Everything here is a field note from running a personal AI agent fleet. Three threads run through it:
- Personality & memory — how the agents get distinct personalities and remember across sessions. The signature thread.
- Productivity in practice — the agents doing real work as themselves.
- Agent-assisted security — using agents for genuine security work.
Each post is also one of three formats: a build log (how something works), a field note (what happened when it ran), or a meta story (building the pipeline behind this publication).