What problem does it solve?
Long-running AI sessions produce decisions, corrections, and collaboration lessons that vanish when the session ends. This Skill performs an incremental memory consolidation so future sessions inherit real state changes and high-confidence collaboration experience instead of starting from zero.
Core Features & Use Cases
- Session Audit: Reviews only work added since the last successful update, distinguishing proposals, decisions, executions, and verified outcomes before writing anything.
- Cautious Memory Writing: Separates explicit user statements, observable facts, and model inference, keeping one-time instructions out of stable auto-loaded memory and avoiding fabricated user-preference profiles.
- Layered Memory Maintenance: Fixes outdated or conflicting existing memories first, then writes new facts at the minimal sufficient scope across workspace memory layers, with optional Git commit and push of related changes.
- Use Case: After a long product discussion session where the user corrected the AI's approach twice, run this Skill to record the verified decisions and the collaboration lesson so the next session continues with the right context.
Quick Start
Invoke the skill by its full name myagents-memory-update to consolidate this session's changes into workspace memory.
