What problem does it solve? Important project knowledge like architectural decisions, conventions, and pitfalls gets lost between sessions, forcing repeated re-investigation. This Skill defines when and how to read and write persistent project memory so durable knowledge survives across sessions. ## Core Features & Use Cases - Read-before-acting triggers: Lists and reads relevant memories before non-trivial tasks, unfamiliar errors, or questions about project conventions. - Write-after-learning rules: Captures non-obvious invariants, pitfalls, and architectural decisions under stable slash-separated topic paths like pitfalls/ and decisions/. - Generic memory operations: Covers list, read, write, edit, delete, rename, and project activation across any memory driver configured in docs/memory-providers.md. - Use Case: After debugging a tricky module lifecycle issue, write the findings under pitfalls/placements-lifecycle so the next session reads it before touching that code instead of re-investigating from scratch. ## Quick Start Read the project memory providers spec, list existing memories relevant to my current task, and write a new memory under decisions/ documenting the architectural choice we just made.
