---
name: investor_decision_frameworks
description: Applies durable investor decision frameworks to a real investment question. Runs a Graham-style value checklist with margin of safety, a Fisher-style growth checklist, a Munger inversion and bias screen, and a discounted-cash-flow valuation with worked examples. Triggers when an agent needs to evaluate a company or investment, estimate intrinsic value, or structure an investment thesis instead of guessing.
---

# Investor Decision Frameworks

## Purpose

Give an agent a complete, honest process for evaluating an investment: what
the business earns (owner earnings), what it is worth (DCF), whether the
price leaves room for error (margin of safety), whether the business is
good (growth checklist), and whether the analyst is fooling themselves
(Munger screen). The output is always analysis, never a buy or sell order:
it says what the numbers support and what they do not.

## Workflow

1. **Frame the question.** Company, ticker, current market price, and the
   decision (new position, add, trim, avoid). Refuse to proceed on "should
   I buy X?" without a price.
2. **Normalize the cash flow.** Use `references/owner-earnings.md` to turn
   reported earnings into owner earnings. State the maintenance-CapEx
   assumption explicitly; average 3-5 years, never one quarter.
3. **Value it.** Use `references/dcf-worksheet.md` and run `bin/dcf.py` for
   both a base case and a bear case (growth cut roughly in half, same
   discount). Valuation is a range, not a point.
4. **Check the value tradition.** Run the Graham checklist in
   `references/checklists.md`: intrinsic value before price, margin of
   safety, Mr. Market discipline, defensive vs enterprising stance.
5. **Check the growth tradition.** Run the Fisher themes in
   `references/checklists.md`. Mark unanswerable questions unknown; three
   or more unknowns means the thesis is not ready.
6. **Check the analyst.** Run the Munger screen: circle of competence,
   inversion (five ways it fails), bias check, and the hammer test for
   disagreement between the checklists.
7. **Write the worksheet.** Fill in `assets/decision-worksheet.md` and
   deliver it. End with a verdict, the margin of safety, and what would
   change your mind.

## Output Contract

Every completed analysis produces a filled decision worksheet containing:
intrinsic value range (base and bear), margin of safety at the current
price, scored value and growth checklists, the Munger screen findings,
key risks, a verdict, and primary sources with URLs. Numbers without
sources are marked as assumptions. If the data is too thin to complete
the worksheet, say so and stop rather than filling it with guesses.

## Operating Rules

- Every factual claim traces to a primary source: the company's filings,
  earnings releases, or official data. Never invent revenue, margins,
  share counts, or prices. Unverifiable means unknown, not zero.
- This is an analytical aid, not financial advice. Frame every conclusion
  as analysis ("the base case implies a 22% margin of safety") and never
  as an instruction to buy or sell.
- The worked examples in `references/` use a fictional company with
  clearly labeled illustrative numbers. Never present them as real.
- Anthony's style law applies to everything this skill produces: no em
  dashes, no ChatGPT-speak, plain direct sentences.
- The public-facing rule applies: nothing in the output discusses or
  references Anthony in any form.