Articles
-
Notes on Spec Drift
The types of drift that show up between Spec Kit specs and the code they describe. What to watch for, and what to do about each.
-
Building an Evaluation Harness for a Fleet of Agents
How we built an evaluation harness for conversational agents in an Employee Assistance Programme context, starting from SIM-VAIL and Petri and adding the domain-specific measures the published suites don't cover.
-
Software Team When the Spec Is the New Code
AI-first teams are getting smaller and the definition work is getting heavier. The PM-to-developer ratio is the least interesting part; the boundary between definition and delivery is what needs redesigning.
-
An AI-First Autonomous Build Environment for Flutter and AWS
How we set up a development environment where a coding agent goes from user story to pull request unsupervised: devcontainers, LocalStack, an Android emulator, and Marionette MCP.