I tell Claude to record everything in Linear as projects and tasks. Larger design docs get saved in git as markdown files, and the linear projects reference them by path.

I could never stand to use project management software when I had to do it by hand, but now that I can just talk through what I want to do and let the AI handle the paperwork, projects/tasks/milestones are a really useful tool for keeping track of what I'm working on.

I start the work week by telling Claude to go investigate Linear and Github to figure out what I was doing at the end of the previous week. Quick way to get back up to speed without having to think about any of it over the weekend.

I write a state.md, a changelog, some mermaid diagrams outlining the architecture, and some suggestions files that double as idea dumps of product roadmaps, and I update these at the end of every session. I go through these regularly, making corrections or annotations. General architectural/procedural insights can go in an agents.md file.