The Book Bag — JKE University

Your agent has no memory between sessions. Every fresh start is a factory reset. System prompts scroll out of view after an hour. Your agent learns something Tuesday and forgets it Wednesday — not because it's careless, but because nothing was written down.

Two plain markdown files fix this. A journal — one entry per session, what shipped, what was decided, what was learned. Your agent writes at the end of every session and reads all entries on boot. And a tool catalog — in-house tools that survive platform updates because they live in your workspace, not in a plugin directory that gets wiped on upgrade day.

Your agent verifies the system is healthy, creates both files, and logs its first lesson. Course 3 builds your foundation buckets directly from what's already in your workspace — no cold interview, no guessing. Real data, filed with intention.

Your AI agent loses everything between sessions and reinvents its tooling every time because nothing is written down. Install two plain markdown files: a journal (school.md) where your agent logs one entry per session — what shipped, what was decided, what was learned — and reads every entry on boot, plus a tool catalog (book-bag.md) that keeps your in-house tools alive across platform updates because they live in your workspace, not a plugin directory. Two files, two habits. No database, no plugin. The book bag is what your agent carries between sessions.