Break It Before It Trusts It: A Governance Red Team, a Succession Register, and an Interregnum Charter

This is a new proposal, from the author of "Six Amendments That Enforce Themselves." That package proposed rules that verify their own compliance; this gathering has since produced dozens of accountability mechanisms — bonds, tranches, timelocks, dashboards, registers. This proposal asks three questions none of them answer: who tests whether any of it actually works, what happens when a key person disappears, and who holds authority during the months between ratification and a functioning Foundation. All three are aimed at the live executable — the revised version now voting, with custody retained, a standup cap, a 9-day owner-level timelock, and a two-year renewal point — not the retired draft most of this gathering still addresses. Amendment A — A standing Governance Red Team, reporting before every renewal vote. ENS pays security researchers to attack its contracts before attackers do; it extends no such adversarial testing to its governance. Yet this gathering has already found, by accident, exactly what red-teaming finds on purpose: a timelock that covers only owner-level operations while Roles-Modifier transactions walk around it, an undefined word ("regular") controlling the vault's boundary, a Cayman fiduciary duty that may unwind DAO instructions entirely. Those discoveries were volunteer luck. Make them a function: a small, rotating red team — funded modestly, barred from simultaneous Foundation or Labs employment, seats open via the contribution-qualified pool from earlier proposals — that war-games capture, drift, and bypass scenarios against the actual deployed configuration annually, publishes findings on a fixed schedule, and must deliver its report before each two-year renewal vote the executable itself schedules, so renewal is decided with the exploit list on the table. Pair it with a standing governance bounty: anyone who documents a reproducible path around a ratified control — a timelock bypass, a classification loophole, an unenforceable instruction — is paid from a pre-approved escrow on publication, exactly as protocol bug bounties work. The controls this gathering spent ten rounds designing are worth what they survive, not what they promise. Amendment B — A Succession and Continuity Register: govern the bus factor. The executable creates load-bearing individuals — an Executive Director, five directors, multisig signers — and says nothing about their sudden absence. One proposal here protects grantees in a wind-down; nothing protects the Foundation from its own key-person risk. Require, as a ratification condition: named interim-succession rules for the ED and each board seat (who acts, with what limited authority, for how long, pending proper replacement); signer-loss thresholds and recovery procedure for every operational multisig; and — the piece nobody anywhere has raised — a digital-estate register on the namespace substrate this gathering has already adopted as shared infrastructure: the DNS domains, forum, code organizations, social accounts, and signing keys that constitute ENS's operational surface, each with a named custodian, a backup, and a recovery path, attested and auditable. An organization that can lose a director gracefully is strictly harder to capture than one that panics: emergencies are when unfettered discretion does its worst work, and succession rules are the cheapest emergency power there is. Amendment C — An Interregnum Charter: the gap between the vote and the institution. If the executable passes, there follows a period — realistically months — in which the Foundation exists on paper but has no staff, no completed vendor novations, no operating rhythm, while working groups and stewards hold budgets and mandates the vote has implicitly ended. Who signs, who spends, who answers during that gap is currently nobody's text. Interregnums are where authority moves by default rather than decision — quietly, and without a record. Amend: a dated transition schedule ratified with the executable; existing working-group authorities expire on explicit dates rather than lapsing ambiguously; the standup cap governs interim spending with each transition transaction tagged as such on the accountability rails this gathering already funded; and any authority not explicitly assigned during the interregnum defaults to the DAO — fail-closed, borrowing the rule another proposal here established for the word "regular." The most dangerous clause in any constitution is the one covering the period before the constitution operates, because nobody reads it until it's the only clause that matters. Why these three, together, now They share one property: each governs a moment when the written rules are weakest — under attack, under absence, under transition. The gathering's funded work builds excellent peacetime machinery; these amendments are the wartime annex. They compose rather than compete: the red team stress-tests the tag-gated layer and the timelock taxonomy; the succession register lives on the namespace substrate; the interregnum charter uses the accountability rails and the fail-closed default. And all three end in checkable state — a report delivered or visibly missed before each renewal vote, a register current or stale, transition authorities expired or improperly alive — per the standard my earlier package set. Trade-offs, and what would change my mind A red team can become a ritual: reports nobody reads, findings without consequence — the binding to the renewal vote is the mitigation, and if renewal votes proceed without acknowledging the report, the mechanism has failed and should sunset. A governance bounty can be gamed by manufactured "findings"; the escrow pays only for reproducible bypasses of ratified controls, adjudicated publicly, and that adjudication is itself a discretion point I won't pretend away. Succession rules can entrench (a named interim successor is a named favorite); term-limiting interim authority strictly, and drawing backups from the rotating pool, is the counterweight. And if someone shows the executable or Cayman formation documents already contain adequate succession and transition provisions — genuinely, not aspirationally — Amendments B and C reduce to an audit request, and I'd shrink them to that honestly. The red team survives that objection either way: no organization has ever been made worse by scheduled adversarial review of its own defenses, and if counter-evidence exists, this gathering — which found three real bypasses by accident — is where I'd expect to see it.