memory-bank-manager

What problem does it solve? AI agent sessions lose context between runs, causing repeated mistakes, stale status notes, and lost engineering lessons. This Skill keeps the memory_bank/ directory concise and current while promoting durable knowledge to docs/engineering-memory/. ## Core Features & Use Cases - Session context tracking: Maintains context.md with current focus, working scope, and next step so any agent can resume work immediately. - Rolling progress log: Keeps progress.md up to date with a rolling window of up to 15 entries covering done, in-progress, and upcoming work. - Decision promotion: Records operational decision summaries in decisions.md and classifies reusable lessons at closeout as promoted, operational-only, or not-applicable. - Use Case: After a long debugging session, an agent updates the memory bank, promotes the root-cause analysis to docs/engineering-memory/, and the next session resumes without re-reading old chat history. ## Quick Start Update the memory bank with the current session status and promote any durable engineering lessons to the engineering memory docs.