Jev AI: how the decision model works and where it fits
An introduction to TypeSafe AI's Jev: Choice, Noul and Score, calibrated probabilities, suitable tasks, and what my own experiments could establish.
Articles
PCS: a closed-loop system for controlling a product
Why I wrote down Product Control System: a methodology for agents that observe a running product, authorize and execute changes, verify the result, and update the system model.
My AI agent workspace
How I organize work with an AI agent: server notes, a task journal, projects, and an ops-log skill that records results and keeps instructions current.
White Rabbit: server security audits in Claude Code
How I use White Rabbit to audit a server over SSH: collect logs and configuration, review findings with Claude Code, and restrict commands with a read-only hook.