The Executable Harness
How scripts and checks can enforce safety rules that are easy to forget.
Notes from building real products.
Practical lessons from building software, running products, and using AI coding tools.
RSS feedThe Agentic Harness in Practice
A real example of the files, scripts, and safety rules that help AI coding tools work reliably.
Handling Context Compaction
How to keep long AI coding tasks accurate when the tool loses part of the conversation.
Multi-Agent Workflow Patterns
Four practical ways to divide research, coding, documentation, and security work among AI tools.
The Agentic Project Harness
A simple four-step guide to making a software project easier and safer for people and AI tools to work on.