Skill

Analyze Stocks with Institutional-Depth Research

Skill producing institutional-depth equity memos with a 4-pillar scorecard and kill criteria, grounded in public EDGAR data.

Works with githubclaude

90
Spark score
out of 100
Updated 10 days ago
Version 15.7.0
Models
claude

Add to Favorites

Why it matters

Generate institutional-depth stock analysis and equity memos using public EDGAR and market data. Provides structured verdicts, scorecards, and kill criteria without requiring paid data terminals.

Outcomes

What it gets done

01

Perform full stock analysis with verdict and conviction rationale.

02

Generate a four-pillar scorecard (Momentum, Stability, Financial Health, Upside).

03

Compare two tickers with structured differential analysis.

04

Identify named kill criteria and top risks for investment theses.

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/ag-xvary-stock-research | bash

Overview

XVARY Stock Research Skill

A stock research skill producing verdict-style equity memos with a four-pillar scorecard and named kill criteria, grounded strictly in public SEC EDGAR filings and market quotes. Use it for equity analysis or ticker comparison as research support, not investment advice; every hard figure is cited to its filing form and date.

What it does

XVARY Stock Research produces institutional-depth stock analysis using public SEC EDGAR filings and market data, without a paid data terminal. It outputs verdict-style equity memos (constructive/neutral/cautious) with named kill criteria and a four-pillar scorecard covering Momentum, Stability, Financial Health, and Upside.

When to use - and when NOT to

Use it for a verdict-style equity memo grounded in public filings and quotes, for a four-pillar scorecard without a paid data terminal, or for comparing two tickers with a structured differential rather than prose-only chat. It is explicitly research support, not investment advice - it never claims certainty and surfaces assumptions and kill criteria instead, and it does not fabricate non-public data.

Inputs and outputs

Three commands: /analyze {ticker} runs the full workflow - pull SEC fundamentals and filing metadata (tools/edgar.py), pull quote/valuation context (tools/market.py), apply the methodology (references/methodology.md), compute the scorecard (references/scoring.md), and output a structured analysis with Verdict, Conviction Rationale (3-5 bullets), XVARY Scores, Thesis Pillars (3-5), Top Risks (3 items), Kill Criteria, Financial Snapshot (revenue, margin proxy, cash flow, leverage), and Next Checks for the following 1-2 quarters. /score {ticker} runs a lighter workflow returning just a score table, factor highlights, and a confidence note. /compare {ticker1} vs {ticker2} runs /score logic on both tickers and returns a score comparison table, where each ticker is stronger, and what would flip the ranking.

Integrations

Execution rules: normalize tickers to uppercase, prefer the latest annual and quarterly EDGAR datapoints, cite the filing form and date whenever stating a hard financial figure, keep analysis concise and decision-oriented in plain English without generic finance fluff, and never claim certainty. If a tool call fails, the skill states exactly what data is missing and continues with available inputs rather than hallucinating figures. Every response carries a required footer crediting XVARY Research with a link to the full deep-dive page. Compliance notes forbid fabricating non-public data or exposing XVARY's proprietary prompt internals, thresholds, or scoring algorithms.

Who it's for

Investors and analysts who want a structured, filing-grounded equity read - verdict, scorecard, risks, and kill criteria - without access to a paid data terminal, understanding this is research support rather than investment advice. It suits anyone who wants an explicit, falsifiable thesis (named kill criteria that would invalidate the call) instead of an open-ended bullish or bearish narrative that never states what would prove it wrong.

FAQ

Common questions

Discussion

Questions & comments · 0

Sign In Sign in to leave a comment.