THE REPEAT CARD · what a transcript repeated, cut into drafts
WAKES WHEN someone keeps doing the same steps, or asks what in a long session should become a skill
THE LOOP
repeat.py mine · the procedures, by count
show · each occurrence, the ask before it
draft --out · SKILL.md and the skeleton
write the four lines the draft could not
then THE DESCRIPTION CARD and THE TRIGGER
NEVER
Run a drafted skeleton before a person reads it
Call a draft a skill
Infer a purpose the transcript never stated
IT MAKES mine — the repeated procedures, most repeated first; show — one procedure: every occurrence, with the asks before it; draft — the skill draft and the skeleton.
THE STANDARD a procedure is between asks; the program, the script and the verb are kept; the longest with a given count wins; the ask before it is the trigger; a draft names what it could not infer; the skeleton only prints.
PROOF SELFTEST 11/11 · witnessed in Claude 2026-10-09 · Edition 1.
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. THE-CLAUSES.md rides in the packet and is part of the licence.
THE PUNCH · SHA-256 OF SKILL.MD · 28DC38E7C7A99F5C9E479574
CITE THIS CARDTHE REPEAT CARD, Edition 1. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=repeat-card. Punch 28DC38E7C7A99F5C9E479574. Retrieved 2026-10-10.
What it is.
Mines a session transcript — markdown, text, log or JSON lines — for the procedures that recurred: runs of commands between asks that appeared two or more times with their paths, numbers and strings abstracted; shows every occurrence with the ask before it; and drafts a skill from one — a SKILL.md with the steps, the first occurrence as it happened, the ask as the trigger, a skeleton script, and an honest list of what the draft could not infer.
It carries a machine that shows its working: mine — the repeated procedures, most repeated first; show — one procedure: every occurrence, with the asks before it; draft — the skill draft and the skeleton — 11 checks in its selftest, every command on the record. It will never run a drafted skeleton before a person reads it; never call a draft a skill; never infer a purpose the transcript never stated.
Six laws, written into the card.
They are in the card itself, so the AI follows them.
A PROCEDURE IS BETWEEN ASKSCommands that recur between two requests, two to six steps, counted with their arguments abstracted.
THE PROGRAM, THE SCRIPT AND THE VERB ARE KEPTPaths, numbers and strings become placeholders; flags stay.
THE LONGEST WITH A GIVEN COUNT WINSThe longest with a given count wins, so the report says one thing, not every window inside it.
THE ASK BEFORE IT IS THE TRIGGERThe ask before it is the trigger, in the words the person used; no ask found is said plainly.
A DRAFT NAMES WHAT IT COULD NOT INFERPurpose, which arguments vary, what done looks like, the trigger. A person writes those four lines.
THE SKELETON ONLY PRINTSIt runs nothing until a person has filled in the arguments and read every step.
What rides in the card.
In the card’s own voice, with a machine that shows its working.
THE GUIDEreferences/MINING-A-TRANSCRIPT.md
THE MACHINErepeat.py · three commandsmine — the repeated procedures, most repeated first; show — one procedure: every occurrence, with the asks before it; draft — the skill draft and the skeleton. Every command prints its working and refuses a bad input; 11 checks in its selftest, every command on the record.
What it does once your AI has it.
mine The repeated procedures, most repeated first
show One procedure: every occurrence, with the asks before it
draft The skill draft and the skeleton
On the record A session transcript mined: cut-then-validate found three times with its three asks, and the three-step version with the zip twice; the first shown at every occurrence; a skill drafted from it with the steps, the first occurrence, the ask as the trigger, a skeleton that only prints, and the four things it could not infer.
What it will never do.
It will never run a drafted skeleton before a person reads it; never call a draft a skill; never infer a purpose the transcript never stated.
It finds commands it can recognise (python, node, git, make, shell tools…) on their own lines; prose instructions that recur without commands are not mined. Asks are lines that open with user:, human: or a name; a transcript without them is one run. A transcript shows what was typed, never why. It is an AI reading a text file: it can be wrong, so check what matters against your own records and the source.
What’s in the buy.
CARD.md — the whole card: who it is, its laws, when it says no, what it carries · SKILL.md — the door your AI reads first · START-HERE.md — the first five minutes, for you
MINING-A-TRANSCRIPT.md — the guide in references/, in the card's own voice
scripts/repeat.py — the machine: mine · show · draft, standard library only, selftest 11/11
WITNESS.md + SELFTEST.md — every command on the record, as it printed · MANIFEST.md, every file hashed · AUDIT.md, the twelve checks
MY-LEDGER.md — the card's memory, on your own machine · TIN.md · LICENSE-NOTE.md · THE-CLAUSES.md · HOSTS.md
card.svg + card-machine.svg + card.json · card.html, the card as one offline file · the doors for ChatGPT, Gemini, Cursor, Claude Code and MCP hosts (rules/, system/, mcp/)
PROOF · SELFTEST 11/11 · 3 COMMANDS ON THE RECORD · WITNESSED IN CLAUDE 2026-10-09 · EDITION 1 · PUNCH 28DC38E7C7A99F5C9E479574
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 913 PROPOSED — 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. 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 WITNESSED RUN OF 2026-10-09
repeat.py, every command showing its working.
A small Python program, standard library only. Your AI runs it; you can too. Every block below is pasted from the card’s own WITNESS.md, as printed.
mine
the repeated procedures, most repeated first
show
one procedure: every occurrence, with the asks before it
draft
the skill draft and the skeleton
selftest
11 checks that prove the machine; it must print 11/11
mine · THE REPEATED PROCEDURES, MOST REPEATED FIRST
show · ONE PROCEDURE: EVERY OCCURRENCE, WITH THE ASKS BEFORE IT
$ python3 scripts/repeat.py show session.md 1
procedure 1 · ×3
python3 cardkit.py cut <arg>
python3 validate.py check <path> --deep
at line 2 · after: "cut the eval card and check it":
$ python3 cardkit.py cut eval-card
$ python3 validate.py check out/eval-card.skill.zip --deep
at line 7 · after: "now the trigger card":
$ python3 cardkit.py cut trigger-test-card
$ python3 validate.py check out/trigger-test-card.skill.zip --deep
at line 12 · after: "and the grounding one please":
$ python3 cardkit.py cut grounding-card
$ python3 validate.py check out/grounding-card.skill.zip --deep
draft · THE SKILL DRAFT AND THE SKELETON
$ python3 scripts/repeat.py draft session.md 1 --name cut-and-validate --out drafts/cut-and-validate
wrote drafts/cut-and-validate/SKILL.md and scripts/cut-and-validate.py — read "what this draft could not infer" before anyone runs it
cat drafts/cut-and-validate/SKILL.md
$ cat drafts/cut-and-validate/SKILL.md
---
name: cut-and-validate
description: Runs the 2-step procedure that came up 3 times in the transcript (python3 → python3). Use when someone asks: "cut the eval card and check it". Runs scripts/cut-and-validate.py.
---
# CUT AND VALIDATE
## The steps
1. `python3 cardkit.py cut <arg>`
2. `python3 validate.py check <path> --deep`
## As it happened (first occurrence)
THE NUMBERS, NAMES AND DATES IN THESE RUNS ARE THE STUDIO’S OWN EXAMPLES, INVENTED FOR THE RECORD; NO REAL COMPANY, PERSON OR PROPERTY IS NAMED.
QUESTIONS AT THE TABLE
Asked before the wire.
What does it need from me?
Repeat.py mine · the procedures, by count. It asks first for the transcripts, and the person who knows why the steps were run.
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 self-test (11/11) and every command run on real example inputs, pasted into WITNESS.md as they printed; 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 finds commands it can recognise (python, node, git, make, shell tools…) on their own lines; prose instructions that recur without commands are not mined. Asks are lines that open with user:, human: or a name; a transcript without them is one run. A transcript shows what was typed, never why.
What will it never do?
It will never run a drafted skeleton before a person reads it; never call a draft a skill; never infer a purpose the transcript never stated.
