What problem does it solve? Agents lose context between sessions, leading to repeated mistakes, stale assumptions, and re-done work. This Skill enforces a mandatory session-start protocol that restores memory before any work begins. ## Core Features & Use Cases - Bounded memory loading: Reads context.md, corrections, session index, and pattern cards with strict token budgets so nothing is silently truncated. - Vault infrastructure verification: Checks symlinks, git remotes, disk space, and shared-memory routing before work starts. - Learning journal creation: Opens or creates a dated learning journal for real-time logging of mistakes, corrections, and wins. - Use Case: At the start of a coding session, run the protocol to load your agent's context, the other agent's context, unresolved corrections, and recent session history, then reply with a one-line memory-loaded receipt. ## Quick Start Tell the agent to run the memory-loader session start protocol and confirm what it remembers before beginning any work.
