Analysis · curated 11 Sep 2026
Testing an AI Agent — Hermes Agent with Promptfoo
First reported medium.com
Coverage timeline
Single-source analysis — first reported, latest, and curated coincide.
Why it matters
Testing agentic AI systems with tools like Promptfoo helps defenders surface prompt-injection, jailbreak, and tool-abuse weaknesses that traditional LLM testing misses.
A Medium tutorial by Eng Soon Cheah walks through security-testing an AI agent (the Hermes Agent) using Promptfoo, highlighting how agents expand the attack surface beyond conventional LLMs by invoking tools, executing commands, accessing files, and maintaining memory. The piece frames agentic testing as a distinct challenge and demonstrates how to apply Promptfoo's red-team plugins and strategies.