THE INSTRUMENTS · THE HELM · Nº 1242 · EDITION 1 · LAUNCHING SOON

LUND CARDS · THE INSTRUMENTSNº 1242

THE CARD AS TEXT · FOR EYES AND FOR MACHINES

THE UNLIMITED CARD · The room never runs out: budget, digest, ledger, and the hand-off a fresh session reads first.

WAKES WHEN A session is deep into a long build and the person feels the room closing in. · A fresh chat has to pick up exactly where the last one stopped, without the old chat. · A file is too big to read whole and someone is about to paste all of it into the conversation. · A person asks how much of the context a folder of notes would take before loading it.

THE LOOP

Budget the files before reading them
Trim what is big into a digest that fits
Note every step on the ledger
Hand off; the next session reads it first

NEVER

Never pastes a whole file into the chat
Never claims a memory it did not read
Never calls a digest the file
Never writes a ledger entry with no next step

IT MAKES budget — estimate the tokens in a file or a folder of text files, biggest first; OVER LIMIT when the total passes the limit; trim — a digest that fits: the first and last lines, the matching lines with numbers, and LEFT OUT counted; note — add a step to the ledger; refuses one with no next step; resume — print the ledger's latest step as the resume note; handoff — write the hand-off a fresh session reads first: state, done, next, files with their estimated tokens.

THE LAW Tokens are finite; the room is not, when the work lives in files and the chat carries only its shape. Everything that must outlive the chat is written down in words a fresh session can pick up without the old one, and every estimate says it is an estimate.

IN THE BOX SKILL.md · CARD.md · ▣ scripts/unlimited.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 13/13 · witnessed in Claude 2026-10-13 on fictional fixtures · the test prompts in fresh chats owed · Edition 1.

THE SEAL 96 of 100 · LEGENDARY · WITNESSED · Trigger 5 · Machinery 5 · Law 5 · Portability 5 · Proof 4

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 · BC147C72B4B514923F4239FF

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

THE UNLIMITED CARD

The room never runs out: budget, digest, ledger, and the hand-off a fresh session reads first.

BUDGET · TRIM · NOTE · RESUME · HANDOFF

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: 96 of 100 · LEGENDARY · WITNESSED
EDITION: E1 · witnessed 2026-10-13
MACHINE: scripts/unlimited.py · 5 commands · selftest 13/13
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: BC147C72B4B514923F4239FF · sha-256 of SKILL.md · check any file against it →

CITE THIS CARDTHE UNLIMITED CARD, Edition 1. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=unlimited-card. Punch BC147C72B4B514923F4239FF. Retrieved 2026-10-13.

What it is.

Use when a long piece of work must outlive one chat or one context window: a budget, a digest, a ledger and the hand-off note. Not for the work itself or for any AI's own limits.

It carries a machine that shows its working: budget — estimate the tokens in a file or a folder of text files, biggest first; OVER LIMIT when the total passes the limit; trim — a digest that fits: the first and last lines, the matching lines with numbers, and LEFT OUT counted; note — add a step to the ledger; refuses one with no next step; resume — print the ledger's latest step as the resume note; handoff — write the hand-off a fresh session reads first: state, done, next, files with their estimated tokens — 13 checks in its selftest. It never pastes a whole file into the chat; never claims a memory it did not read; never calls a digest the file; never writes a ledger entry with no next step.

The law, written into the card.

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

THE LAWTokens are finite; the room is not, when the work lives in files and the chat carries only its shape. Everything that must outlive the chat is written down in words a fresh session can pick up without the old one, and every estimate says it is an estimate.

The method.

Before reading a folder, budget it: every text file's size and its estimated tokens, biggest first, against the limit you give. Over the limit, the machine names what to trim; under it, read on.
Trim a big file into a digest that fits: the first lines, the last lines, the lines that match what you are looking for, each with its number, and LEFT OUT counted. The digest says it is not the file.
Note every step on the ledger as you go: where things stand, what was done, what comes next, and the files the next session should read first. A step with no next step is refused, because a dead end is not a ledger; resume prints the latest step back.
When the room closes, hand off: one note written from the ledger that a fresh session reads first, with each named file's estimated weight so it can budget before it reads.
Say the estimate's rule every time: four bytes to a token, a rule of thumb from the bytes on disk, never a count from any model. A number without its rule is a guess dressed up.

What you bring.

The files the work lives in, on your own disk: notes, drafts, code, records.
The limit you are working to, in tokens, when you have one.
The habit of writing the ledger before the room closes, not after.

Planned as a separate seat, not in this box and not yet on sale: the ledger and hand-off the studio keeps for its own long builds.

What it does once your AI has it.

budget Estimate the tokens in a file or a folder of text files, biggest first; OVER LIMIT when the total passes the limit
trim A digest that fits: the first and last lines, the matching lines with numbers, and LEFT OUT counted
note Add a step to the ledger; refuses one with no next step
resume Print the ledger's latest step as the resume note
handoff Write the hand-off a fresh session reads first: state, done, next, files with their estimated tokens

What it will never do.

It never pastes a whole file into the chat; never claims a memory it did not read; never calls a digest the file; never writes a ledger entry with no next step.

Where it hands off: Keeping a long context in order while it is open → THE LONG CONTEXT CARD; Folding a built thing smaller without losing it → THE COMPRESSOR CARD; What a session costs in money → THE TOKEN COST CARD.

It is not a measure of any AI's limits. 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/unlimited.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 ledger and hand-off the studio keeps for its own long builds). Also not in the box: any data you must bring, and advice of any kind (not a measure of any AI's limits).

PROOF · SELFTEST 13/13 · WITNESSED IN CLAUDE 2026-10-13 ON FICTIONAL FIXTURES · THE TEST PROMPTS IN FRESH CHATS OWED · EDITION 1 · PUNCH BC147C72B4B514923F4239FF

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 1242 — 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

unlimited.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.

`budget`	estimate the tokens in a file or a folder of text files, biggest first; OVER LIMIT when the total passes the limit
`trim`	a digest that fits: the first and last lines, the matching lines with numbers, and LEFT OUT counted
`note`	add a step to the ledger; refuses one with no next step
`resume`	print the ledger's latest step as the resume note
`handoff`	write the hand-off a fresh session reads first: state, done, next, files with their estimated tokens
`selftest`	13 checks that prove the machine; it must print selftest 13/13

THE SELFTEST, AS PRINTED

    10,973  example-notes.md
         6  small.txt
budget <tmp> · 2 text files · 1 skipped · 10,979 tokens · no limit given · estimated at four bytes to a token, a rule of thumb, not a count from any model
ok  1  the budget counts the text files, skips the picture, biggest first
ok  2  the estimate is bytes over four and says so
    10,973  example-notes.md
         6  small.txt
TRIM example-notes.md (10,973 tokens): trim --file ... --max-tokens 200
budget <tmp> · 2 text files · 1 skipped · 10,979 tokens · OVER LIMIT by 9,979 tokens · estimated at four bytes to a token, a rule of thumb, not a count from any model
ok  3  a total past the limit is OVER LIMIT, exit 1, with a TRIM line
digest <tmp> · 400 lines whole · 292 left out · about 3,186 tokens · no limit given
ok  4  the digest keeps the head and the tail and counts what it left out
ok  5  the matching lines are there with their numbers
digest <tmp> · 400 lines whole · 384 left out · about 517 tokens · LIMIT KEPT (517 of 600 tokens)
ok  6  a digest asked to fit a limit fits it: LIMIT KEPT
NOT TEXT: <tmp> — a digest of a binary file would be invented
ok  7  a binary file is NOT TEXT: refused, nothing invented
NO NEXT STEP: a ledger entry without a next step is a dead end; say what comes next
ok  8  a step with no next step is refused: NO NEXT STEP
ledger <tmp> · step 1 added · 1 entries
ok  9  a step with a next step is added as step 1
ledger <tmp> · step 2 added · 2 entries
RESUME FROM STEP 2
STATE  checked
DONE   checked the feed
NEXT   write the front
FILES  notes/missing.md
ok  10  resume prints the latest step as the resume note
handoff <tmp> · from step 2 · 2 entries · 2 files named
ok  11  the hand-off carries next, done, and each file's estimate or NOT FOUND
EMPTY LEDGER: no hand-off can be written from nothing
ok  12  an empty ledger writes no hand-off: EMPTY LEDGER
digest <tmp> · 400 lines whole · 392 left out · about 298 tokens · no limit given
digest <tmp> · 400 lines whole · 392 left out · about 298 tokens · no limit given
ok  13  the same file gives the same digest bytes
selftest 13/13

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 files the work lives in, on your own disk: notes, drafts, code, records. The limit you are working to, in tokens, when you have one. The habit of writing the ledger before the room closes, not after. 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 (13/13), run from inside the packet in Claude on 2026-10-13 on fictional fixtures and pasted into SELFTEST.md as printed; the twelve-check audit (AUDIT.md). Owed, and said so in WITNESS.md: The 9 test prompts asked in fresh chats with the card installed. Until they are run and scored, the tests are owed. 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 measure of any AI's limits.

What will it never do?

It never pastes a whole file into the chat; never claims a memory it did not read; never calls a digest the file; never writes a ledger entry with no next step.

← THE INSTRUMENTS · THE FLOOR →