heurist-finance: Skill for Claude Code

Event-Driven

Shared plumbing and output contract: ../_shared/analyst-core.md. The governing fact — say it to the user when relevant: most tradable-looking news carries no edge by the time retail reads it. Liquid names price headlines in minutes. The events below are the exceptions, and each is an exception for a structural reason (forced flows, institutional constraints, slow diffusion) — when the reason doesn't apply, neither does the edge.

Per-event playbooks

M&A arbitrage. Edge test first: the spread compensates for break risk, and the pros price legal/regulatory risk with counsel and scenario trees. Retail's viable corner is small ($500M–$2B), simple, cash deals — base rate: announced cash deals close at ~95%+ (Mitchell & Pulvino 2001); the annualized spread on a clean one is the price of break risk, not free money. Workflow: deal terms from web + the merger 8-K via sec filing_timeline; compute spread, annualize over expected close; check financing condition, regulatory overlap, and go-shop. Skip: horizontal deals with antitrust tails, stock-for-stock (you're implicitly short the acquirer), anything where the spread looks generous — the market knows something. Position asymmetry is brutal (earn the spread, lose the pre-deal gap), so size accordingly. Sourcing: no tool enumerates pending deals — build the candidate set iteratively via web queries (e.g. "pending all-cash acquisitions under $2B"), then verify each candidate's terms in the merger 8-K/proxy via filing_timeline.

Spinoffs — the best structural edge on this list. Parent+spinco outperform benchmarks by ~10–20% over 12–24 months in the classic studies (Cornell/Qiao and predecessors) because the selling is forced: index funds and mandate-constrained holders must dump a spinco that doesn't fit (too small, wrong sector, no dividend), regardless of price. The dislocation window is roughly days 30–180 post-separation. Workflow: Form 10 via filing_timeline; who's forced to sell (parent's index membership vs spinco's); insider/management incentives in the spinco (equity grants struck post-spin are a tell); let the forced selling exhaust (volume decay + base formation on technical_snapshot) before entering. Garbage-barge spinoffs (debt-stuffed liability dumps) exist — read the Form 10 balance sheet, don't buy "it's a spinoff" as a thesis.