learning-capture

To capture reusable patterns and domain knowledge, recognize specific learning moments like novel problem-solving or workflow optimizations, evaluate them against a decision framework, and structure the captured insights into a draft skill for future reuse.

You should capture workflow optimizations when high ROI is expected, specifically identifying patterns that will save 500 or more tokens per reuse across at least 10 future uses, while ensuring the knowledge is not redundant with existing capabilities.

The pattern recognition process triggers on five learning moment types: novel problem-solving, repeated patterns, domain-specific knowledge, effective reasoning patterns, and workflow optimizations encountered during active work sessions.

You evaluate a captured pattern using a structured decision framework that checks for high context window ROI, strong reusability signals, and confirms the knowledge is not redundant with existing capabilities before offering to structure it into a draft skill.

The skill capture process provides specific templates for structuring different types of domain knowledge based on the identified learning moment type, alongside detailed examples and decision threshold guidelines to guide the capture.

The best way to avoid capturing redundant workflows is to apply a conservative approach within the decision framework, verifying strong reusability signals and confirming the knowledge is not redundant with existing capabilities before offering capture.