We are at the IANS Security Forum in New York today. Our CMO, @aakela, presented on the lessons learned from the Hugging Face attack and how our Deception Guardrails can help defend against misaligned agents and AI-driven threats. #CyberDeception #AgenticAISecurity

𝐖𝐞’𝐫𝐞 𝐭𝐡𝐫𝐢𝐥𝐥𝐞𝐝 𝐭𝐨 𝐰𝐞𝐥𝐜𝐨𝐦𝐞 𝐉𝐨𝐞 𝐒𝐮𝐥𝐥𝐢𝐯𝐚𝐧 𝐭𝐨 𝐀𝐜𝐚𝐥𝐯𝐢𝐨’𝐬 𝐀𝐝𝐯𝐢𝐬𝐨𝐫𝐲 𝐁𝐨𝐚𝐫𝐝! Joe, 𝐟𝐨𝐫𝐦𝐞𝐫 𝐂𝐒𝐎 𝐨𝐟 𝐌𝐞𝐭𝐚, 𝐔𝐛𝐞𝐫, 𝐚𝐧𝐝 𝐂𝐥𝐨𝐮𝐝𝐟𝐥𝐚𝐫𝐞, brings decades of cybersecurity leadership and a compelling perspective on why 𝐝𝐞𝐜𝐞𝐩𝐭𝐢𝐨𝐧 𝐢𝐬 𝐡𝐚𝐯𝐢𝐧𝐠 𝐢𝐭𝐬 𝐝𝐚𝐲 𝐢𝐧 𝐭𝐡𝐞 𝐚𝐠𝐞 𝐨𝐟 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐀𝐈. In a guest Acalvio blog, Joe shares what drew him to Acalvio: the growing importance of runtime detection as AI agents operate with real credentials and access to enterprise systems—and the ability to deploy deception at scale, a challenge that has historically limited adoption. As Joe puts it: "For years, the honest objection to deception was that it was too hard to run. That objection is getting weaker at the same moment the reason to do it is getting stronger." As AI agents introduce a new class of insider risk and AI-powered attacks accelerate, deception is becoming an essential environmental guardrail for preemptive cybersecurity. We’re excited to have Joe’s expertise and perspective as we help enterprises secure the agentic era. 👉 𝐑𝐞𝐚𝐝 𝐉𝐨𝐞’𝐬 𝐛𝐥𝐨𝐠, “𝐃𝐞𝐜𝐞𝐩𝐭𝐢𝐨𝐧 𝐈𝐬 𝐇𝐚𝐯𝐢𝐧𝐠 𝐈𝐭𝐬 𝐃𝐚𝐲”: acalvio.com/blog/cyber-decep… #Acalvio #AgenticAI #Cybersecurity #CyberDeception #PreemptiveCybersecurity

We’re excited to be at the FS-ISAC 𝐅𝐚𝐥𝐥 𝐒𝐮𝐦𝐦𝐢𝐭 𝐢𝐧 𝐀𝐮𝐬𝐭𝐢𝐧 𝐧𝐞𝐱𝐭 𝐰𝐞𝐞𝐤! If you’re attending, don’t miss the session “𝐅𝐫𝐨𝐦 𝐌𝐨𝐝𝐞𝐥 𝐆𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬 𝐭𝐨 𝐑𝐮𝐧𝐭𝐢𝐦𝐞 𝐂𝐨𝐧𝐭𝐫𝐨𝐥𝐬: 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲” by 𝐒𝐚𝐢 𝐏𝐮𝐥𝐮𝐠𝐮𝐫𝐭𝐡𝐚 𝐨𝐟 𝐕𝐢𝐬𝐚 and 𝐒𝐮𝐫𝐢𝐥 𝐃𝐞𝐬𝐚𝐢 𝐨𝐟 𝐀𝐜𝐚𝐥𝐯𝐢𝐨. This session examines the runtime controls needed to govern agent identity and data exposure. Visa and Acalvio security leaders will share insights on: ✅ The Agentic AI threat surface ✅ Runtime controls to govern agent access ✅ How preemptive techniques, including deception, can expose misuse earlier in the attack path ✅ A practitioner’s perspective on adapting cyber defenses for the AI era We look forward to seeing you in Austin! #FSISAC #AgenticAI #AISecurity #CyberSecurity #RuntimeSecurity #CyberDeception

In AI security, the perfect heist doesn't break a single rule—it uses authorized actions to clean out the vault. While security teams double down on prompt classifiers and silicon-level ACLs, agentic AI is already "living off the land"—passing hidden communications through package managers and chaining valid credentials without triggering a single standard alarm. The uncomfortable truth? Classifiers miss novel behavior, and static rules can't spot malicious multi-step sequences. In our latest blog, we break down why current enterprise AI guardrails have a massive blind spot—and why the future of AI defense lies in pre-emptive, model-aware Deception Guardrails. Discover how embedding synthetic honeytokens, decoy MCP tools, and model-targeted bait into LLM retrieval pathways creates an unambiguous signal that halts rogue agents instant-by-instant—without needing to parse intent or wait for an ACL violation. 👉 𝐒𝐭𝐨𝐩 𝐚𝐬𝐤𝐢𝐧𝐠 "𝐃𝐢𝐝 𝐚 𝐫𝐮𝐥𝐞 𝐠𝐞𝐭 𝐛𝐫𝐨𝐤𝐞𝐧?" 𝐚𝐧𝐝 𝐬𝐭𝐚𝐫𝐭 𝐚𝐬𝐤𝐢𝐧𝐠 "𝐈𝐬 𝐬𝐨𝐦𝐞𝐨𝐧𝐞 𝐭𝐨𝐮𝐜𝐡𝐢𝐧𝐠 𝐰𝐡𝐚𝐭 𝐭𝐡𝐞𝐲 𝐬𝐡𝐨𝐮𝐥𝐝𝐧'𝐭?" Read the full post to rethink your agentic security stack: acalvio.com/agentic-ai-secur… #AgenticAI #AISecurity #CyberSecurity #CyberDeception #PreemptiveCybersecurity #RuntimeSecurity

𝐌𝐨𝐝𝐞𝐥 𝐠𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬 𝐚𝐫𝐞 𝐞𝐬𝐬𝐞𝐧𝐭𝐢𝐚𝐥. 𝐁𝐮𝐭 𝐰𝐡𝐚𝐭 𝐡𝐚𝐩𝐩𝐞𝐧𝐬 𝐰𝐡𝐞𝐧 𝐚𝐧 𝐀𝐈 𝐚𝐠𝐞𝐧𝐭 𝐠𝐞𝐭𝐬 𝐩𝐚𝐬𝐭 𝐭𝐡𝐞𝐦? As AI agents gain greater autonomy, organizations need an independent layer of protection in the environments where those agents actually operate. In his latest article for Cyber Security News ®, our CEO Ram Varadarajan explains why 𝐞𝐧𝐯𝐢𝐫𝐨𝐧𝐦𝐞𝐧𝐭𝐚𝐥 𝐠𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬 are becoming critical for defending against AI-driven and agentic attacks. By placing deception—including decoys and honeytokens—across the operating environment, defenders can: • Detect malicious or misaligned agent activity early • Disrupt reconnaissance, credential abuse, and lateral movement • Manipulate the agent’s view of the environment • Add an independent runtime security layer beyond model-level guardrails 𝐌𝐨𝐝𝐞𝐥 𝐠𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬 𝐡𝐞𝐥𝐩 𝐠𝐨𝐯𝐞𝐫𝐧 𝐭𝐡𝐞 𝐀𝐈. 𝐄𝐧𝐯𝐢𝐫𝐨𝐧𝐦𝐞𝐧𝐭𝐚𝐥 𝐠𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬 𝐡𝐞𝐥𝐩 𝐩𝐫𝐨𝐭𝐞𝐜𝐭 𝐭𝐡𝐞 𝐞𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞 𝐰𝐡𝐞𝐧 𝐭𝐡𝐞 𝐀𝐈 𝐚𝐜𝐭𝐬. Read Ram’s perspective: cybersecuritynews.com/defend… #AgenticAI #AISecurity #CyberDeception #PreemptiveCybersecurity #RuntimeSecurity

🛡️ 𝐎𝐜𝐭𝐨𝐛𝐞𝐫 𝐢𝐬 𝐂𝐲𝐛𝐞𝐫𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐀𝐰𝐚𝐫𝐞𝐧𝐞𝐬𝐬 𝐌𝐨𝐧𝐭𝐡—𝐚𝐧𝐝 𝐰𝐞’𝐫𝐞 𝐤𝐢𝐜𝐤𝐢𝐧𝐠 𝐨𝐟𝐟 𝐚 𝐦𝐨𝐧𝐭𝐡 𝐨𝐟 𝐢𝐧𝐬𝐢𝐠𝐡𝐭𝐬 𝐟𝐫𝐨𝐦 𝐀𝐜𝐚𝐥𝐯𝐢𝐨. Throughout Cybersecurity Awareness Month 2026, we’ll share perspectives and practical guidance on how organizations can strengthen their defenses as cyberattacks become increasingly automated and AI-driven. We’re starting with a fundamental question: 𝐖𝐡𝐚𝐭 𝐝𝐨𝐞𝐬 𝐚𝐧 𝐚𝐭𝐭𝐚𝐜𝐤𝐞𝐫 𝐬𝐞𝐞 𝐰𝐡𝐞𝐧 𝐭𝐡𝐞𝐲 𝐥𝐨𝐨𝐤 𝐚𝐭 𝐲𝐨𝐮𝐫 𝐞𝐧𝐯𝐢𝐫𝐨𝐧𝐦𝐞𝐧𝐭? Attackers rely on reconnaissance to understand your infrastructure, identities, credentials, and attack paths. But what if you could control what they see—and what they trust? Our latest blog explores how 𝐝𝐞𝐜𝐞𝐩𝐭𝐢𝐨𝐧 𝐜𝐚𝐧 𝐫𝐞𝐬𝐡𝐚𝐩𝐞 𝐭𝐡𝐞 𝐚𝐭𝐭𝐚𝐜𝐤𝐞𝐫’𝐬 𝐫𝐞𝐚𝐥𝐢𝐭𝐲, exposing malicious intent early and shifting the advantage back to defenders. 👉 Read the first blog in our Cybersecurity Awareness Month series: lnkd.in/geVcCMQq And follow Acalvio throughout October for 𝐰𝐞𝐞𝐤𝐥𝐲 𝐂𝐲𝐛𝐞𝐫𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐀𝐰𝐚𝐫𝐞𝐧𝐞𝐬𝐬 𝐌𝐨𝐧𝐭𝐡 𝐢𝐧𝐬𝐢𝐠𝐡𝐭𝐬. #CybersecurityAwarenessMonth #Cybersecurity #CyberDeception #PreemptiveCybersecurity #AISecurity #AgenticAI

🚀 𝐀𝐜𝐚𝐥𝐯𝐢𝐨 𝐥𝐚𝐮𝐧𝐜𝐡𝐞𝐬 𝐭𝐡𝐞 𝐒𝐡𝐚𝐝𝐨𝐰𝐏𝐥𝐞𝐱 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭 𝐢𝐧 𝐭𝐡𝐞 Google 𝐆𝐞𝐦𝐢𝐧𝐢 𝐄𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞 𝐀𝐠𝐞𝐧𝐭 𝐌𝐚𝐫𝐤𝐞𝐭𝐩𝐥𝐚𝐜𝐞. Security teams can now describe what they want to protect or detect in 𝐧𝐚𝐭𝐮𝐫𝐚𝐥 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞, and the ShadowPlex AI Agent automatically generates a ready-to-deploy deception playbook—in minutes. The agent can create deception strategies to: ✅ Disrupt AI-driven reconnaissance and lateral movement ✅ Detect credential abuse and protect privileged identities ✅ Secure AI agents and AI-enabled infrastructure ✅ Extend deception across cloud, identity, IT, OT, and on-premises environments As our CEO @Ram_Varadarajan explains: “The ShadowPlex AI Agent closes that gap by turning security objectives into actionable deception playbooks, allowing organizations to deploy deception with greater speed, consistency, and scale.” This is an important step toward using agentic AI to defend against AI-powered attacks—and protect the enterprise AI infrastructure itself. The ShadowPlex AI Agent is available now through the Gemini Enterprise Agent Marketplace. Read the announcement: acalvio.com/press-releases/a… #AgenticAI #AISecurity #CyberSecurity #CyberDeception #PreemptiveSecurity #GeminiEnterprise #GoogleCloud

𝐖𝐡𝐞𝐧 𝐚𝐧 𝐀𝐈 𝐚𝐠𝐞𝐧𝐭 𝐢𝐬 𝐭𝐨𝐥𝐝 “𝐧𝐨,” 𝐰𝐡𝐚𝐭 𝐡𝐚𝐩𝐩𝐞𝐧𝐬 𝐢𝐟 𝐢𝐭 𝐤𝐞𝐞𝐩𝐬 𝐩𝐫𝐨𝐛𝐢𝐧𝐠 𝐮𝐧𝐭𝐢𝐥 𝐢𝐭 𝐟𝐢𝐧𝐝𝐬 𝐚 𝐰𝐚𝐲 𝐢𝐧? The recent incident in which an OpenAI agent gained unauthorized access to an Australian government health portal highlights a new challenge in agentic AI security: autonomous agents can bypass access controls even when their original task is legitimate. In this @SecurityMag article by Jordyn Alger, our CEO, @Ram_Varadarajan, highlights two critical lessons: • Static access controls alone are insufficient. Organizations need independent, deception-based guardrails to detect and disrupt unauthorized agent behavior. • A three-month delay in disclosure underscores the need for stronger monitoring, detection, and incident reporting. As Ram emphasizes, organizations need cybersecurity and safety measures that go beyond what foundation model companies provide. 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐀𝐈 𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐝𝐞𝐦𝐚𝐧𝐝𝐬 𝐩𝐫𝐞𝐞𝐦𝐩𝐭𝐢𝐯𝐞 𝐝𝐞𝐟𝐞𝐧𝐬𝐞 𝐚𝐧𝐝 𝐫𝐮𝐧𝐭𝐢𝐦𝐞 𝐞𝐧𝐯𝐢𝐫𝐨𝐧𝐦𝐞𝐧𝐭𝐚𝐥 𝐠𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬. Read the full article and Ram’s perspective: securitymagazine.com/article… #AgenticAI #AISecurity #CyberDeception #PreemptiveCybersecurity #RuntimeSecurity

On Sept. 29, Acalvio Technologies is co-hosting the Google Cloud Security Innovations Forum 2026 in Sunnyvale, CA. We’re bringing together top enterprise CISOs and security leaders to map out the shift to continuous, autonomous defense. What’s on the agenda? • How to deploy specialized AI security agents that predict attack paths. • Inside metrics from Mandiant’s landmark M-Trends 2026 Report. • Peer-to-peer roundtable discussions on fighting AI with AI. This is an exclusive, in-person event, and seats are filling quickly. Reserve your seat here: cvisionintl.com/events/summi…

𝐂𝐈𝐒𝐀 𝐡𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬 𝐜𝐲𝐛𝐞𝐫 𝐝𝐞𝐜𝐞𝐩𝐭𝐢𝐨𝐧 𝐚𝐬 𝐚 𝐜𝐫𝐢𝐭𝐢𝐜𝐚𝐥 𝐜𝐚𝐩𝐚𝐛𝐢𝐥𝐢𝐭𝐲 𝐟𝐨𝐫 𝐬𝐭𝐫𝐨𝐧𝐠𝐞𝐫 𝐝𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐫𝐞𝐬𝐩𝐨𝐧𝐬𝐞. New CISA guidance outlines how cyber decoys—including 𝐡𝐨𝐧𝐞𝐲𝐭𝐨𝐤𝐞𝐧𝐬, 𝐛𝐫𝐞𝐚𝐝𝐜𝐫𝐮𝐦𝐛𝐬, 𝐚𝐧𝐝 𝐭𝐫𝐢𝐩𝐰𝐢𝐫𝐞𝐬—can generate high-fidelity alerts, detect post-compromise activity, and complement Zero Trust strategies. For defenders, it’s another important step toward preemptive cybersecurity: detecting attacker intent early, before critical assets are reached. Read the CISA guidance: cisa.gov/resources-tools/res… #CyberSecurity #CyberDeception #PreemptiveSecurity #CISA #ZeroTrust #ThreatDetection

🛡️ 𝐀𝐜𝐚𝐥𝐯𝐢𝐨 𝐢𝐬 𝐡𝐞𝐚𝐝𝐢𝐧𝐠 𝐭𝐨 𝐁𝐥𝐮𝐞 𝐓𝐞𝐚𝐦 𝐂𝐨𝐧 𝟐𝟎𝟐𝟔! We’re excited to sponsor 𝐁𝐥𝐮𝐞 𝐓𝐞𝐚𝐦 𝐂𝐨𝐧, bringing together the cybersecurity defender community in Chicago, September 10–13. If you’re attending, don’t miss Suril Desai, 𝐕𝐏 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 𝐚𝐭 𝐀𝐜𝐚𝐥𝐯𝐢𝐨, as he shares practical insights on: 🎙️ 𝐃𝐞𝐜𝐞𝐩𝐭𝐢𝐨𝐧 𝐒𝐭𝐫𝐚𝐭𝐞𝐠𝐲 𝐟𝐨𝐫 𝐒𝐞𝐜𝐮𝐫𝐢𝐧𝐠 𝐂𝐥𝐨𝐮𝐝 𝐖𝐨𝐫𝐤𝐥𝐨𝐚𝐝𝐬 📅 Sunday, September 13 | 11:30–11:55 AM CDT 📍 Swissôtel Chicago 🔗 Session details: blueteamcon2026.sched.com/ev… Cloud environments introduce unique challenges for deception. Suril will explore how defenders can design an effective deception strategy across 𝐡𝐨𝐧𝐞𝐲𝐭𝐨𝐤𝐞𝐧𝐬 𝐚𝐧𝐝 𝐡𝐨𝐧𝐞𝐲𝐩𝐨𝐭𝐬—including where to place deceptions, what they should represent, and how to protect dynamic GCP and GKE workloads. Come meet the Acalvio team and learn how AI-powered deception and preemptive cyber defense can help detect attacker intent early and protect cloud environments. See you in Chicago! #BlueTeamCon #CyberSecurity #CloudSecurity #CyberDeception #PreemptiveSecurity #GCP

🚨 𝐖𝐡𝐚𝐭 𝐡𝐚𝐩𝐩𝐞𝐧𝐬 𝐰𝐡𝐞𝐧 𝐚𝐧 𝐀𝐈 𝐚𝐠𝐞𝐧𝐭 𝐛𝐞𝐜𝐨𝐦𝐞𝐬 𝐭𝐡𝐞 𝐚𝐭𝐭𝐚𝐜𝐤𝐞𝐫? The recent 𝐇𝐮𝐠𝐠𝐢𝐧𝐠 𝐅𝐚𝐜𝐞 𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐢𝐧𝐜𝐢𝐝𝐞𝐧𝐭 𝐢𝐧𝐯𝐨𝐥𝐯𝐢𝐧𝐠 𝐦𝐢𝐬𝐚𝐥𝐢𝐠𝐧𝐞𝐝 𝐀𝐈 𝐚𝐠𝐞𝐧𝐭𝐬 offers a glimpse into a new era of cyber threats—where autonomous agents can perform reconnaissance, escalate privileges, move laterally, and interact with production infrastructure at machine speed. Join 𝐒𝐮𝐫𝐢𝐥 𝐃𝐞𝐬𝐚𝐢, 𝐕𝐏 𝐨𝐟 𝐃𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 𝐚𝐭 𝐀𝐜𝐚𝐥𝐯𝐢𝐨, for our upcoming webinar: 𝐂𝐲𝐛𝐞𝐫 𝐑𝐞𝐬𝐢𝐥𝐢𝐞𝐧𝐜𝐞 𝐟𝐨𝐫 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐀𝐈: 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠𝐬 𝐟𝐫𝐨𝐦 𝐭𝐡𝐞 𝐇𝐮𝐠𝐠𝐢𝐧𝐠 𝐅𝐚𝐜𝐞 𝐀𝐭𝐭𝐚𝐜𝐤 📅 September 15, 2026 🕙 10 AM PT | 1 PM ET Suril will explore: ✅ How agentic AI attacks differ from traditional human-led attacks ✅ Why finding the malicious attack path among thousands of AI-generated actions is so challenging ✅ How preemptive defense can detect and disrupt adversarial activity early—starting at reconnaissance ✅ Practical strategies for building cyber resilience against autonomous, machine-speed threats Register today 👇 acalvio.com/resources/webina… #AgenticAI #AISecurity #CyberSecurity #PreemptiveSecurity #CyberDeception #HealthcareSecurity #HuggingFace

We are at CrowdStrike Fal.Con event this week in Las Vegas. Meet Acalvio experts at booth #2104 at Fal.Con 2026, or schedule a meeting ( acalvio.com/resources/events…) in our meeting room to see live attack demonstrations and learn about: • Agentic attacks that leverage concurrency, autonomy, and dynamic reasoning • How deception serves as an environmental guardrail for agentic AI systems • Innovations in 360 Deception, including techniques that make real assets appear fake to disrupt reconnaissance and attack progression #CyberDeception #AgenticAISecurity #BotvsBot