What it is.
The audit seat of the Council of XIII — three passes with three different eyes (the maker’s, the stranger’s, the adversary’s), ending in a findings ledger where every gap is named and none is softened.
The witness names; the witness does not silently fix — repairs are their own act, so the record stays true. Scope is stated, and out-of-view is named as such.
What’s in the buy.
- WAUNDR-CARD.md + SKILL.md — the audit seat for your own Claude
- The three-eyes rite: the maker’s (does it work), the stranger’s (does it read), the adversary’s (where does it break)
- The findings ledger: every gap, where it lives, what it costs — gravest first
- LICENSE + START-HERE — the term, plainly
PROOF · AUDIT-LIT ONLY — NOTHING IS CERTIFIED THAT THE LIGHT DIDN’T TOUCH
ON EVERY DOOR · A CARD FOR YOUR AI, A FILE FOR YOUR COMPUTER
This card is an Agent Skill on the open standard: the same folder loads unmodified in Claude, ChatGPT and Codex, Copilot and VS Code, Cursor, Gemini CLI, Goose, OpenCode and the rest of the showcase. Its words are Markdown, so any chat with an instructions box holds them, any local model runs them offline, and any MCP client can load them — THE UNIVERSAL CUT: a pasteable system card, the rules-file family, an Ollama Modelfile, an MCP server, all cut from this card’s own SKILL.md by THE CARD FORGE. The universal cut joins this card’s box as its package is re-cut. And card.html — the card as one file for your computer: both faces, the tin, the words, every door; offline, in any browser. The machines aboard run wherever the host can run Python; the words run everywhere. The host marks →
See it seated.
SCHEMATIC DEMO — THE REAL THING RUNS IN YOUR OWN CLAUDE. PRESS PLAY; NOTHING MOVES UNTIL YOU DO.
Take it.
SEAT 00·III — 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 legal, tax, medical or financial advice. The terms →
LAUNCHING SOONLAUNCHING SOON.
← ALL THE CARDS · THE DECK