compound-learn

To document project learnings, you extract decisions, patterns, and pitfalls from completed work and write them into a knowledge base. This captures insights under ~/knowledge/projects/ or ~/knowledge/patterns/ for future cross-project reuse.

The best way to capture decisions from a post-mortem is to run an extraction step after task completion but before closing. This files the documented patterns and best practices directly into the memory system for future reuse.

You should document insights into a knowledge base immediately after reviewing a completed task and before closing it. This locks in learnings from post-mortems, task reviews, and project retrospectives while the context is fresh.

Yes, this knowledge extraction process is applicable across software, product, and research contexts. It captures actionable learnings from various project types and stores them in a centralized memory system for cross-project reuse.

Storing extracted learnings in a memory system improves workflows by preventing the same problem from arising again. It captures decisions and patterns in the knowledge base, making insights available for future reuse across projects.

No specific dependencies are required to file learnings under a knowledge base directory. The process implements destination-specific frontmatter and writing rules to organize files under ~/knowledge/projects/ or ~/knowledge/patterns/ autonomously.