Blog
Insights on AI security, prompt injection defense, and cyber deception.
AI SecurityWhitepaperMethodology
Introducing the AI Security Lifecycle: A Four-Phase Methodology
Our new whitepaper presents a comprehensive framework for AI security: Test, Defend, Deceive, Learn.
AI SecurityVendor NeutralityPromptfoo
Why Your AI Security Should Not Be Owned by an AI Model Vendor
OpenAI acquired Promptfoo. Here is why vendor-neutral AI security matters more than ever.
ReleaseSecurityMCP
Shield v1.1: 18 Security Fixes, MCP Servers, and License Metering
Oubliette Shield v1.1 hardens the codebase with 18 security fixes from our own red team assessment, adds MCP server integration for both Shield and Dungeon, and introduces a soft license/metering layer for the Pro tier.
AI SecurityPrompt InjectionCyber Deception
Stop Blocking Prompt Injections. Start Trapping Attackers.
Why traditional AI firewalls fail, and how cyber deception turns prompt injection defense into an intelligence-gathering operation.