session-learn

To extract reusable knowledge from a conversation into wiki notes, the Skill identifies seven signal types including concepts, entities, corrections, patterns, ideas, decisions, and gaps, then persists them as durable records.

Saving chat session learnings as a decision log involves capturing decisions, corrections, and open gaps from the work session, applying conflict annotation to preserve contradictions rather than silently overwriting them.

Tracking knowledge gaps after a work session requires separating confirmed learning from open questions, logging unanswered gaps into a review queue for follow-up, and ensuring closure logging preserves low-confidence items.

Syncing extracted session patterns to existing wiki pages supports conflict annotation, preserving contradictions, updates, and low-confidence items explicitly rather than silently overwriting previous durable notes.

Conflict-aware closure for session knowledge capture is needed during end-of-session reviews and multi-step problem solving, ensuring deduplication and destination path resolution preserve contradictions and low-confidence items.

Session knowledge capture works for multi-step problem solving and end-of-session reviews by extracting durable wiki notes, logging corrections, and isolating open gaps for follow-up action.