Agent Wrapped: Build a Year-in-Review of Everything Your AI Agent Remembers

Coding agents are brilliant, fast, and have the memory of a goldfish. Every new session starts from zero, and that amnesia quietly costs teams time, tokens, and trust. It shows up in three ways: You explain the same decisions again and again. Conventions, constraints, and “we tried that, it broke prod” live in old sessions that the next agent never sees. Agents burn tokens rediscovering your codebase. Before writing a line, a fresh agent spends its budget figuring out what the last one already knew. Stale memory is worse than none. When code changes but remembered facts don’t, agents act confidently on information that’s no longer true, and reviewers can’t tell. In this hands-on workshop, you’ll build Agent Wrapped: a shareable year-in-review card showing what your AI agent learned, what it recalls, and what it had to unlearn. Along the way, your agent will “die” every round and come back with no context except what Entire remembers. You’ll learn how to turn sessions into lasting project memory with Entire Brain, give fresh agents a head start without wasting tokens, catch and fix memory that has gone stale after a refactor, and use Trails to check that changes respect what the project already knows. Walk out with a working tool, a card worth posting, and an agent that remembers why you made every weird decision. This workshop is for engineers, tech leads, and anyone tired of re-explaining their codebase to a coding agent. All experience levels are welcome; bring a laptop and a coding agent.