learn

To capture learnings into a memory store, instruct the AI to identify non-obvious session insights, verify their novelty against existing records, formulate actionable patterns, and append them to MEMORY.md for future reference.

The best way to avoid duplicate insights in knowledge management is to compare newly identified session learnings against existing MEMORY.md entries and current documentation before formulating and saving the pattern.

You should persist debugging quirks and patterns into memory after finishing a coding, debugging, or problem-solving session to ensure non-obvious takeaways are captured for future reference and continuous improvement.

Identifying non-obvious session insights works by applying disciplined prompts to extract takeaways from a session, verifying their novelty against the memory store, and formatting them as actionable patterns before saving them.

You need an existing memory store like MEMORY.md to save retrospective patterns, as the process requires appending validated insights to appropriate sections within the file for long-term recall and reuse.