-and-move-on situation. If this package touched your environment, treat it as compromised. More info at https://lnkd.in/dCFWHMYN #AISecurity #ProactiveSecurity #infosec #cybersecurity

    • No alternative text description for this image
  • AI gives attackers the advantage, but Horizon3.ai’s AI-Native Proactive Security Platform enables organizations to level the playing field and fight AI with AI. That’s why we were recognized by Fast Company as one of the most innovative companies in 2026, alongside Sublime Security, Cyera, Chainguard, and more. Learn how we’re turning one of cybersecurity’s biggest challenges into one of its biggest strengths at https://lnkd.in/eWnxib6K. And if you're at #RSAC2026, come by and see us in person at booth N-4201! #ProactiveSecurity #AISecurity #FastCompany

    • No alternative text description for this image

Similar pages

Browse jobs

Funding