THE METHODS · THE CARD BENCH · Nº 435 · EDITION 2 · LAUNCHING SOON

LUND CARDS · THE METHODSNº 435

THE CARD AS TEXT · FOR EYES AND FOR MACHINES

THE HONEST CARD · The plain-speaking seat: the claims in a text pulled out and asked for their source.

WAKES WHEN A page says 'the fastest', 'proven', 'guaranteed' and nobody can say from where. · A number appears in a listing with nothing it came from. · A maker wants their copy read the way a careful buyer reads it. · A text hedges on every line and says nothing.

THE LOOP

Pull every claim out with its line and kind
Ask each for its source
Mark FROM MEMORY, NO WORKING, OVERPROMISE
Count the certainties against the hedges

NEVER

Never marks a claim verified with no source
Never takes a say-so as a source
Never rewrites a claim into a softer untruth
Never claims a certainty it does not have

IT MAKES read — pull every claim out of the text into a sheet with the line number, the kind and an empty source column; check — each claim against the sources on file: SOURCED, FROM MEMORY (no source), NO WORKING (a number with none), OVERPROMISE (a guarantee); mark — put a source on a claim; an empty source or one that only says "trust me" is refused; hedge — count the certainties and the hedges and print the ratio; HEDGE-HEAVY and CERTAINTY-HEAVY are flags, not failures; sources — write an empty sources sheet with the columns named.

THE LAW A claim is a certainty, a number, a guarantee or a comparison. Each either stands on a source a reader could follow or is said as a claim. A guarantee stands on nothing, whatever its source, because no writing can keep it. The machine marks; the person writes.

IN THE BOX SKILL.md · CARD.md · ▣ scripts/honest.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 14/14 · 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 · 5A0131C82A21A562BC56AD16

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

THE HONEST CARD

The plain-speaking seat: the claims in a text pulled out and asked for their source.

READ · CHECK · MARK · HEDGE · SOURCES

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/honest.py · 5 commands · selftest 14/14
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: 5A0131C82A21A562BC56AD16 · sha-256 of SKILL.md · check any file against it →

CITE THIS CARDTHE HONEST CARD, Edition 2. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=honest-card. Punch 5A0131C82A21A562BC56AD16. Retrieved 2026-10-14.

What it is.

Use when writing must be read for claims that outrun their proof: certainties with no source, numbers with no working, guarantees, comparisons. Not for witnessing runs or checking file lists.

It carries a machine that shows its working: read — pull every claim out of the text into a sheet with the line number, the kind and an empty source column; check — each claim against the sources on file: SOURCED, FROM MEMORY (no source), NO WORKING (a number with none), OVERPROMISE (a guarantee); mark — put a source on a claim; an empty source or one that only says "trust me" is refused; hedge — count the certainties and the hedges and print the ratio; HEDGE-HEAVY and CERTAINTY-HEAVY are flags, not failures; sources — write an empty sources sheet with the columns named — 14 checks in its selftest. It never marks a claim verified with no source; never takes a say-so as a source; never rewrites a claim into a softer untruth; 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 claim is a certainty, a number, a guarantee or a comparison. Each either stands on a source a reader could follow or is said as a claim. A guarantee stands on nothing, whatever its source, because no writing can keep it. The machine marks; the person writes.

The method.

Read the text: every sentence with a certainty, a number, a promise or a comparison goes into the claims sheet with its line and its kind.
Put a source on each claim: a file, a run, a document, a person who was there. 'Trust me' is refused; an empty source is refused.
Check: SOURCED when a source is on file; SOURCED IN LINE when the sentence itself says where its number came from; FROM MEMORY and NO WORKING when nothing does; OVERPROMISE for any guarantee.
Count the certainties against the hedges: a text of nothing but certainties is CERTAINTY-HEAVY; a text that hedges every line is HEDGE-HEAVY, which is its own untruth.
Rewrite by hand: say what was measured, say what is from memory, and drop what cannot be said either way.

What you bring.

The text as it stands.
The sources you have: files, runs, documents, people.
The will to cut a line that stands on nothing.

Planned as a separate seat, not in this box and not yet on sale: the reading the studio gives every room and every tin it writes.

What it does once your AI has it.

read Pull every claim out of the text into a sheet with the line number, the kind and an empty source column
check Each claim against the sources on file: SOURCED, FROM MEMORY (no source), NO WORKING (a number with none), OVERPROMISE (a guarantee)
mark Put a source on a claim; an empty source or one that only says "trust me" is refused
hedge Count the certainties and the hedges and print the ratio; HEDGE-HEAVY and CERTAINTY-HEAVY are flags, not failures
sources Write an empty sources sheet with the columns named

What it will never do.

It never marks a claim verified with no source; never takes a say-so as a source; never rewrites a claim into a softer untruth; never claims a certainty it does not have.

Where it hands off: Keeping the runs that would be the sources → THE WITNESS CARD; Checking the file list against the box → THE TIN CARD; The numbers themselves → the person who measured them.

It is not a fact-checking service. 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/honest.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 reading the studio gives every room and every tin it writes). Also not in the box: any data you must bring, and advice of any kind (not a fact-checking service).

PROOF · SELFTEST 14/14 · 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 5A0131C82A21A562BC56AD16

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 435 — 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

OFFLINE · ONE LICENCE · PUNCH-CHECKABLE · THE STANDARD

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

honest.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.

`read`	pull every claim out of the text into a sheet with the line number, the kind and an empty source column
`check`	each claim against the sources on file: SOURCED, FROM MEMORY (no source), NO WORKING (a number with none), OVERPROMISE (a guarantee)
`mark`	put a source on a claim; an empty source or one that only says "trust me" is refused
`hedge`	count the certainties and the hedges and print the ratio; HEDGE-HEAVY and CERTAINTY-HEAVY are flags, not failures
`sources`	write an empty sources sheet with the columns named
`selftest`	14 checks that prove the machine; it must print selftest 14/14

THE SELFTEST, AS PRINTED

  1  line 3    certainty            It is the fastest tool of its kind.
  2  line 3    number               Our tests ran 1,200 jobs in 4 hours, measured on the sample set.
  3  line 3    promise+certainty    It is guaranteed to never fail.
  4  line 3    comparison           It is better than the old one.
  5  line 3    number               About 40% of users said so in the survey (n=52).
read <tmp> · 5 claims · sources column empty; put one on each or say it is from memory
NO SOURCE: a claim is never marked verified with an empty source; refused
NOT A SOURCE: 'trust me' is a say-so, not a source; refused
mark claim 1 · source: benchmarks/2026-01-run.csv · <tmp>
NO CLAIM 99 in <tmp>
  1  SOURCED: benchmarks/2026-01-run.csv
  2  SOURCED IN LINE: the sentence says where the number came from
  3  OVERPROMISE: a guarantee no writing can keep; say what was measured instead
  4  FROM MEMORY: said without a source; say so in the text or find one
  5  SOURCED IN LINE: the sentence says where the number came from
check <tmp> · 3 sourced · 2 not
hedge <tmp> · 7 sentences · 3 certainties · 1 hedges · CERTAINTY-HEAVY: say what was measured and what is from memory
hedge <tmp> · 5 sentences · 6 certainties · 0 hedges · CERTAINTY-HEAVY: say what was measured and what is from memory
ok   1  read pulls out certainties, numbers, promises and comparisons, and leaves the plain sentence alone
ok   2  read writes the claims sheet with an empty source column
ok   3  a certainty with no source is FROM MEMORY
ok   4  a number whose sentence says where it came from is SOURCED IN LINE
ok   5  a guarantee is OVERPROMISE whatever the source
ok   6  a bare number is NO WORKING
ok   7  a claim with a real source on file is SOURCED
ok   8  marking a claim with an empty source is NO SOURCE: refused
ok   9  marking with a say-so is NOT A SOURCE: refused
ok   10  mark puts a real source on the claim
ok   11  a claim id that is not there is NO CLAIM
ok   12  check writes the report and exits 1 while any claim stands on nothing
ok   13  hedge prints the counts and a flag or balanced
ok   14  a text of certainties is CERTAINTY-HEAVY, a flag not a failure
selftest 14/14

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 text as it stands. The sources you have: files, runs, documents, people. The will to cut a line that stands on nothing. 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 (14/14), 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 fact-checking service.

What will it never do?

It never marks a claim verified with no source; never takes a say-so as a source; never rewrites a claim into a softer untruth; never claims a certainty it does not have.

← THE METHODS · THE FLOOR →