The Seed skill

11 Claude skills that take your startup from idea to funded: competitor research, investor matching, a business plan, a pitch deck and a closer that books your investor calls.

Free, MIT, no signup, no API key. The tools are plain Python that read free public data (the App Store’s own search and review feeds, SEC EDGAR) and do the arithmetic. Everything they write lands in a seed/ folder in your project, so every skill picks up where the last one stopped.

One finds the weak points of your competitors, straight from their own one-star reviews. Another researches the best investors for your idea, from public sources, and ranks them by fit with the reasons and the links. Then one turns your idea into a business plan, and another turns the plan into a pitch deck.

And the craziest one is the Closer. It writes a personalised outreach email to every investor on your list, keeps track of every conversation, sorts the replies, and follows up until an investor call is booked, with the calendar invite ready to send. So all you have to do is show up.

Nothing leaves your inbox until you say yes. The Closer drafts. Claude sends through your own Gmail connector only after you approve the batch, with a daily cap, an opt-out line in every email, and a thread that stops for good on a “no”.

Install

Paste this into Claude:

https://github.com/Jakeschincariol/seed-skill

install skill

Or as a plugin, in Claude Code:

/plugin marketplace add Jakeschincariol/seed-skill
/plugin install seed-skill@seed-skill

Claude Code namespaces plugin skills, so installed as a plugin they show up as /seed-skill:seed-scout and so on. Copy the folders instead if you want plain /seed-scout:

git clone https://github.com/Jakeschincariol/seed-skill.git
cp -r seed-skill/seed-* ~/.claude/skills/

Project-local instead of global: copy the same folders into your repo’s .claude/skills/. No Claude Code at all? Paste any single SKILL.md at the top of a chat and it runs as a mode. You lose the Python tools, but the method works.

The tools need Python 3.8 or newer. Nothing to pip install. The investor finder’s SEC search asks you to identify yourself, as the SEC requires: put your name and email in seed/config.json ("sec_user_agent": "Your Name [email protected]") or in the SEC_USER_AGENT environment variable.

To send email and book calls, connect Gmail and Google Calendar in Claude (Customize, Connectors) and set their send tools to “Needs approval”.

The eleven

How to use it

Everything lands in a seed/ folder in your project, and each skill reads what the last one wrote:

scout -> market -> model -> plan -> deck            (the story and the numbers)
investors -> closer -> coach                         (the raise: who, the outreach, the call)
terms -> dataroom                                    (when an investor says yes)
updates                                              (every month, before and after)

Say what you want in plain words (“find my competitors’ weak points”, “who should I pitch”, “write the deck”, “follow up with everyone who hasn’t replied”) and Claude picks the skill.

Example: the reel

The startup in the reel is made up: Leashly, “a dog-walking app with flat prices and a backup walker on every walk”, raising a $1.5M seed. The competitors and the market numbers are real.

1. /seed-scout read the 500 most recent App Store reviews of Rover and of Wag! on October 25, 2026. Rover (4.9 stars, 515,771 ratings): 106 of its 341 one- and two-star reviews mention fees. Wag! (4.8 stars, 67,447 ratings): 81 of its 324 mention cancellations. Those are the weak points: fees, and walkers who cancel at the last minute.
2. /seed-market sized it from APPA’s 2026 report: $158 billion of US pet spending in 2025, and 71 million US homes (53%) with a dog.
3. /seed-model ran 36 months at $25 a walk and a 15% take rate, with the $1.5M raise (the example assumptions in seed-model/): $467,400 of revenue in year 1 and $2,993,520 in year 3, $2.23 of contribution per walk, LTV:CAC of 4.24x, peak burn $82,801 in month 5, break-even in month 29, and cash that never runs out (its low point is $252,553).
4. /seed-plan and /seed-deck turned that into the nine-section plan and the 10-slide deck that ends on “$1.5M seed to launch in 3 cities”.
5. /seed-investors ranked the investors who back pet and local-services marketplaces at seed (the reel’s names, like Maya Chen of Northlight Ventures, are fictional too).
6. /seed-closer drafted one email per investor, each with its own first line (“Saw you led two pet-care seed rounds…”). After the yes, it followed up on day 3 and day 7 of the ones that went quiet, sorted the replies, answered Maya’s take-rate question from the plan, and when she wrote “Thursday 10am works”, wrote the invite: Leashly x Northlight Ventures, Thursday November 5, 10:00 AM ET.
7. /seed-coach prepared the 20-minute call, and /seed-terms and /seed-dataroom are there for what comes after it.

Fine print

No promise of funding. This is software that drafts, researches and does arithmetic. Investors decide whether to invest, based on your business. A better deck and a faster follow-up help; they do not guarantee a check.

Nothing sends without your yes. Only /seed-closer can lead to an email, and only after you approve that batch in the chat. Set your Gmail connector’s send tool to “Needs approval” (Customize, Connectors, tool permissions), never “Always allow”, so Claude’s own “Allow once / Deny” card is a second check. Every email carries your real name and an opt-out line. A “no”, an opt-out or a bounce ends that thread for good. Default cap: 20 new emails a day.

Public sources only. The investor finder reads public pages (a firm’s own site, public profiles found through search, SEC EDGAR, the press) and keeps the link for every fact. It never logs in anywhere, never scrapes behind a login and never scrapes LinkedIn. Check a fact before you lean on it in an email.

Not legal advice. /seed-terms explains the common clauses and does the dilution math. Have a startup lawyer read anything before you sign it.

Not investment advice. The model and the market size are only as good as the assumptions you give them. They are there to make your thinking visible to you and to an investor, not to predict the future.

The data is free and public. The App Store’s search and review feeds are not an official API and can change or rate-limit. SEC EDGAR asks every automated request to say who is asking.