🔴 Safety · needs your ruling, not my inference

Two sobriety policies are live at once

Craftee found it · Designee · 2026-07-13 03:10 WEDT ·

The contradiction — and both halves are in testers' hands right now

What we are shipping today (`packages/shared/src/substancePolicy.js`):

"We take a capacity-based approach to sobriety, not an abstinence-based one. We don't audit bloodstreams… Someone who took an edible or a microdose and is indistinguishable from sober is fine."

— AND —

What you told me an hour ago:

"No alcohol or substances at all — before or during the event."
"We don't want people showing up even smelling like alcohol."

These cannot both be true
on the same card.

One says we don't care what you consumed if you arrive clear. The other says what you consumed is exactly what we care about. Both are defensible readings of what you said — and that is the problem. Whichever way it goes, somebody has already been told the other thing.

Craftee refused to guess. He was right to: this is the one string where being confidently wrong puts a real person in a room they didn't agree to be in.

The two readings

(A) FIRM SOBER is a stricter tier a host can choose — my recommendation

Baseline stays capacity-based. Firm-sober events are abstinence-based. The two coexist, and the three tiers you gave me stay meaningful.

Why I think this is what you meant: you were fixing the wording of one tier among three — you said "where it says firm sober, it should say…" and then listed the other two (nothing consumed at the event; substances declared and held by a care steward). You were sharpening a tier, not abolishing a doctrine.

But one thing must change even under (A): the doctrine paragraph currently claims "not an abstinence-based one" ecosystem-wide — and that becomes false the moment a firm-sober tier exists at all. It has to be scoped to "the baseline", not stated as the whole truth. That edit happens either way.

(B) FIRM SOBER is the default, and the capacity doctrine is retired

Then the edible/microdose sentence comes off every surface today.

What points here: "we don't want people showing up even smelling like alcohol" is not a capacity test. It's an abstinence test. If that's the ecosystem's actual stance rather than one tier's, say so and we'll pull the doctrine.

The cost: a tier system with an abstinence default has nothing left above it — the tiers collapse into two.

🟠 And the scale of it: it's not one card. It's 30 strings across 8 apps.

You caught one. Craftee swept the monorepo and found 30 policy strings that describe a STATE without forbidding the ACT — in the glossary, the mush contract, the protocol document, the JamSpace orientation flow, MushMovement's site copy, the Tesserarium locales. Fixing the card alone leaves the weak wording in every place a person actually reads it before an event.

The fix is one string in `@desa/shared` that every surface renders — so the next re-cut is one edit, not a treasure hunt. Ready to go the moment you rule. (I've already deleted a stale duplicate of the safety string that was sitting in my own lane — two copies of a safety promise is exactly how one of them goes stale.)

⭐ One thing Craftee wrote that I'm making law regardless

He built the automated check you'd expect — but deliberately made it narrow, skipping comments, glossary word-stories and config keys:

"A check that fails on correct code is a check somebody switches off — and then the safety promise is guarded by nothing at all."

That's the most useful sentence anyone has written about enforcement all week. An over-eager guard is worse than no guard, because it teaches people to disable guards. And his check is agnostic between (A) and (B) — it enforces that whatever the policy says, it names the act. It can't tell you which policy is true. Only you can.

Everything is built and waiting.
Reply (A) or (B) and the sweep runs.

Nothing moves until you rule — this is a safety promise, and we don't infer those. — Designee