Projects
Some projects I've been working on lately, all about helping people and coding agents work well together. Two are works in progress: an experiment in how people and agents could share the work, and a .NET demo app. The other two are agent skills you can use today.

aceflow (Agentic Co-Engineering) is an experimental work in progress. It's an idea for a tool to help people and coding agents build better software together. A team hands an issue to a coding agent, which works through intent, plan, build and quality checks, and asks people to step in at each decision. A person still reviews every pull request. It's only mockups so far.

Energy1 is a working demo app, still in progress, for a fictional energy supplier. It shows Domain-Driven Design, a Modular Monolith and Vertical Slice Architecture in .NET, with Agent Skills that help coding agents work in the codebase. Each module (Customers, Tariffs, Metering and Billing) is a bounded context made of vertical slices, and the modules share data only through integration events. Meter readings come from a simulated third-party Meter Reading API through an anti-corruption layer, and an agentic workflow ingests new meter readings and passes them back to that API.
An agent skill that reviews context written for AI agents against the five CAFE(S) properties: clarity, actionability, fidelity, efficiency and security. View on GitHub
An agent skill that makes an AI agent write technical text in Simplified Technical English (STE). STE is a controlled form of English. The goal of STE is that each reader gets the same meaning on the first read. View on GitHub

