Salt Security @SaltSecurity

AI agents don’t create risk just by hallucinating. They create risk when they take the wrong action. That action often moves through MCP, tools, and APIs—far beyond the model layer. To secure agentic AI, you need visibility from prompt to API. Read more: salt.security/blog/deconstru… #AISecurity #APISecurity #AgenticAI

@CTOPERU_ 🙌🌐

Salt Code is an absolute game changer 🧂 Don’t know about it? Don’t worry! Salt Security is hosting a Webinar June 16th with @cybrrnick that dives into all things Salt Code and AI security. Join us for the live product launch and see how Salt governs AI-generated code from the first prompt through runtime, without slowing your developers down → events.salt.security/saltcode?utm_m…

Zero wait. Agentless. With Salt Cloud Connect for AWS, organizations can instantly discover APIs, MCP servers, and agentic activity across AWS environments (without deploying sensors or routing traffic). In the agentic era, you can’t secure what you can’t see. Check how it’s done → youtu.be/gH9uFXrqbpM #cybersecurity #apisecurity #AWS #CloudConnect #AIsecurity @awscloud

AI guardrails protect the conversation. Agentic security protects the consequence. Agents don’t just respond... they call tools, access APIs, and take action across your business. The real question: can you see the full path from prompt → tool → API → action? Read more: salt.security/blog/everyone-… #cybersecurity #AIsecurity #APIsecurity #AgenticAI #AgenticSecurity #APIs #technology

We also have a cool page on Salt Code → salt.security/code?utm_mediu…

Or check out our recent blog → salt.security/blog/your-ai-c…

POV: Your developer asks an AI assistant to build an API. The AI delivers. Fast, clean-looking, functional, and everyone quickly moves to the next thing. But…. Did it enforce authentication? Apply rate limiting? Protect PII? Follow internal security standards? Avoid introducing a shadow API? Our latest research found that "29% of security leaders identify insecure coding patterns as the leading risk introduced by AI assistants." That’s the risk of AI-generated code: it can look right while missing the policies that make it safe. Salt Code was built for this exact shift. It enforces security policies inside AI coding assistants and carries governance across code, CI/CD, and runtime. Because “looks functional” is not the same as “safe for production.” Curious to know more about Salt Code? Tune into our webinar with @cybrrnick on June 16 → events.salt.security/saltcode?utm_m… #AICodeSecurity #APISecurity #AISecurity #DevSecOps #SaltCode

AI is writing more enterprise code than ever. The problem? AI coding assistants aren’t trained on your internal security policies, compliance requirements, or industry frameworks. That’s why we’re excited to launch Salt Code, the first agentic security solution to enforce security policies inside AI coding assistants. Salt Code brings policy-driven security to the moment code is created, helping developers generate compliant code by default from prompt to production. Secure code starts here: salt.security/press-releases… #cybersecurity #apisecurity #SaltCode #aisecurity

You can’t secure your AI with hope. In fact, your own agents are already exposing data that your security team forgot about ⚠️ AI agents are constantly discovering data, accessing systems, and taking action across clouds, SaaS, models, MCP servers, and APIs. And a single-platform view won’t show the full risk. In our blog, we break down why the Agentic Security Graph matters, and why hope is not an AI security strategy: salt.security/blog/even-goog… #cybersecurity #AIsecurity #AIagents #APIsecurity #AgenticAI #MCP #APIs

Gartner is almost here! Join Salt Security for our AI Agent Security Workshop on June 1, 2026 at the Residence Inn – National Harbor at the Gartner Security & Risk Management Summit 🌐 This hands-on workshop (facilitated by Lead Demo Engineer 😎) will give attendees practical experience identifying, analyzing, and securing real agentic attack paths. Breakfast is included, and attendees will also receive exclusive access to the Salt Lounge for the rest of the day 💜🧂 Register here: events.salt.security/AIsecuritythre… #AIAgentSecurity #APISecurity #Cybersecurity #AgenticAI #aisecurity

Heading to Gartner Security & Risk Management Summit? Stop by and see the Salt Security team at Booth #929 👋 We’ll be talking all things API+AI security, Agentic AI, MCP, and the evolving attack surface shaping the future of enterprise security. And this year, we’re also opening up the Salt Lounge at the Residence Inn! It’s a quiet space designed for attendees looking to recharge between sessions and meetings after our AI Agent Security workshop → events.salt.security/AIsecuritythre… Conferences can get hectic. Consider this your reset space. Looking forward to seeing everyone in National Harbor! #gartner #AIsecurity #AIagents #APIsecurity #cybersecurity #APIs #AgenticAI

Do you have visibility into your AI security risks? If not, you need to play catch-up quickly. As enterprises shift from conversational to agentic AI, the real risk moves from model outputs to the action layer; the MCP servers and APIs through which agents execute real-world tasks. This is where the Agentic Security Graph comes into play. If you missed our recent webinar, you can watch it here to learn more about the Agentic Security Graph, Salt’s three-step operational approach, and how to secure the AI action layer before it becomes the breach your organization is reacting to → youtu.be/Ek85tePzdN0 #cybersecurity #APIsecurity #AIsecurity #AgenticAI #AIAgents #MCP #ActionLayer

Q: "I thought we just needed to secure the LLM. That's not enough?" A: Not even close. The assumption is that if you lock down the model, you're safe. But the model isn't where the damage happens. Agents don't cause harm by thinking. They cause harm by acting. And those actions flow through a stack that most security tools have never been designed to see. LLM at the top. Orchestration layer in the middle. APIs and data systems at the bottom. Each layer is a potential attack surface. Each one can be exploited independently. Securing the LLM and ignoring the rest is like locking your front door and leaving the windows open. Salt sees the entire agentic stack, from model to action, so you understand where risk actually lives. Not where you assumed it did. #AgenticSecurity #AIRisk #CISOs #SecurityStrategy

Where does the real AI risk live? This episode of @EM360Tech podcast series "The Security Strategist", featuring Salt Security, dives into: -The role of AI Agents in security -Transitioning from discovery to governance -Protecting sensitive data in AI systems -Identifying weak points in AI security Learn how AI agents, shadow AI, and API security challenges are impacting enterprise cybersecurity: em360tech.com/podcasts/your-… #cybersecurity #APIsecurity #AISecurity #TechTrends #AgenticAI #CISOTips

“The organizations building their Agentic Security Graph now are not doing it because something went wrong. They are doing it because they understand that the cost of building it before an incident is a fraction of the cost of needing it after one.” Learn more about the Agentic Security Graph and what it takes to stay secure in today’s evolving AI landscape: salt.security/blog/your-ai-a… #AgenticSecurityGraph #AIsecurity #AIagents #AgenticAI #apisecurity #cybersecurity #AIsecurity

AI agents are creating new opportunities for innovation… as well as new security risks that teams need to understand ⚠️ Join Salt Security for our AI Agent Security Workshop on June 4, 2026 at the Residence Inn – National Harbor at the Gartner Security & Risk Management Summit 🌐 This hands-on workshop (facilitated by our Lead Demo Engineer) will give attendees practical experience identifying, analyzing, and securing real agentic attack paths. Breakfast is included, and attendees will also receive exclusive access to the Salt Lounge for the rest of the day 💜🧂 Timeline: -Breakfast 7:45 AM -Workshop 8:00 AM – 9:15 AM -Networking 9:15 AM - 10:30 AM Seats are limited (only 15 spots available), so save your spot now! Register here: events.salt.security/AIsecuritythre… #AIAgentSecurity #APISecurity #Cybersecurity #AgenticAI #aisecurity #GartnerSRM

Secure the AI action layer before it becomes the breach your organization is reacting to. Not sure how? We've got you covered. Join our upcoming webinar to learn more about… -The Agentic Security Graph: LLM (brain), MCP servers (hands), APIs (action layer). -Why MCP servers and API activity are the highest-risk, least-monitored surfaces. -A three-step operational approach: Discover, Govern (AG-SPM), Detect & Respond (AG-DR). -Identifying practical controls, telemetry sources, and workflows to reduce agentic risk and meet compliance. 🕘 May 19th at 9AM PST (12 PM EST) 🗣️ @cybrrnick Save your spot today → events.salt.security/theagenticsecu… #cybersecurity #webinar #aisecurity #agenticAI #apisecurity #AIagents #actionlayer

Q: "Shouldn't we be focused on securing the AI model itself?" A: It's the right instinct, but it's the wrong layer. The model doesn't move money. The model doesn't delete records. The model doesn't exfiltrate data. Agents do. And agents operate one layer below the model, at the action layer, where instructions become real-world consequences. The entire security industry is staring up at the LLM. Meanwhile, attackers are already working the layer below it. They're manipulating the tools agents use, compromising the MCP servers that broker actions, and exploiting the APIs that agents call. None of that requires touching the model at all. If your security strategy stops at the model, you have a plan for the wrong problem. Salt Security is built for the action layer. That's where risk is manufactured. That's where we focus. #AgenticSecurity #ActionLayer #AIRisk #cybersecurityStrategy #cybersecurity #AIsecurity #APIsecurity

Amazing discussion on security and agentic AI with Roey Eliyahu (Salt Security), Irina Tsukerman (Scarab Rising), Shlomi Beer (ImpersonAlly), and hosted by Kristian McCann (@uctodaynews) → uctoday.com/security-compl… Find out what agentic AI is, where the risks are emerging, and how organizations can manage systems designed to act on their behalf. Thank you for including Roey in this insightful roundtable! #cybersecurity #apisecurity #aisecurity #AgenticAI #AIagents #roundtable #thoughtleaders #thoughtleadership