Laptop screen required

Please use a laptop or larger screen to view the writing.

These essays are designed around wide, interactive technical layouts.

Writing

Technical reflections from real systems work.

Notes on AI agent systems, evaluation infrastructure, full-stack product engineering, and the operating discipline that makes automation trustworthy.

AI agentsSystems architectureSecurityEvaluation

How Does OpenClaw Work?

A production-minded systems tour of OpenClaw: Gateway architecture, context assembly, hooks and plugins, file-backed memory and recall, sandboxed tool execution, agent runs, QA scenarios, security audits, and red-team patterns you can reuse in your own agent platform.

2026-06-09 / 13 min read

Read article