agent-security

Preventing AI Agent Security Incidents: A Pre-Production Evaluation Playbook

A VentureBeat survey found 54% of enterprises have already hit an AI agent security incident — and most still let agents share credentials. Here's a practical playbook to evaluate agents against realistic adversarial conditions before they reach production.

07/21/2026 · Model Evaluation · 9 min read

The Three Gaps Killing Enterprise AI Agents in Production: Security, Evaluation, and Context

54% of enterprises have already had an AI agent security incident — and most still let agents share credentials. Here are the three gaps behind failing AI agents in production, and a checklist to close them.

07/19/2026 · Industry Trends · 8 min read

When AI Browsers Dream: How Prompt Injection Breaks Agentic Browsing (and How to Defend)

AI browser prompt injection isn't a one-off bug — it's a structural attack class that gets worse as agents gain autonomy. Here's how the new "dream-world" jailbreak works, and a concrete defense checklist for users and builders.

07/07/2026 · Research · 9 min read

How to Secure an AI Agent: Prompt Injection, Role Confusion, and Red-Teaming in 2026

In one week of June 2026, three independent sources reframed agent security — Willison's role-confusion model, the RIFT-Bench red-teaming benchmark, and the MosaicLeaks secret-leak demo. Here's how to secure an AI agent as a trust-boundary problem, not a string-filtering one.

06/25/2026 · Research · 9 min read