seed-market — Claude Code skill for market sizing
seed-market — a skill for Claude Code estimates TAM, SAM, and SOM. It activates on market-sizing requests or after /seed-scout, before /seed-model and /seed-plan.
MIT ★ 197 updated 2026-11-01
/plugin marketplace add Jakeschincariol/seed-skill
/plugin install seed-skill@seed-skillWhat seed-market does
The skill calculates TAM from population, share, and annual spend; SAM from reachable customers, purchase frequency, and price; and SOM as a reasoned share of SAM over a stated period. Every figure needs a source: a URL, a named publication, or an assumption with its rationale. Vague labels such as “google” and “TBD” are rejected.
The market.py script writes results to seed/market.json, which the plan, deck, and pitch coach read. The check command validates sources and the SOM ≤ SAM ≤ TAM order. Writes that break this order are refused. It requires python3.
What it helps with
- estimate total addressable market from population data
- calculate reachable market from customers, purchase frequency, and price
- justify a SOM share with a timeframe and rationale
- keep sourced market facts for a pitch deck
Example prompts for the agent
- Estimate this product’s market as TAM, SAM, and SOM.
- Calculate SAM from reachable customers, purchase frequency, and price.
- Check the sources and the TAM, SAM, SOM order.
How to install seed-market
1Plugin marketplace
/plugin marketplace add Jakeschincariol/seed-skill/plugin install seed-skill@seed-skillThis installs the whole seed-skill plugin with all its skills
2Skills CLI (Vercel)
npx skills add Jakeschincariol/seed-skill --skill seed-market -a claude-code3Manually: put the files in place
~/.claude/skills/seed-market/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/seed-market/
1Skills CLI (Vercel)
npx skills add Jakeschincariol/seed-skill --skill seed-market -a codex2Manually: put the files in place
.agents/skills/seed-market/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add Jakeschincariol/seed-skill --skill seed-market -a cursor2Manually: put the files in place
.cursor/skills/seed-market/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install Jakeschincariol/seed-skill seed-market2Manually: put the files in place
.github/skills/seed-market/SKILL.md1Skills CLI (Vercel)
npx skills add Jakeschincariol/seed-skill --skill seed-market -a gemini-cli2Platform CLI
gemini skills install https://github.com/Jakeschincariol/seed-skill --path seed-marketManually: put the files in place
.devin/skills/seed-market/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about seed-market
How do I install seed-market in Claude Code?
Add the marketplace with /plugin marketplace add Jakeschincariol/seed-skill, then install the plugin with /plugin install seed-skill@seed-skill.
What does seed-market do?
It calculates TAM, SAM, and SOM, records the sources, and saves the results in seed/market.json for the plan, deck, and pitch coach.
What does seed-market need to run?
It requires python3. Each input figure also needs a URL, a publication name, or a reasoned assumption.
Source text
SKILL.md · MIT
---
name: seed-market
description: >-
Market Sizer. Sizes your market the way investors check it: TAM top-down
(population x share x spend a year), SAM bottom-up (customers you can
reach x purchases a year x price) and SOM (the share of SAM you can win in
N years, with a stated reason). Every input carries a source (a URL, a
named publication, or an assumption said out loud), vague sources are
refused, and it refuses any result that breaks SOM <= SAM <= TAM. Records
sourced facts the deck quotes. Writes seed/market.json for the plan, the
deck and the pitch coach. Use when the user says "how big is my market",
"TAM SAM SOM", "size the market", "market size slide", "is this market
big enough", or after /seed-scout, before /seed-model and /seed-plan.
argument-hint: "[top-down | bottom-up | som | cite | show | check]"
---
# seed-market
The Market Sizer turns "a huge market" into three numbers an investor can
check. The total market (TAM) is everyone who could buy, top-down from
public data. The market you can serve (SAM) is built bottom-up from the
customers you can actually reach, how often they buy and your price. The
share you can win (SOM) is a slice of SAM over a few years, with the reason
written next to it. Every input says where it came from, and the guesses are
labelled as guesses so nobody mistakes them for facts.
`${CLAUDE_SKILL_DIR}` below is this skill's folder (where this SKILL.md lives).
```bash
python3 "${CLAUDE_SKILL_DIR}/market.py" top-down --population 71M --population-name "US homes with a dog" \
--population-source "APPA 2026 State of the Industry report (americanpetproducts.org)" \
--share 10% --share-source "assumption: about 1 in 10 dog homes pays for walks; check with a survey" \
--spend 2400 --spend-source "assumption: 8 walks a month at the \$25 price" --write
python3 "${CLAUDE_SKILL_DIR}/market.py" bottom-up --customers 400K --customers-source "..." \
--frequency 96 --frequency-source "..." --price 25 --price-source "..." --unit walk --write
python3 "${CLAUDE_SKILL_DIR}/market.py" som --share 2.5% --years 3 --reason "who switches, and why" --write
python3 "${CLAUDE_SKILL_DIR}/market.py" cite --name "US pet industry spend, 2025" --value 158B --source "APPA ..." --write
python3 "${CLAUDE_SKILL_DIR}/market.py" show
python3 "${CLAUDE_SKILL_DIR}/market.py" check # exit 1 if anything is out of order or unsourced
```
Other flags: `--level tam|sam` (either method can fill either level),
`--population-name`, `--share-name`, `--spend-name`, `--customers-name`,
`--currency`, `--dir`, `--json`, `--today`.
## What it reads and writes
| file | what |
| --- | --- |
| your sources | public reports, government data, your own pricing; or an assumption with its reason |
| `seed/market.json` | with `--write`: `tam`, `sam`, `som`, `currency`, `method`, the formulas, `inputs` (name, value, unit, source, kind, level), the SOM share, years and reason. `/seed-plan`, `/seed-deck` and `/seed-coach` read it |
## The method and the rules
- **TAM, top-down:** population x share x spend a year. **SAM, bottom-up:**
customers x purchases a year x price. **SOM:** a share of SAM within N
years (1 to 10). All yearly, exact (Decimal), rounded half up to the dollar.
- **SOM <= SAM <= TAM.** A write that breaks it is refused (exit 1) and the
file stays as it was. When SAM changes, SOM is recomputed from its share.
- **A source on every input.** A URL, a named publication, or
"assumption: <why>". "google", "TBD", "ChatGPT", a single word or an
assumption without its reason are refused. Assumptions are listed every
time the numbers print: say so when you quote them.
- **A reason on every SOM**, a real sentence: who switches, why, how you
reach them. Over 10% of SAM gets a note: be ready to defend it.
- **Facts are not math.** `cite` keeps a sourced number (an industry total)
that the plan and deck quote, so every number on a slide has a home.
- Estimates from your inputs, not a forecast. Not investment advice.
## Steps
1. Find the real numbers first: a trade association report, census or
government data, a public filing. Open the page and keep its name or URL.
Never use a number you could not open.
2. Run `top-down ... --write` for the TAM with those numbers. Where there is
no public number, write an assumption with its reason, and tell the user.
3. Run `bottom-up ... --write` for the SAM from the customers you can reach
at launch, how often they buy (match `/seed-model`) and your price.
4. Run `som --share X% --years N --reason "..." --write`. Keep it in line
with the model's customers in year N.
5. `cite` the headline facts the deck will show. Run `check` (it must exit
0) and `show`, then read the three numbers and every assumption back.
6. Hand off to `/seed-model` (the 36-month model) and `/seed-plan`, which
puts the Market section together from `seed/market.json`.
## Example: the reel
Leashly (made up) is a US dog-walking app at $25 a walk. The real numbers
are APPA's: 71 million US homes (53%) have a dog, and Americans spent $158
billion on pets in 2025 (APPA 2026 State of the Industry report). The rest
are Leashly's assumptions, labelled as such:
```
Market size (USD a year), top-down TAM, bottom-up SAM, SOM 2.5% of SAM in 3 years
TAM $17.04B 71,000,000 US homes with a dog x 10% x $2,400 = $17,040,000,000
SAM $960M 400,000 paying dog homes in the 3 launch cities x 96 walks a year x $25 = $960,000,000
SOM $24M 2.5% of SAM $960,000,000 = $24,000,000
Fact: US pet industry spend, 2025 = $158B (APPA 2026 State of the Industry report (americanpetproducts.org))
Fact: Share of US homes with a dog = 53% of US homes (APPA 2026 State of the Industry report (americanpetproducts.org))
```
The 10% share, the $2,400 a year (8 walks a month at $25) and the 400,000
launch-city homes are assumptions to replace with real data. The SOM's
reason: owners leaving Rover over fees and Wag! over cancellations switch to
flat $25 walks with a backup walker, 10,000 of the 400,000 homes by year 3.
## What it never does
It never makes up a statistic, never passes off an assumption as a fact,
never inflates a share to make the slide bigger, and never promises that the
market will turn into revenue.
© Jake Schincariol · MIT · Jakeschincariol/seed-skill