THE ICP CARD · who really buys, from the deals you won
WAKES WHEN someone asks who the ideal customer is, which accounts to call first, or whether one fits
THE LOOP
icp.py new · accounts and the won deals
profile · from the won deals, with weights
score · every account, tiered, reasoned
A tier gets the call; C tier a question
re-profile each quarter as deals close
NEVER
Refuse an account on a score alone
Call a profile from five deals a profile
Score a new segment low and call it evidence
IT MAKES new — the account sheet (example rows) and won.csv beside it; profile — the profile from won deals; score — every account scored and tiered, reasons shown.
THE STANDARD from won deals, not from wishes; who bought is not who should have; weights are arithmetic; under ten won deals is a sketch; the score orders the list; it refuses nobody; re-profile each quarter.
PROOF SELFTEST 11/11 · witnessed in Claude 2026-10-09 · Edition 1.
THE SEAL 98 of 100 · LEGENDARY · WITNESSED · Trigger 5 · Machinery 5 · Law 5 · Portability 5 · Proof 5
IT CAN BE WRONG an AI reading a text file: it can be wrong and can invent a number, a rule, a name or a date; check what matters against your own records and the source; the risk is yours. THE-CLAUSES.md rides in the packet and is part of the licence.
THE PUNCH · SHA-256 OF SKILL.MD · D94A4EE8319E7BE19F402564
CITE THIS CARDTHE ICP CARD, Edition 1. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=icp-card. Punch D94A4EE8319E7BE19F402564. Retrieved 2026-10-10.
What it is.
Draws the ideal customer profile from the deals a team actually won — the values that cover most won deals in each field, the size band from the middle half, with weights from how concentrated the wins were — and scores every account in a list against it, 0 to 100, tiered A, B and C, with the reason for every point shown and unknowns named; and says plainly when the profile is a sketch.
It carries a machine that shows its working: new — the account sheet (example rows) and won.csv beside it; profile — the profile from won deals; score — every account scored and tiered, reasons shown — 11 checks in its selftest, every command on the record. It will never refuse an account on a score alone; never call a profile from five deals a profile; never score a new segment low and call it evidence.
Six laws, written into the card.
They are in the card itself, so the AI follows them.
FROM WON DEALS, NOT FROM WISHESThe profile is the values that cover most wins and the middle half of their size.
WHO BOUGHT IS NOT WHO SHOULD HAVEA segment never sold to scores low by construction; the card says so on every run.
WEIGHTS ARE ARITHMETICWeights are arithmetic, from how concentrated the wins were on each field; a starting point for the sellers, not a verdict.
UNDER TEN WON DEALS IS A SKETCHUnder ten won deals is a sketch, and the machine labels it one.
THE SCORE ORDERS THE LIST; IT REFUSES NOBODYA tier gets the first call; C tier gets a human question.
RE-PROFILE EACH QUARTERThe customer changes as the product does.
What rides in the card.
In the card’s own voice, with a machine that shows its working.
THE GUIDEreferences/WHO-BOUGHT-IS-NOT-WHO-SHOULD.md
THE MACHINEicp.py · three commandsnew — the account sheet (example rows) and won.csv beside it; profile — the profile from won deals; score — every account scored and tiered, reasons shown. Every command prints its working and refuses a bad input; 11 checks in its selftest, every command on the record.
What it does once your AI has it.
new The account sheet (example rows) and won.csv beside it
profile The profile from won deals
score Every account scored and tiered, reasons shown
On the record Eight example won deals drawn into a profile (software, help-desk X, 75 to 150 employees — labelled a sketch at eight deals); five accounts scored against it, the look-alike tier A, the account unlike any win tier C, every reason shown.
What it will never do.
It will never refuse an account on a score alone; never call a profile from five deals a profile; never score a new segment low and call it evidence.
A profile from fewer than ten won deals is a sketch and the machine says so. The profile describes who bought, not who should have: a new segment never sold to scores low by construction. The weights are arithmetic on the sheet, not judgement. It is an AI reading a text file: it can be wrong, so check what matters against your own records and the source.
What’s in the buy.
CARD.md — the whole card: who it is, its laws, when it says no, what it carries · SKILL.md — the door your AI reads first · START-HERE.md — the first five minutes, for you
WHO-BOUGHT-IS-NOT-WHO-SHOULD.md — the guide in references/, in the card's own voice
scripts/icp.py — the machine: new · profile · score, standard library only, selftest 11/11
WITNESS.md + SELFTEST.md — every command on the record, as it printed · MANIFEST.md, every file hashed · AUDIT.md, the twelve checks
MY-LEDGER.md — the card's memory, on your own machine · TIN.md · LICENSE-NOTE.md · THE-CLAUSES.md · HOSTS.md
card.svg + card-machine.svg + card.json · card.html, the card as one offline file · the doors for ChatGPT, Gemini, Cursor, Claude Code and MCP hosts (rules/, system/, mcp/)
PROOF · SELFTEST 11/11 · 3 COMMANDS ON THE RECORD · WITNESSED IN CLAUDE 2026-10-09 · EDITION 1 · PUNCH D94A4EE8319E7BE19F402564
ON EVERY DOOR · A CARD FOR YOUR AI, A FILE FOR YOUR COMPUTER
In Claude: turn on code execution under Settings › Capabilities, then Customize › Skills › + › Create skill, and upload the zip as it is — START-HERE.md in the box walks you through it, and tells you how to keep the ledger. In ChatGPT, Gemini and the rest: a Project, a GPT or a Gem with the card’s files attached and the short door in the box pasted as its instructions. The words run anywhere an AI reads Markdown; the machines run wherever the host runs Python. Witnessed in Claude only. The host marks →
Take it.
Nº 849 PROPOSED — LAUNCHING SOON
Sold as is under the term printed. A card instructs an AI you operate; results depend on your model, your files and your judgement. Not medical, legal, tax or financial advice, never the person’s boss, and not a licence to practise. The terms →
One payment. The file is yours to keep and keeps working at its edition; new editions issued in the year after you buy reach you by email as replacement files. Tax is added at checkout where it applies. Nothing auto-renews.
A TERM, NOT A SUBSCRIPTION: NOTHING RENEWS ITSELF AND NOBODY IS CHARGED AGAIN. WHEN A TERM ENDS THE FILES STAY ON YOUR COMPUTER AND STAY READABLE — FREEZE, NOT LOSS.
THE MACHINE · IN THE OPEN · FROM THE WITNESSED RUN OF 2026-10-09
icp.py, every command showing its working.
A small Python program, standard library only. Your AI runs it; you can too. Every block below is pasted from the card’s own WITNESS.md, as printed.
new
the account sheet (example rows) and won.csv beside it
profile
the profile from won deals
score
every account scored and tiered, reasons shown
selftest
11 checks that prove the machine; it must print 11/11
new · THE ACCOUNT SHEET (EXAMPLE ROWS) AND WON.CSV BESIDE IT
$ python3 scripts/icp.py new accounts.csv
sheets written · accounts.csv (5 example accounts) · won.csv beside it (8 example won deals) · all invented: replace them with your own
profile · THE PROFILE FROM WON DEALS
$ python3 scripts/icp.py profile won.csv --out profile.json
PROFILE · from 8 won deals · a sketch: only 8 won deals
industry software · covers 75% of won deals · weight 1.25
region north america · covers 75% of won deals · weight 1.25
tech helpdesk-x · covers 88% of won deals · weight 1.38
trigger new support lead, hiring support agents · covers 75% of won deals · weight 1.00
segment mid-market · covers 75% of won deals · weight 1.25
employees 75–150 (the middle half of won deals)
NOTE · who bought, not who should have: a segment never sold to scores low by construction
profile written · profile.json
score · EVERY ACCOUNT SCORED AND TIERED, REASONS SHOWN
$ python3 scripts/icp.py score accounts.csv --profile profile.json --report ranking.md
ICP SCORE · 5 accounts · profile from 8 won deals · A 3 · B 0 · C 2
A 100 Harbor (example) industry software ✓ · region north america ✓ · tech helpdesk-x ✓ · trigger new support lead ✓ · segment mid-market ✓ · employees 120 ✓ (75–150)
A 75 Atlas (example) industry software ✓ · region north america ✓ · tech helpdesk-x ✓ · trigger hiring support agents ✓ · segment small ✗ (won: mid-market) · employees 40 ~ (near 75–150)
A 75 Summit (example) industry fintech ✗ (won: software) · region north america ✓ · tech helpdesk-x ✓ · trigger new support lead ✓ · segment mid-market ✓ · employees 300 ~ (near 75–150)
C 0 Pioneer (example) industry retail ✗ (won: software) · region europe ✗ (won: north america) · tech helpdesk-y ✗ (won: helpdesk-x) · trigger unknown · segment enterprise ✗ (won: mid-market) · employees 2500 ✗ (75–150)
C 0 Meadow (example) industry manufacturing ✗ (won: software) · region asia ✗ (won: north america) · tech none ✗ (won: helpdesk-x) · trigger unknown · segment enterprise ✗ (won: mid-market) · employees 900 ✗ (75–150)
report written · ranking.md
THE NUMBERS, NAMES AND DATES IN THESE RUNS ARE THE STUDIO’S OWN EXAMPLES, INVENTED FOR THE RECORD; NO REAL COMPANY, PERSON OR PROPERTY IS NAMED.
QUESTIONS AT THE TABLE
Asked before the wire.
What does it need from me?
Icp.py new · accounts and the won deals. It asks first for the won deals with industry, size, region, tech, trigger and segment, and the account list to score.
Does it send anything anywhere?
No. The machine is a standard-library Python program that runs on your own computer and sends nothing anywhere; the words run wherever an AI reads Markdown. No server, no account, no key.
What is on the record?
The self-test (11/11) and every command run on real example inputs, pasted into WITNESS.md as they printed; the punch on the face is the SHA-256 of SKILL.md, so a downloaded copy can be checked against this room.
Can it be wrong?
Yes — it is an AI reading a text file. A profile from fewer than ten won deals is a sketch and the machine says so. The profile describes who bought, not who should have: a new segment never sold to scores low by construction. The weights are arithmetic on the sheet, not judgement.
What will it never do?
It will never refuse an account on a score alone; never call a profile from five deals a profile; never score a new segment low and call it evidence.
