@msm-core/learning

Auto-improvement for agents — capture goal-pursuit experiences, reflect them into reusable lessons (winning approaches + anti-patterns), and retrieve the relevant ones to seed the next plan. Storage- and brain-agnostic (inject a store port + a judge).