Cranking out Good Code 
Hey, I'm Joseph
I've built software that tested Olympic athletes for doping and pipelines adopted by 1,000+ engineers at Salesforce. These days I'm deep in agentic engineering — designing AI that actually ships work. I write about developer tooling, agents, and the craft of engineering at scale. More about me →
Recent Posts
-
Give Your Agents a Task Queue, Not a Task List
The simplest trick to make AI agents run longer, think better, and actually ship production code.
-
Hermes vs OpenClaw: Why I'm Running Both
After months running an OpenClaw agent swarm, I added Hermes as a standalone agent. Here's what each does well, where they break, and why your security team would block both tomorrow.
-
The First eBay API Skill on ClawHub — and Why Agent Marketplaces Are About to Explode
We just published the first skill on ClawHub that calls real eBay APIs. Here's how we built it, what the agentic skill economy looks like, and why smart buying and selling is one of the most compelling use cases for autonomous agents.
-
Fighting AI-Powered Fraud with AI-Powered Investigators
How typed tool orchestration, agentic reasoning, and human-in-the-loop design turn a manual fraud investigation process into a trustworthy, observable AI workflow.