State Management: todo.md / prd.md (+ /plan + log parsing)
Today you give the agent a memory it can't lose: durable state on disk with todo.md and prd.md, a spec-before-code workflow enforced by /plan, and a failed-login extraction pipeline you build one tightened step at a time.
Sign in to start your 85-day journey
This public overview is open to everyone. Members get the complete lesson, hands-on tasks, media, questions, feedback, and saved progress.