THE METHODS · CLAUDE CODE · Nº 488 · EDITION 2 · LAUNCHING SOON

LUND CARDS · THE METHODSNº 488

THE CARD AS TEXT · FOR EYES AND FOR MACHINES

THE CREW CARD · The subagent seat: work carved into briefs, each helper owning its files, every report reviewed.

WAKES WHEN A job is too big for one session and someone wants several helpers on it. · The helpers keep stepping on each other's files. · Someone is writing a subagent brief or definition and wants it to hold. · Reports have come back from helpers and someone has to decide what merges.

THE LOOP

Decide whether to split at all
Carve the work into briefs with owned files
Give each helper the fewest tools it needs
Review every report before anything merges

NEVER

Never puts two helpers on one file
Never gives every helper every tool
Never lets a helper merge its own work
Never claims a certainty it does not have

IT MAKES carve — write DIR/<helper>.md briefs from the breakdown; SAME FILE and FULL PERMISSIONS are refused (exit 1); check — check the breakdown without writing: collisions, permissions, NO DONE-WHEN, NO FILES; decide — say whether to split at all: one helper with everything is not a crew, and prints DO NOT SPLIT when the work is one file; review — read each helper's report against its brief; a report with no listed files or no proof line is NOT DONE; list — print every helper with its files and tools.

THE LAW A crew works when ownership is clean: one file, one helper; one brief, one done-when; the fewest tools that do the job; and a reviewer who reads every report before a line merges. A crew of one is a session, and the work of one file is one helper's.

IN THE BOX SKILL.md · CARD.md · ▣ scripts/crew.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 2026-10-14 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 · E6E58CD687ED43F11DBE5C5E

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

THE CREW CARD

The subagent seat: work carved into briefs, each helper owning its files, every report reviewed.

CARVE · CHECK · DECIDE · REVIEW · LIST

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/crew.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: E6E58CD687ED43F11DBE5C5E · sha-256 of SKILL.md · check any file against it →

CITE THIS CARDTHE CREW CARD, Edition 2. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=crew-card. Punch E6E58CD687ED43F11DBE5C5E. Retrieved 2026-10-14.

What it is.

Use when splitting work across helper agents: who owns which files, which tools each gets, how each reports, and the review before any merge. Not for planning the change or undoing what agents did.

It carries a machine that shows its working: carve — write DIR/<helper>.md briefs from the breakdown; SAME FILE and FULL PERMISSIONS are refused (exit 1); check — check the breakdown without writing: collisions, permissions, NO DONE-WHEN, NO FILES; decide — say whether to split at all: one helper with everything is not a crew, and prints DO NOT SPLIT when the work is one file; review — read each helper's report against its brief; a report with no listed files or no proof line is NOT DONE; list — print every helper with its files and tools — 13 checks in its selftest. It never puts two helpers on one file; never gives every helper every tool; never lets a helper merge its own work; 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 crew works when ownership is clean: one file, one helper; one brief, one done-when; the fewest tools that do the job; and a reviewer who reads every report before a line merges. A crew of one is a session, and the work of one file is one helper's.

The method.

Decide first. One helper is not a crew and one file cannot be shared: the machine says DO NOT SPLIT and the work stays in one session.
Write the breakdown as a sheet: helper, task, the files it owns, the tools it gets, and the line that says when it is done.
Carve: the machine writes a brief per helper with the fixed report shape and the files it must not touch. Two helpers on one file is refused; every helper with every tool is refused.
Run the helpers from their briefs, each with only its tools.
Review: each report against its brief; a report with no proof or no files listed is NOT DONE, one that touched a file it does not own is NOT OWNED; nothing merges until every verdict reads done and a person has read the files.

What you bring.

The work, broken into parts a person can name.
The files each part touches.
The tools the helpers may use, by name.

Planned as a separate seat, not in this box and not yet on sale: the crews the studio runs on its own builds, briefs and all.

What it does once your AI has it.

carve Write DIR/<helper>.md briefs from the breakdown; SAME FILE and FULL PERMISSIONS are refused (exit 1)
check Check the breakdown without writing: collisions, permissions, NO DONE-WHEN, NO FILES
decide Say whether to split at all: one helper with everything is not a crew, and prints DO NOT SPLIT when the work is one file
review Read each helper's report against its brief; a report with no listed files or no proof line is NOT DONE
list Print every helper with its files and tools

What it will never do.

It never puts two helpers on one file; never gives every helper every tool; never lets a helper merge its own work; never claims a certainty it does not have.

Where it hands off: Planning the change before anyone starts → THE PLAN CARD; Undoing what the helpers did → THE CHECKPOINT CARD; A workflow one person repeats → THE COMMAND CARD.

It is not a guarantee of merge safety. 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/crew.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 crews the studio runs on its own builds, briefs and all). Also not in the box: any data you must bring, and advice of any kind (not a guarantee of merge safety).

PROOF · SELFTEST 13/13 · 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 E6E58CD687ED43F11DBE5C5E

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 488 — 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

crew.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.

`carve`	write DIR/<helper>.md briefs from the breakdown; SAME FILE and FULL PERMISSIONS are refused (exit 1)
`check`	check the breakdown without writing: collisions, permissions, NO DONE-WHEN, NO FILES
`decide`	say whether to split at all: one helper with everything is not a crew, and prints DO NOT SPLIT when the work is one file
`review`	read each helper's report against its brief; a report with no listed files or no proof line is NOT DONE
`list`	print every helper with its files and tools
`selftest`	13 checks that prove the machine; it must print selftest 13/13

THE SELFTEST, AS PRINTED

carve <tmp> · 3 briefs · 5 files owned · 0 flags
SAME FILE: src/x.py is owned by a and b — one file, one helper
carve refused · 2 helpers · fix the breakdown first
DO NOT SPLIT: one helper is not a crew; do the work in one session (2 files)
decide <tmp> · split: 3 helpers over 5 files, no file shared · each helper reviewable on its own files
review <tmp> · 3 reports · 2 not standing
review <tmp> · 1 reports · 2 not standing
ok   1  a clean breakdown checks clean: no shared file, each helper its own tools
ok   2  carve writes one brief per helper with its files, tools and done-when
ok   3  a brief names the fixed report shape and the files it must not touch
ok   4  two helpers on one file is SAME FILE: refused
ok   5  every helper with every tool is FULL PERMISSIONS: refused
ok   6  one helper with every tool is WIDE TOOLS, flagged not refused
ok   7  a helper with no done-when is NO DONE-WHEN and one with no files is NO FILES
ok   8  carve exits 1 and writes no brief when the breakdown is refused
ok   9  decide says DO NOT SPLIT for one helper
ok   10  decide says split for three helpers over separate files
ok   11  review marks a report that touched a file it does not own NOT OWNED and one with no proof NOT DONE
ok   12  review lets the clean report stand against its done-when
ok   13  a helper that never reported is NO REPORT
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 work, broken into parts a person can name. The files each part touches. The tools the helpers may use, by name. 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 2026-10-14 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 guarantee of merge safety.

What will it never do?

It never puts two helpers on one file; never gives every helper every tool; never lets a helper merge its own work; never claims a certainty it does not have.

← THE METHODS · THE FLOOR →