learning-capture

Capturing patterns means identifying repeated workflows, problem-solving approaches, and domain knowledge from completed work, then formalizing them into reusable skills. After finishing a task, review moments where you solved a novel problem or optimized a process—if the pattern will save tokens and effort across 10+ future uses, capture it as a skill artifact with clear reasoning for why it's valuable.

A pattern qualifies for capture when it meets ROI criteria: expected to save over 500 tokens per use, apply to 10+ future tasks, and require less than 5000 tokens to load and execute. This ensures you're investing capture effort only in high-impact, repeatable knowledge that genuinely reduces cognitive load and token waste.

Learning moments include novel problem-solving approaches, recurring tasks you've automated or optimized, domain-specific knowledge you've synthesized, effective reasoning patterns that solve a class of problems, and workflow optimizations that cut processing time. Capture these when they meet ROI thresholds and will reduce repetition in future work.

Yes. Tacit knowledge—reasoning patterns, domain expertise, and contextual understanding built through experience—can be formalized into skill artifacts that reduce cognitive load and token consumption. The skill consolidates this knowledge into a structured process others (or future you) can apply consistently to similar problems.

Recognition identifies that a repeatable approach exists in your work. Capturing transforms that insight into a documented, reusable skill artifact with clear logic for when and how to apply it. Capture happens only when the pattern meets ROI criteria and produces a draft compatible with skill templates.