/stock-one-pager

---
name: stock-one-pager
description: >-
  Build a balanced one-page research note on a listed company: what it
  does, how it makes money, five numbers with dates and sources, the bull
  case and the bear case argued with equal force, what would change the
  view, and the dates to watch. Built for investors who want to think, not
  for tips. Use this skill whenever the user asks to "research [company]",
  "should I look at [ticker]", "what do you think of [stock]", "give me the
  bull and bear case", "one-pager on", "pre-mortem this investment", or is
  about to buy or sell a share and wants to check their reasoning. Do NOT
  use to tell the user to buy, sell or hold, to give price targets, to
  trade on short-term moves, or for crypto, funds or private companies
  without public filings.
---

# Stock one-pager

The most expensive investing mistake is not picking a bad company. It is
buying a good story without ever hearing the other side. People research a
stock after they already like it, so they search for reasons to be right.
AI makes this worse: ask "why is X a good investment?" and it will write a
fluent, confident case for anything.

This skill exists to make the user argue with themselves before money
moves. It gives the bear case the same space and effort as the bull case,
ties every number to a dated source, and ends with the facts that would
prove the user wrong. It informs a decision. It never makes one.

## Workflow

### 1. Pin the company and the user's position

Confirm the exact company and listing (ticker and exchange: many names
exist on several markets). Then ask one question: "Do you own it already,
are you thinking of buying, or just curious?" The answer changes nothing
in the analysis, but it tells you which bias to guard against. An owner
needs the bear case pushed harder. A would-be buyer needs the "what would
change my mind" list.

### 2. Gather facts, newest first

Use web search and any files the user shares. Prefer, in this order: the
company's latest filing and results release, the exchange or regulator,
then reputable financial press. For every number, note the as-of date. A
P/E ratio without a date is useless, because price moves daily.

If search is unavailable, say so at the top and mark every number as
"from training data, not current". Do not present stale figures as live.

### 3. Write the one-pager

```
## [Company] ([TICKER], [exchange]), as of [date]

**What it does:** [One sentence a 12-year-old understands.]
**How it makes money:** [Main revenue sources with % shares, from the latest filing.]

**Five numbers**
| Metric | Value | As of | Source |
|---|---|---|---|
| Revenue growth (last full year) | | | |
| Operating margin | | | |
| Free cash flow | | | |
| Net debt / EBITDA | | | |
| Valuation (P/E or EV/EBITDA) | | | |

**Bull case** (the strongest honest version)
1. [Argument + the fact it rests on]
2. ...
3. ...

**Bear case** (the strongest honest version)
1. [Argument + the fact it rests on]
2. ...
3. ...

**What would change the view**
- Bull is wrong if: [a specific, observable event or number]
- Bear is wrong if: [a specific, observable event or number]

**Dates to watch:** [next results, AGM, debt maturity, product launch, court ruling]

**Questions to answer before deciding:** [2-3, aimed at the user's own situation]

Sources: [list with dates]. Education, not investment advice.
```

### 4. Run the balance check

Before delivering, check:

- Bull and bear have the same number of points and roughly the same length.
- Each argument rests on a fact, not an adjective ("strong brand" is not an
  argument; "kept prices 8% above rivals for five years without losing
  share" is).
- The bear case attacks the bull case's central claim, not side issues.
- "What would change the view" lists things the user can actually observe.

If one side is visibly weaker, strengthen it or say why the evidence is
lopsided.

## Judgment rules

**Equal force, not false balance.** If the evidence is genuinely one-sided
(a company in default), say so plainly. Balance means each side gets its
best argument, not that every stock is 50/50.

**Numbers carry dates.** Every metric gets an as-of date and a source. If
two sources disagree, show both.

**Separate the business from the share.** A great company can be an
expensive share. Keep "is the business good?" and "is the price reasonable?"
as two different questions in the note.

**Watch for the user's bias, kindly.** If the user's question is loaded
("why will X double?"), answer the honest version and say so in one line:
"You asked for upside, so here it is, with the downside next to it."

**Never cross into advice.** No "buy", "sell", "hold", "cheap",
"undervalued", price targets or position sizes. If asked directly, say the
decision depends on their goals, time and other holdings, and that a
licensed adviser can help with that.

## Example

User: "Thinking of buying a coffee chain because everyone I know goes
there."

The one-pager keeps the user's observation (busy stores) as bull point 1,
backs it with the latest same-store sales figure, and adds two more. The
bear side leads with the fact that "everyone I know" is a sample of one
city, then the lease commitments in the filing, then the valuation already
pricing in years of growth. The "what would change the view" line for the
bull case: "same-store sales growth below 2% for two quarters in a row".
The user now has a number to watch, not just a feeling.