memory-system

What problem does it solve? Agents lose all context between sessions, forcing users to re-explain preferences, coding conventions, and past decisions every time. This Skill provides persistent, searchable memory so agents recall what was already learned instead of re-discovering it. ## Core Features & Use Cases - Structured Memory Index: Maintains a lightweight MEMORY.md index (max 200 lines) pointing to topic files for user preferences, project conventions, tech decisions, and feedback history. - Four Memory Operations: Save, recall, search, and prune memories triggered by phrases like "remember this" or "what do you remember about X". - Session Start Protocol: Automatically reads the memory index at session start and silently applies relevant context without reciting it. - Use Case: A user says "remember that we use bun instead of npm". The agent saves it to project-conventions.md, updates the index, and in future sessions automatically uses bun in all commands without being told again. ## Quick Start Ask the agent to remember a preference or convention, such as telling it to remember that you prefer concise responses and use PowerShell on Windows.