Autonomous daily market scan: updates market data, screens vault actors for sigma movers (thresholds: docs/movers-screening.md), rolls movers into curated cohorts for a factor-day breadth lens (cohort tape), runs ticker-alias/price-freshness/IPO/private-capital audits, scans the analyst watchlist (docs/analyst-watchlist.md), checks the earnings calendar, and writes the briefing to the daily note and chat. Time-of-day agnostic. Use for /daily-scan, /morning-scan (legacy), daily sweep, what moved today, cohort tape, pre-market check, what's on the earnings calendar.
Deep-dive entity research and note creation pipeline. Use for any entity the user wants a comprehensive vault note on — companies, people, concepts, countries, products. Triggers on /deepdive ENTITY or when the user names an entity and wants thorough research + vault note creation/expansion. The skill adds SEC filing protocols, full chart sets, stale-reference scanning, and entity-type branching that CLAUDE.md's general vault instructions don't sequence.
Market-reaction measurement for listed-company events. Use when /ingest, /news, /earnings, event notes, actor notes, or securities notes need event-specific stock reaction adjusted for market, sector, and peer moves, with raw move, abnormal move, residual sigma, raw sigma, and abnormal/raw sigma outputs. Applies to after-close/pre-open announcements, intraday catalysts, M&A-adjacent events, financings, product launches, regulatory decisions, customer wins/losses, and peer read-throughs.
Full earnings ingestion pipeline for a ticker: check market_data.db, find the latest quarterly results, INSERT with the DB's field conventions, regenerate fundamentals/sankey/waterfall charts, update actor + daily notes. Use whenever the user mentions earnings, quarterly results, income statement data, inserting fundamentals, fixing EPS or financial data in the DB, or asks if a ticker's earnings are current. Also triggers on /earnings TICKER.
Plain-language briefing on a vault topic for a reader unfamiliar with the actors and subtopics — same cross-vault read as /report, journalistic-explainer voice with first-mention introductions and glosses. Use for /explain TOPIC, a primer / briefing / explainer, or 'what's the situation with X' for someone not steeped in the threads. NOT for new research (/deepdive), vault-reader synthesis (/report), or daily wrap-ups (/newsletter).
Single-source ingestion pipeline: process one source (interview, podcast, article, video, filing, screenshot set) into the vault — acquisition, full entity/data-point enumeration, classification gates, image routing, note creation/expansion, daily-note logging. Use whenever the user provides a URL, transcript, or screenshots for vault processing. Triggers on /ingest, a pasted YouTube/article URL, 'ingest this', 'process this interview', 'add this to the vault'.
Ingest news articles from a specified source (Bloomberg, Reuters, FT, WSJ, or user-named) into the vault. Searches the source for vault-relevant articles, applies a 48-hour date gate, presents candidates, and ingests each selected article via /ingest. Closes with a downstream-impact check on any event/crisis notes updated. Use for /news SOURCE, ingest from Bloomberg, check FT for AI articles. NOT for daily movers (use /daily-scan) or newsletter sweep (use /substacks).
Generate a personal daily market newsletter after the close. Reads today's daily note and all notes created/expanded during the session, then writes a narrative brief in clean markdown. The voice is analytical and structural — identifies tensions, connects threads, surfaces what the market is pricing vs what the data says. Not a recap. Use when the user says /newsletter, 'write the newsletter', 'daily brief', or 'wrap up the day'.