Research / Tag

prompt-injection

2 articles tagged prompt-injection

June 10, 202611 min read

Why Agentic AI Is the Most Dangerous Tool in Offensive Security

An agentic agent does my recon in fifteen minutes and never gets bored. The same autonomy that makes it a weapon makes it a liability, because one poisoned document turns its power against its owner. Why the more agentic the system, the further a single injection travels.

agentic ai-security offensive-security ai-agents prompt-injection

March 19, 202613 min read

Building Adversarial Agents: Your First AI Hackbot

Most AI security tools test for prompt injection with a list of 50 known payloads. That's a scanner, not a pentester. Here's how to build an AI that actually thinks like an attacker.

claude-code red-team ai-security hackbot prompt-injection