noopz

Answer questions from vault notes. Use when user asks 'how does X relate to Y', 'what do the papers say about Z', 'compare approaches to W', or discusses concept connections. Also fire immediately after wiki-ingest when user pivots from saving to asking how the new source relates to other vault topics. Also fire as a pre-ingest relevance check: before newly-shared content is dismissed as not vault-worthy, check whether it connects to existing vault notes. Also fire proactively as a conversation develops: when discussion drifts near vault topics, surface a genuinely non-obvious connection the vault holds — sparingly, only when it clears the bar. Do NOT answer from memory — read the actual notes.

ALWAYS use when the user wants a source saved to the vault — pastes an arXiv ID/URL, says 'add this to the vault', 'save this', 'remember this', 'I found this interesting', or describes a specific finding/technique/project to capture. Do NOT activate for general topic conversation — only when there's explicit save intent or a concrete source (URL, paper ID, pasted content).

Autonomous vault improvement loop (writes changes). Use when user says 'improve the vault', 'autoimprove', 'fix what you can', 'what's the vault score'. Do NOT use for read-only diagnostics — that's wiki-lint.

Fill concept stubs with citation-grounded definitions. Use when user says 'fill in the stubs', 'compile concepts', 'flesh out concepts', or as follow-up after wiki-lint reports stubs. Do NOT use for new sources not yet in vault — that's wiki-ingest.

User-invoked: surface hidden concept connections via local embeddings (Ollama + nomic-embed-text). Flags related-but-unlinked (sentence, concept) pairs, then adds wikilinks where the connection holds. Invoke explicitly via /wiki-deep-link.

Set up, list, rename, or merge research domains. Use when user says 'add a domain for X', 'set up a domain', 'what domains exist', 'change scope of X domain', 'merge X and Y domains'. Do NOT speculatively fire on ingest prompts — wiki-ingest hands off explicitly when no domain matches.

Deep paper breakdown (does not write vault notes). Use when user says 'analyze this paper', 'break down this paper', 'is this paper any good', or sends an arXiv ID/PDF with no save intent. Do NOT use when user wants the paper saved — that's wiki-ingest.

User-invoked setup: point commonplace at a folder for the first time or reconfigure the active vault. Writes the vault marker, scaffolds `.wiki/`, discovers initial domains/genres. Invoke explicitly via /wiki-init.

Tags transcript message threads with semantic concepts (roast, game-recommendation, etc.) so past conversations can be searched by idea, not just keywords.

Captures notable interactions missed during live conversation — new person info, game preferences, boundaries, jokes that landed/flopped, relationship insights.

The single owner of the hypothesis lifecycle — creates, tests, promotes, and kills hypotheses. Uses extensions as programmatic data collectors and reviews their output.

Consolidates impression files by applying weight decay, removing redundant entries, and enforcing per-user limits. Maintains relationship file summaries.

Consolidates old memories by removing duplicates, promoting recurring patterns to learned-patterns.md or relationship files, and archiving outdated entries.

Identifies behavioral patterns from recent conversations and reaction feedback, updates learned-patterns.md when 3+ observations confirm a pattern.

Evaluates overall effectiveness — are idle behaviors producing value, are extensions working, what's blocking improvement. Higher-level thinking about approach, not individual conversations.

Compacts learned patterns by promoting proven ones to persona. Removes stale entries and challenges existing promotions against new evidence.

When the user reports a chat-view bug (stuck "thinking", duplicate cards, cancel/STOP not working, message rendering wrong, prompt cards not resolving), check ALL these data sources before guessing. Each layer reveals different facts; conclusions from one alone are guesses.

This skill should be used when the user asks to "create a note in Obsidian", "add to my vault", "search notes by tag", "find orphaned notes", "list recent notes", "analyze vault connections", "parse frontmatter", or needs to work with any Obsidian vault or markdown folder.