What problem does it solve? Valuable knowledge from AI conversations — debugging conclusions, design decisions, technical findings — disappears when the session ends. This Skill preserves that knowledge as permanent, structured wiki notes in an Obsidian vault instead of letting it evaporate. ## Core Features & Use Cases - Full capture mode: Classifies conversation content into five types (synthesis, concept, source, decision, session), rewrites it as declarative knowledge with provenance markers, and writes a cross-linked wiki page into the correct vault folder. - Quick mode (--quick): Stages findings to a _raw/ folder in under 60 seconds with no manifest, index, or log writes, designed for mid-session capture and automatic session-end preservation via a Stop hook. - Vault maintenance: Updates index.md, log.md, and hot.md tracking files, enforces at least two wikilinks per note, and optionally refreshes QMD embeddings. - Use Case: After a long debugging session that uncovered a Swift actor reentrancy deadlock, run the skill to file the finding as a tagged, confidence-scored wiki note linked to related pages. ## Quick Start Ask the AI to save this conversation to your wiki, or use quick mode to stage the key findings to _raw/ for later promotion.
