THE METHODS · THE STUDIO ROOM · Nº 436 · EDITION 2 · LAUNCHING SOON

LUND CARDS · THE METHODSNº 436

THE CARD AS TEXT · FOR EYES AND FOR MACHINES

THE AFTER-ACTION CARD · The look-back seat: what happened against the plan.

WAKES WHEN Something went wrong and three people remember it three ways. · A launch is over and the team wants what to keep and what to change, on paper. · A plan said five steps would happen and nobody knows which did. · Actions were promised at the last look-back and nobody can find them.

THE LOOP

Build the timeline from the record, in order
Diff it against what was expected
Write actions; each with an owner and a day
Report: from the log, or as remembered

NEVER

Never writes a cause the record does not hold
Never lists an action without an owner
Never names a person as the fault
Never claims a certainty it does not have

IT MAKES timeline — every event in time order, each tagged "from the log", "from a ticket" or "as remembered"; a silence longer than the threshold is printed as GAP n min; diff — each expected step DONE on time, DONE · LATE by n min or MISSED, matched when every word of the step begins a word of an event's what (case-insensitive; post matches posted); events matching no step are EXTRA; whys — every why beside the event it cites; a why whose event_id is blank or not in the record is NO EVENT CITED (exit 1 if any); actions — every action with its owner and due day; NO OWNER or NO DAY named and exit 1 if any; open ones listed with OVERDUE by n days or due in n days; report — the one page: what happened, what was expected, why (cited causes only; the rest as questions for the room), actions (owned and dated only; the rest listed as not on the list); header "A record, not a verdict. No line names a person as the fault.".

THE LAW The look-back is a record, not a verdict. Every line of the timeline says where it came from — the log or someone's memory — and the two are never mixed. A cause is written only when an event in the record shows it; an action is written only with an owner and a day. No line names a person as the fault.

IN THE BOX SKILL.md · CARD.md · ▣ scripts/afteraction.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 16/16 · 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 · 976FD724600855216DF2ED47

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

THE AFTER-ACTION CARD

The look-back seat: what happened against the plan.

TIMELINE · DIFF · WHYS · ACTIONS · REPORT

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/afteraction.py · 5 commands · selftest 16/16
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: 976FD724600855216DF2ED47 · sha-256 of SKILL.md · check any file against it →

CITE THIS CARDTHE AFTER-ACTION CARD, Edition 2. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=after-action-card. Punch 976FD724600855216DF2ED47. Retrieved 2026-10-14.

What it is.

Use when a launch, an outage or a project needs its look-back: the timeline from the record, the gap against the plan, actions with owners. Not for assigning blame or for a live incident.

It carries a machine that shows its working: timeline — every event in time order, each tagged "from the log", "from a ticket" or "as remembered"; a silence longer than the threshold is printed as GAP n min; diff — each expected step DONE on time, DONE · LATE by n min or MISSED, matched when every word of the step begins a word of an event's what (case-insensitive; post matches posted); events matching no step are EXTRA; whys — every why beside the event it cites; a why whose event_id is blank or not in the record is NO EVENT CITED (exit 1 if any); actions — every action with its owner and due day; NO OWNER or NO DAY named and exit 1 if any; open ones listed with OVERDUE by n days or due in n days; report — the one page: what happened, what was expected, why (cited causes only; the rest as questions for the room), actions (owned and dated only; the rest listed as not on the list); header "A record, not a verdict. No line names a person as the fault." — 16 checks in its selftest. It never writes a cause the record does not hold; never lists an action without an owner; never names a person as the fault; 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 LAWThe look-back is a record, not a verdict. Every line of the timeline says where it came from — the log or someone's memory — and the two are never mixed. A cause is written only when an event in the record shows it; an action is written only with an owner and a day. No line names a person as the fault.

The method.

Collect events into one sheet: the time, who, what, and the source (log, ticket or memory); sort them; a gap longer than the house's threshold is marked so it can be asked about.
Lay the expected steps beside the timeline: each step is DONE, LATE, MISSED or EXTRA, and the times say by how much.
Write the 'why' lines only where an event id can be cited; everything else is a question for the room, listed as such.
Write the actions: each with an owner and a day, or it does not go on the list (NO OWNER is printed and the line stays open).
Report in one page: what happened, what was expected, where they parted, the actions; each fact tagged 'from the log' or 'as remembered'.

What you bring.

The events, with times and where each came from.
What was expected: the plan or the steps.
The people who will own the actions.

Planned as a separate seat, not in this box and not yet on sale: the studio's own look-backs on its builds and launches.

What it does once your AI has it.

timeline Every event in time order, each tagged "from the log", "from a ticket" or "as remembered"; a silence longer than the threshold is printed as GAP n min
diff Each expected step DONE on time, DONE · LATE by n min or MISSED, matched when every word of the step begins a word of an event's what (case-insensitive; post matches posted); events matching no step are EXTRA
whys Every why beside the event it cites; a why whose event_id is blank or not in the record is NO EVENT CITED (exit 1 if any)
actions Every action with its owner and due day; NO OWNER or NO DAY named and exit 1 if any; open ones listed with OVERDUE by n days or due in n days
report The one page: what happened, what was expected, why (cited causes only; the rest as questions for the room), actions (owned and dated only; the rest listed as not on the list); header "A record, not a verdict. No line names a person as the fault."

What it will never do.

It never writes a cause the record does not hold; never lists an action without an owner; never names a person as the fault; never claims a certainty it does not have.

Where it hands off: Running the next launch to a plan → THE ROAD CARD; A rule so it never happens again → THE HOOK CARD; Deciding who was responsible → the people in the room.

It is not a finding of fault. 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/afteraction.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 look-backs on its builds and launches). Also not in the box: any data you must bring, and advice of any kind (not a finding of fault).

PROOF · SELFTEST 16/16 · 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 976FD724600855216DF2ED47

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 436 — 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

afteraction.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.

`timeline`	every event in time order, each tagged "from the log", "from a ticket" or "as remembered"; a silence longer than the threshold is printed as GAP n min
`diff`	each expected step DONE on time, DONE · LATE by n min or MISSED, matched when every word of the step begins a word of an event's what (case-insensitive; post matches posted); events matching no step are EXTRA
`whys`	every why beside the event it cites; a why whose event_id is blank or not in the record is NO EVENT CITED (exit 1 if any)
`actions`	every action with its owner and due day; NO OWNER or NO DAY named and exit 1 if any; open ones listed with OVERDUE by n days or due in n days
`report`	the one page: what happened, what was expected, why (cited causes only; the rest as questions for the room), actions (owned and dated only; the rest listed as not on the list); header "A record, not a verdict. No line names a person as the fault."
`selftest`	16 checks that prove the machine; it must print selftest 16/16

THE SELFTEST, AS PRINTED

ok   1  timeline sorts the record by time and tags every line from the log, from a ticket or as remembered, never untagged
ok   2  timeline prints GAP n min for a silence over the threshold, between the two events, and counts it
ok   3  a wider --gap-minutes drops the GAP: the threshold is the person's
ok   4  a source outside log|ticket|memory and a malformed hour are bad input (exit 2): an event that cannot be tagged is never written
ok   5  diff: DONE on time with the minutes early, DONE · LATE by n min, MISSED with the words looked for, EXTRA for events matching no step; a step word matches an event word that begins with it
ok   6  diff: a step with no time on the plan is DONE · NO TIME ON THE PLAN, never called on time or late
ok   7  whys: a cited why is printed beside its event, time and source; a blank or unknown event_id is NO EVENT CITED; exit 1
ok   8  whys exits 0 when every why cites an event in the record
ok   9  actions: NO OWNER and NO DAY are named and kept off the list (exit 1); OVERDUE by n days and due in n days from --today; done is not overdue
ok   10  actions exits 0 when every action has an owner and a day; a status outside open|done is bad input (exit 2)
ok   11  report: the header stands on line two and every happened-line carries a source tag
ok   12  report never prints a cause without its event id: the Why section holds only cited whys, each with ← e-id, and the uncited ones sit under Questions as NO EVENT CITED
ok   13  report never lists an action without an owner or a day as valid: a3 and a4 sit under Not on the list with their flags
ok   14  report --out writes the same page to the file
ok   15  report without --whys or --actions says so and invents neither a cause nor an action
ok   16  an events file without the columns and an expected file without steps are bad input (exit 2)
selftest 16/16

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 events, with times and where each came from. What was expected: the plan or the steps. The people who will own the actions. 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 (16/16), 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 finding of fault.

What will it never do?

It never writes a cause the record does not hold; never lists an action without an owner; never names a person as the fault; never claims a certainty it does not have.

← THE METHODS · THE FLOOR →