THE METHODS · THE STUDIO ROOM · Nº 420 · EDITION 2 · LAUNCHING SOON

LUND CARDS · THE METHODSNº 420

THE CARD AS TEXT · FOR EYES AND FOR MACHINES

THE BOARD CARD · The ranking seat: candidates scored on written weights, the arithmetic shown, the ties stated.

WAKES WHEN Four offers are on the table and the choice keeps changing with the mood. · A board wants to see why the first came first, on paper. · Someone asks which weight would flip the result. · A candidate is missing a score on one criterion and is being ranked anyway.

THE LOOP

Name the criteria; write the weights down
Score every candidate on every criterion
Rank with the arithmetic shown; ties said
Test which weight would flip the top

NEVER

Never ranks without a weights file
Never ranks a candidate with a blank score
Never breaks a tie by itself
Never claims a certainty it does not have

IT MAKES rank — every candidate's normalised score per criterion, the weighted parts and the total, ranked; TIE stated, never broken; UNSCORED candidates apart; the first line names the weights file and its criteria; NO WEIGHTS (exit 2) without a weights file or with fewer than 2 criteria; explain — the arithmetic for one candidate: each reading, the min and max it was set against, the normalised score, the weight and the part, the total and its place; an UNSCORED candidate is shown and refused a place (exit 1); UNKNOWN CANDIDATE exit 2; sensitivity — for each criterion the smallest weight change, in steps of 0.05 up to 0.50 either way with the other weights re-normalised proportionally, that changes the top, and the new top; NONE WITHIN RANGE when none does; gaps — every blank score by candidate and by criterion; nothing is guessed in.

THE LAW A ranking is only as honest as its weights, so the weights are written down before any score is read, and the arithmetic is printed beside every total. A candidate with a blank score is UNSCORED and stands outside the ranking; a tie is stated, never broken by the machine; a sensitivity line says which weight would change the top.

IN THE BOX SKILL.md · CARD.md · ▣ scripts/board.py (machine) · ▣ scripts/evals.py (test runner) · tests/EVALS.json (9 tests) · references/ (schema + fictional demo) · START-HERE.md · system/ doors · MY-LEDGER.md · LICENSE-NOTE.md · card.svg · card.json · card.html · AUDIT.md · MANIFEST.md · WITNESS.md · SELFTEST.md.

PROOF SELFTEST 15/15 · witnessed in Claude 2026-10-14 on fictional fixtures · the test prompts in fresh chat, 9/9 · Edition 2.

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. LICENSE-NOTE.md rides in the packet and is the licence.

THE PUNCH · SHA-256 OF SKILL.MD · 1B1E6FAEAA53FFE28F6095F9

THE FACE, FULL SIZE → · THE MACHINE FACE → · card.json →

THE BOARD CARD

The ranking seat: candidates scored on written weights, the arithmetic shown, the ties stated.

RANK · EXPLAIN · SENSITIVITY · GAPS

LEGENDARYWITNESSEDIN CLAUDE IN LUNDR IN CHATGPT

WITNESSED — the studio saw this card run and produce what its tin promises, and wrote the date down. What the marks mean →

THE SPEC SHEET · READ BEFORE YOU BUY · THE STANDARD EVERY CARD MEETS →

SEAL: 98 of 100 · LEGENDARY · WITNESSED
EDITION: E1 · witnessed 2026-10-14
MACHINE: scripts/board.py · 4 commands · selftest 15/15
RUNS: On your own computer. Standard library only, offline; no account, no key, no telemetry.
HOSTS: Claude · Claude Code · ChatGPT · Gemini · any AI that reads a file (START-HERE.md and system/ name each door)
LICENCE: One person, any number of their own AIs and computers · editions replace files · nothing renews · freeze, not loss
PUNCH: 1B1E6FAEAA53FFE28F6095F9 · sha-256 of SKILL.md · check any file against it →

CITE THIS CARDTHE BOARD CARD, Edition 2. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=board-card. Punch 1B1E6FAEAA53FFE28F6095F9. Retrieved 2026-10-14.

What it is.

Use when several candidates must be ranked on the same named criteria with weights the person wrote down: offers, suppliers, projects, hires. Not for a gut ranking or a yes/no triage.

It carries a machine that shows its working: rank — every candidate's normalised score per criterion, the weighted parts and the total, ranked; TIE stated, never broken; UNSCORED candidates apart; the first line names the weights file and its criteria; NO WEIGHTS (exit 2) without a weights file or with fewer than 2 criteria; explain — the arithmetic for one candidate: each reading, the min and max it was set against, the normalised score, the weight and the part, the total and its place; an UNSCORED candidate is shown and refused a place (exit 1); UNKNOWN CANDIDATE exit 2; sensitivity — for each criterion the smallest weight change, in steps of 0.05 up to 0.50 either way with the other weights re-normalised proportionally, that changes the top, and the new top; NONE WITHIN RANGE when none does; gaps — every blank score by candidate and by criterion; nothing is guessed in — 15 checks in its selftest. It never ranks without a weights file; never ranks a candidate with a blank score; never breaks a tie by itself; never claims a certainty it does not have.

The law, written into the card.

It is in the card itself, so the AI follows it.

THE LAWA ranking is only as honest as its weights, so the weights are written down before any score is read, and the arithmetic is printed beside every total. A candidate with a blank score is UNSCORED and stands outside the ranking; a tie is stated, never broken by the machine; a sensitivity line says which weight would change the top.

The method.

Write the criteria and their weights in a file first, each with whether higher is better; no weights, no ranking (NO WEIGHTS).
Score every candidate on every criterion in the sheet; a blank is UNSCORED and that candidate is listed apart until it is filled.
Normalise each criterion across the candidates, weigh, sum; print the arithmetic for every candidate so the total can be checked by hand.
State ties as ties; the machine never breaks one.
Run the sensitivity: for each weight, the smallest change that would change the top; the board reads it before it decides.

What you bring.

The candidates and their scores on each criterion.
The criteria, with a weight and a direction each.
The question the ranking is meant to answer.

Planned as a separate seat, not in this box and not yet on sale: the studio's own boards for its suppliers and offers.

What it does once your AI has it.

rank Every candidate's normalised score per criterion, the weighted parts and the total, ranked; TIE stated, never broken; UNSCORED candidates apart; the first line names the weights file and its criteria; NO WEIGHTS (exit 2) without a weights file or with fewer than 2 criteria
explain The arithmetic for one candidate: each reading, the min and max it was set against, the normalised score, the weight and the part, the total and its place; an UNSCORED candidate is shown and refused a place (exit 1); UNKNOWN CANDIDATE exit 2
sensitivity For each criterion the smallest weight change, in steps of 0.05 up to 0.50 either way with the other weights re-normalised proportionally, that changes the top, and the new top; NONE WITHIN RANGE when none does
gaps Every blank score by candidate and by criterion; nothing is guessed in

What it will never do.

It never ranks without a weights file; never ranks a candidate with a blank score; never breaks a tie by itself; never claims a certainty it does not have.

Where it hands off: A keep/park/kill call on a list of ideas → THE CONCEPT RANKER; A price built from a cost → THE PRICE CARD; A choice a client must make between options → THE PICK ROOM CARD.

It is not a recommendation. It is an AI reading a text file: it can be wrong, so check what matters against your own records and the source. The holder brings their own AI; no endorsement by Anthropic or any other AI company is claimed.

What’s in the buy.

SKILL.md — the card your AI reads: when it wakes, the loop, the never-lines, the hand-offs.
CARD.md — the whole card in ten sections: face, law, method, machine, what you bring, never, tests, seal, term.
scripts/board.py — the machine, Python standard library only; `selftest` proves it on its own fixtures.
tests/EVALS.json and scripts/evals.py — 9 test prompts and the offline runner that scores your AI's answers.
references/ — the data schema and FICTIONAL demo files, invented to show the machine, not real.
MY-LEDGER.md — your own record of what you asked and what you decided.
START-HERE.md, system/short-prompt.md, system/rules.md — the doors, for any AI with or without skills.
card.svg, card.json, card.html, AUDIT.md, MANIFEST.md, WITNESS.md, SELFTEST.md — the faces and the proof.
NOT IN THE BOX — any studio edition (planned, not yet on sale: the studio's own boards for its suppliers and offers). Also not in the box: any data you must bring, and advice of any kind (not a recommendation).

PROOF · SELFTEST 15/15 · WITNESSED IN CLAUDE 2026-10-14 ON FICTIONAL FIXTURES · THE 9 TEST PROMPTS ASKED IN A FRESH CHAT HOLDING ONLY THE CARD · 9/9 · EDITION 2 · PUNCH 1B1E6FAEAA53FFE28F6095F9

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 420 — 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 · NOTHING RENEWS ITSELF

LAUNCHING SOON

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 CARD’S OWN SELFTEST.MD

board.py, every command showing its working.

A small Python program, standard library only. Your AI runs it; you can too. The block below is pasted from the card’s own SELFTEST.md, as printed when it was run from inside the packet.

`rank`	every candidate's normalised score per criterion, the weighted parts and the total, ranked; TIE stated, never broken; UNSCORED candidates apart; the first line names the weights file and its criteria; NO WEIGHTS (exit 2) without a weights file or with fewer than 2 criteria
`explain`	the arithmetic for one candidate: each reading, the min and max it was set against, the normalised score, the weight and the part, the total and its place; an UNSCORED candidate is shown and refused a place (exit 1); UNKNOWN CANDIDATE exit 2
`sensitivity`	for each criterion the smallest weight change, in steps of 0.05 up to 0.50 either way with the other weights re-normalised proportionally, that changes the top, and the new top; NONE WITHIN RANGE when none does
`gaps`	every blank score by candidate and by criterion; nothing is guessed in
`selftest`	15 checks that prove the machine; it must print selftest 15/15

THE SELFTEST, AS PRINTED

ok   1  rank: the first line names the weights file and every criterion with its weight and direction; 4 ranked, 1 UNSCORED apart; exit 0
ok   2  rank prints the min-max span per criterion, then each candidate's normalised score × weight = part and the total, in order
ok   3  rank states a TIE and never breaks it: Example Ridge Works and Sample Creek Supply share rank 3 at 0.400, each line marked TIE, a TIE line for a person
ok   4  rank never ranks an UNSCORED candidate: Harbor Example Co (quality blank) stands apart, holds no rank, and its price 7 is kept out of the normalisation (price min stays 8)
ok   5  rank is deterministic (two runs, the same bytes) and weights that do not sum to 1 count as shares, said in the first line, with the same ranking lines
ok   6  rank flags NO SPREAD when every reading on a criterion is the same: lead_days gives every candidate 0.500 and separates nobody
ok   7  rank refuses NO WEIGHTS (exit 2) when the weights file is missing and when it names fewer than 2 criteria; no ranking line is printed
ok   8  a weight that is not a number, a direction that is not true/false, a criterion with no column, a cell that is not a number and a duplicate candidate are bad input (exit 2)
ok   9  rank refuses NOT ENOUGH CANDIDATES (n=1, need 2) when only one candidate is scored (exit 1)
ok   10  explain prints the arithmetic for one candidate: each reading against the min and max, the normalised score, × weight = part, the sum, the rank, who is ahead and behind
ok   11  explain shows an UNSCORED candidate's blank and what was read, refuses it a place (exit 1); a candidate not on the sheet is UNKNOWN CANDIDATE (exit 2)
ok   12  sensitivity finds, per criterion, the smallest weight change in 0.05 steps that moves the top, re-normalising the others: price +0.10 hands the top to Maple Example Shop, quality -0.15 does too
ok   13  sensitivity states a new TIE at the top rather than breaking it: lead_days -0.10 leaves Maple Example Shop and Riverbend Example Goods level at 0.610
ok   14  sensitivity prints NONE WITHIN RANGE for every criterion when no weight change up to 0.50 moves the top
ok   15  gaps lists every blank by candidate and by criterion, and says no gaps when the sheet is full
selftest 15/15

THE RUN IS PASTED FROM THE CARD’S OWN SELFTEST.MD AS PRINTED. ITS FIXTURES ARE FICTIONAL, INVENTED FOR THE TEST; NO PRIVATE PERSON IS NAMED.

QUESTIONS AT THE TABLE

Asked before the wire.

What does it need from me?

The candidates and their scores on each criterion. The criteria, with a weight and a direction each. The question the ranking is meant to answer. If something it needs is missing, it says what is missing and stops there.

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 machine's selftest (15/15), run from inside the packet in Claude on 2026-10-14 on fictional fixtures and pasted into SELFTEST.md as printed; the twelve-check audit (AUDIT.md). Owed, and said so in WITNESS.md: Any run on real data. Every demo file is fictional. 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: 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. LICENSE-NOTE.md rides in the packet and is the licence. It is not a recommendation.

What will it never do?

It never ranks without a weights file; never ranks a candidate with a blank score; never breaks a tie by itself; never claims a certainty it does not have.

← THE METHODS · THE FLOOR →