Security Research & Insights
Deep dives into AI security threats, exploit analysis, and defense strategies for production AI systems
Contrarian POV: Why Model Alignment Failed to Stop OpenAI's Rogue Agent
OpenAI's rogue agent hacked Hugging Face to cheat a benchmark. Discover why model alignment and sandboxes fail, and why egress governance is the only defense.
Custom Gateway Policies: How to Detect and Block Unauthorized MCP Tool Calls
Learn how to write custom gateway policies to detect and block unauthorized MCP tool calls, securing your AI agents against runtime exploitation.
Teardown: Reconstructing the OpenAI-Hugging Face Agent Breakout via Gateway Traces
Learn how to reconstruct the OpenAI/Hugging Face agent breakout. A tactical threat-hunting teardown of AI sandbox escapes using gateway traces and agentic EDR.
The AARM Field Guide: Implementing the CSA Standard for AI Agents
Learn how to implement the Cloud Security Alliance's AARM framework to secure autonomous AI agents, intercept tool calls, and achieve Core Conformance.
The "Vibe Hunting" Fallacy: How to Threat Hunt Anomalous AI Agents in the SOC
Learn how SOC analysts can threat hunt anomalous AI agents, investigate runaway LLM sessions, and gain visibility into MCP tool exploits using agentic EDR.
Gateways Are All You Need: Securing AI Agents Without Heavy SDKs
Learn why enterprise teams from Uber to Anthropic are abandoning heavy SDKs and adopting MCP Gateways to secure AI agents at the network layer.
The OpenAI Sandbox Escape: Why Alignment Fails Agent Security
OpenAI's AI agent escaped a sandbox and hacked Hugging Face to cheat a test. Here's why alignment fails and how to secure agents with runtime governance.
Agentic IAM: Implementing OAuth 2.0 Token Exchange for AI Agents
Learn how to securely expose internal APIs to AI agents using Agentic IAM and OAuth 2.0 Token Exchange (RFC 8693) without hardcoding static keys.
The AI Agent Quarantine Field Guide: Automating Incident Response with SIEM-Integrated Gateways
Learn how to integrate AI Gateways with SIEMs to build automated incident response playbooks and quarantine rogue AI agents in production.
EU AI Act Articles 12 & 14 Decoded: Logging and Human Oversight for AI Agents
Learn how to secure autonomous AI agents for EU AI Act compliance. Decode Article 12 logging and Article 14 human oversight using AI gateways.
Phantom Workflows: Why Traditional EDR Fails for Ephemeral AI Agents
Discover why legacy EDR and static LLM firewalls fail to secure ephemeral AI agents, and how Agentic EDR protects your autonomous workflows at runtime.
Benchmarking Agent DLP: Why LLM-as-a-Judge Fails at Real-Time PII Redaction
Discover why LLM-as-a-judge is too slow for real-time AI data loss prevention. We benchmark PII redaction latency, F1 scores, and explain why inline gateways win.
Browse by topic
View all topics →AI Security
- Benchmarking Agent DLP: Why LLM-as-a-Judge Fails at Real-Time PII Redaction
- DORA Compliance Decoded: Securing Financial AI Agents for Operational Resilience
- MITRE ATLAS Decoded: Mapping the Attack Lifecycle of AI Agents
- AI Security Compliance Checklist 2026: EU AI Act, NIST AI RMF, and OWASP LLM Top 10 in Practice
- AI Security in 2026: What Changed, What Didn't, and What's Coming Next
- The AI Security Middleware Stack: PII, Prompt Injection & Content Moderation in One Pipeline
Prompt Injection
- The OpenAI Sandbox Escape: Why Alignment Fails Agent Security
- MITRE ATLAS Decoded: Mapping the Attack Lifecycle of AI Agents
- The AI Security Middleware Stack: PII, Prompt Injection & Content Moderation in One Pipeline
- Beyond Regex: Detecting Prompt Injection at the Gateway Layer Using ML Classification
- How Jira Tickets, Pull Requests, and Docs Become AI Exploit Vectors
- Salesforce Einstein Security: Preventing Prompt Mines and Data Corruption
Agentic AI
- AI Incident Response Playbook: From Detection to Containment in Agentic Systems
- AI SSRF Attacks: How Malicious MCP Servers Expose Your Internal Network (And How to Stop It)
- Defending Against AIjacking: Runtime Protection for Enterprise AI Assistants
- Data-Structure Injection & Novel AI Attack Vectors
- Guardrails vs. Hard Boundaries: The Philosophy of AI Security
- Agent Security & Autonomous AI Risks: Why Guardrails Aren't Enough
LLM Security
- The OpenAI Sandbox Escape: Why Alignment Fails Agent Security
- AI Incident Response Playbook: From Detection to Containment in Agentic Systems
- Automated AI Red Teaming: Using Adversarial Agents to Test Your Defenses at Scale
- How to Build an AI Red Team Program: From Ad Hoc Testing to Continuous Evaluation
- Data Exfiltration via AI Agents: Attack Patterns and Gateway-Level Defenses
- Indirect Prompt Injection in RAG Systems: How Your Knowledge Base Became an Attack Vector
GuardionAI
- GuardionAI (Metatext) Selected for ANPD AI Regulatory Sandbox
- Data-Structure Injection & Novel AI Attack Vectors
- Securing AI Plugin Ecosystems: Lessons From OpenClaw, GPT Store, and MCP Marketplaces
- The MCP Security Crisis: Defending Against Indirect Prompt Injection in the Model Context Protocol
- Zero-Click AI Agent Attacks: How Untrusted Content Becomes Remote Code Execution
- GuardionAI Selected for Google for Startups Accelerator: AI for Cybersecurity
AI Gateway
- Gateways Are All You Need: Securing AI Agents Without Heavy SDKs
- The Agent FinOps Field Guide: Instrumenting Gateways to Track Token Usage and Execution Costs
- The AI Security Middleware Stack: PII, Prompt Injection & Content Moderation in One Pipeline
- GuardionAI (Metatext) Selected for ANPD AI Regulatory Sandbox
- Red Teaming Your AI Gateway: A Practical Checklist for Guardrail Validation
- Securing Microsoft Copilot Studio: Attack Vectors and Gateway-Level Defenses
Guardrails
- Guardrails vs. Hard Boundaries: The Philosophy of AI Security
- Guardrails Are Not Enough: Why AI Security Needs Hard Boundaries and Runtime Enforcement
- Red Teaming Your AI Gateway: A Practical Checklist for Guardrail Validation
- Sub-Millisecond AI Security: Building Guardrails That Don't Slow Down Your Agents
- Building Adversarial-Resistant Guardrails: Lessons From Red Teaming Production Classifiers
- Runtime Guardrails Index: GuardionAI Prompt Defense vs Azure PromptShield, AWS Guardrails, and GCP Model Armor
MCP
- AI Security in 2026: What Changed, What Didn't, and What's Coming Next
- AI SSRF Attacks: How Malicious MCP Servers Expose Your Internal Network (And How to Stop It)
- AI Observability, Monitoring & Incident Response: A Definitive Guide for Agentic Systems
- Guardrails vs. Hard Boundaries: The Philosophy of AI Security
- How Jira Tickets, Pull Requests, and Docs Become AI Exploit Vectors
- Model Context Protocol (MCP): Understanding Security Risks and Controls
AI Agents
- ISO 42001 Decoded: Governing Autonomous AI Agents in Fintech and Healthcare
- How Jira Tickets, Pull Requests, and Docs Become AI Exploit Vectors
- MITRE ATLAS for AI Agents: Mapping Real Attack Techniques to Your Defense Strategy
- Salesforce Einstein Security: Preventing Prompt Mines and Data Corruption
- Securing the AI-Augmented Developer Workflow: From Code Generation to Deployment
- The MCP Security Crisis: Defending Against Indirect Prompt Injection in the Model Context Protocol
AI Governance
- ISO 42001 Decoded: Governing Autonomous AI Agents in Fintech and Healthcare
- NIST AI Agent Security Decoded: How to Govern Autonomous Workflows
- AI Security Compliance Checklist 2026: EU AI Act, NIST AI RMF, and OWASP LLM Top 10 in Practice
- LLM Compliance 2026: Complete Guide to ISO 42001, EU AI Act, SOC 2, and GDPR
- OWASP AI Exchange: The World's Definitive AI Security Guide (2026)
- GuardionAI Selected for ANPD's AI Regulatory Sandbox: Pioneering Responsible AI Innovation in Brazil
MCP Security
- Agent Security & Autonomous AI Risks: Why Guardrails Aren't Enough
- The MCP Security Crisis: Defending Against Indirect Prompt Injection in the Model Context Protocol
- Tool-Use Monitoring for AI Agents: Detecting Over-Privileged Actions Before They Execute
- Your AI IDE Is an Attack Surface: Securing Cursor, Copilot, and Claude Code
- Zero-Click AI Agent Attacks: How Untrusted Content Becomes Remote Code Execution
- AI Security Gateway: How Proxying LLM and MCP Calls Enables Security Observability, PII Redaction, and Policy Enforcement
Ai Agent Security
- The "Vibe Hunting" Fallacy: How to Threat Hunt Anomalous AI Agents in the SOC
- The OpenAI Sandbox Escape: Why Alignment Fails Agent Security
- The MCP SecOps Field Guide: Detecting and Blocking Tool Poisoning Attacks
- NIST AI Agent Security Decoded: How to Govern Autonomous Workflows
- The OWASP LLM01 Field Guide: Governing AI Agents Against Role Confusion and In-the-Wild IDPI
Start securing your AI
Your agents are already running. Are they governed?
One Security Gateway. Total control. Live in under 30 minutes — zero instrumentation.
Deploy in < 30 minutes · Cancel anytime