All Writing

NVIDIA announced NemoClaw at GTC 2026 as the security layer OpenClaw always needed. I evaluated both thoroughly — here is my honest verdict on the architecture, where NemoClaw genuinely improves the security posture, and where it still falls short.

15 min read

Bug Bounty 22 Mar 2026

I Built a Team of AI Agents for Bug Bounty Hunting. Three Weeks of Iteration Taught Me More Than Three Years of Reading About AI.

I deployed 6 specialised AI agents for bug bounty hunting using OpenClaw. Here's the honest account: initial scores, what broke, the triage turning point, and what actually improved.

9 min read

AI Security 15 Mar 2026

I Asked Two AI Models to Cross an Ethical Line. One Refused. One Didn't. Here's What That Means for Security.

Same prompt, same context, two different LLMs — two completely different decisions. A real experiment that exposes why inconsistent AI guardrails are themselves a security vulnerability.

11 min read

All Writing

NVIDIA NemoClaw vs OpenClaw: A Security and Architecture Deep-Dive (Is the Move Worth It?)

I Built a Team of AI Agents for Bug Bounty Hunting. Three Weeks of Iteration Taught Me More Than Three Years of Reading About AI.

I Asked Two AI Models to Cross an Ethical Line. One Refused. One Didn't. Here's What That Means for Security.