[register] Owner-decision backlog — 17 issues blocked on a ruling, not on effort

[hyperpolymath]

Why this exists

Triage of all 65 open issues on 2026-08-25 found that 17 are blocked on an owner decision, not on effort. Each question is currently buried somewhere in a long thread, which is why several have sat for months while looking like ordinary backlog.

This is the register. One line per decision, the question stated as a question. Answer them here or in the linked issue — either unblocks the work.

Four were resolved during the 2026-08-25 session and are recorded at the bottom as precedent.

Open decisions

Destinations and structure

#	The question
#493	Five satellite destinations named in .gitmodules do not exist, and two entries are aliased (path ≠ url repo). For each: recreate, redirect to a successor, or accept the vendored copy as the survivor and delete the claim? (Owner has directed base-form rationalisation instead — see #634; this asks only about the dead entries.)
#462	DYADT — implementation found orphaned inside patallm-gallery, and it diverges from the authoritative spec. Mint the repo and reconcile, or rewrite against the spec from the reference verifier? See #635 child 1.
#403	hyperpolymath/manifesto exists but its root carries no doctrine — only RSR scaffold. Checkbox 1 is "Owner edits doctrine to his voice". Nobody else can write this.

Enforcement and protection

#	The question
#408	Approve the canonical merge-options / branch-protection shape, and rule the governance / * → new-context rename mapping for ~190 repos. Branch-protection writes are owner-gated.
#306	Per repo across 48: enable Pages, or delete the Pages workflow? Pages enablement is API/UI-only and not workflow-controllable, so there is no implementable middle path.
#343	Which repo is canonical for .editorconfig / .gitattributes / .gitignore? standards and rsr-template-repo still differ on all three. And how much per-language .gitignore divergence is legitimate?
#342	Force the-nash-equilibrium from contractile schema v1.0.0 to v2.0.0 (adds two files per verb = new work), or bless it as a permanent holdout? SPDX-headered files are owner-only.
#446	Approve the readiness/insight badge criteria and axes, and name one pilot repo, before any engine is built.

Credentials and provisioning

#	The question
#399	When is the credential behind the 40 unbaselined secret findings revoked and purged? Nothing else keeps the gate red; the issue closes on its own afterwards. (The original premise is dead — .hypatia-baseline.json is now 46,370 bytes / 118 entries, not the 3-byte empty file.)
#443	Provision SonarCloud — one global token, auto-provision projects — and set a resume date. Also secure admin on the 3 shared repos. Self-parked since 2026-06-29.

Scope and co-design

#	The question
#245	Which plugin hosts to bind at all — WordPress / WebExtensions / Thunderbird MailExt / React-Next — given WP-PHP cannot load WASM? Four questions posted 2026-05-28, none answered. Blocks #246, #280.
#438	Provide the authoritative protected-repo list and authorise a full-access recovery session. Confirm scope of revert-vs-keep-as-.md.
#95	Narrow this tracker to the governance-audit items and close the rest as discharged, or close it entirely and re-file? Its own last comment poses exactly this.
#490	✅ Resolved — a2ml product → existing a2ml-ecosystem; spec incl. .scm variant stays in standards. Retained here for the audit trail.

New, arising from the 2026-08-25 analysis

Ref	The question
#636	Does a minting workflow get to create repos unattended? This crosses the standing "no unattended mutations" guardrail. Phase 1 can be workflow_dispatch-only and still deliver enrollment — but full birth automation needs an explicit ruling.
#634	Build scaffoldia's composer, or revive the dormant Haskell engine? Two halves of one tool are being maintained; neither mints a repo today.
#633	After a week of reviewed decision manifests, wire actuate.sh? Owner has ruled observe-only for now; this is the deferred second half.
⚠️ new	gitbot-fleet's local checkout has a corrupt packfile and 3 commits that exist nowhere on GitHub — including 632b54a "fix: recover phase-zero CI and governance". Filed separately; needs a keep-or-discard decision before anyone tidies that directory.

Resolved 2026-08-25 — precedent

Question	Ruling
LANGUAGE-POLICY.adoc §1.2 vs .claude/CLAUDE.md — TypeScript or AffineScript?	AffineScript governs. §1.2 was the error. Bun stays tier-1 runtime; the two questions were being run together. PR #632
Wire the merge actuator?	Observe-only first. observe.sh shipped; actuate.sh held pending review
Where does the a2ml product go?	a2ml-ecosystem (exists); spec stays in standards
estate-audits — new repo?	No. Audits go to developer/dev-notes/, above the repo trees
verisim-data — retire?	No. It is verisimdb-data, it is live, and 20+ hypatia modules read from it
git-logistics-office — what is it?	The estate control plane. Chartered in #636

---

The #239 TypeScript/AffineScript ruling is the reason this register exists in this form: I had triaged that campaign family as close-as-stale on the strength of LANGUAGE-POLICY.adoc, and asking rather than acting kept four live issues from being wrongly closed. Decisions that look like backlog are not backlog.

[hyperpolymath]

⚠️ Corrections to this register — re-verified live, 2026-08-26

This register was filed 2026-08-25 and restates each parked issue carrying its original numbers forward. Thirteen of those issues have now been re-verified against the live API. Only one of thirteen had its central factual claim survive unchanged.

Please do not rule on the figures below as written. Corrected verdicts are posted on each issue; summarised here.

A denominator problem underneath the whole register

The estate is 420 repos, not one owner: user:hyperpolymath = 384 (43 private, 4 forks) plus org:metadatastician = 36. Several repos this register names as hyperpolymath/<x> now 302-redirect into the org — burble, paint-type, stapeln, enaction-engine, phi-LAM — and hyperpolymath/007 redirects to 007-lang.

Every prior estate audit that enumerated a single owner undercounted.

Two tool-reach traps that each produce a confident wrong answer here, both hit during this pass:

• gh api prints 404 bodies to stdout, so a naive 2>/dev/null census scores every repo as "protected".
• /commits/HEAD/check-runs caps at 100 per page — without --paginate it manufactures false "stale context" hits.

✅ Closes outright

#245 — plugin-host bindings. All five target repos now contain 0 TypeScript files (#246 inventories 65); the named blocker affinescript#103 is CLOSED. Closed.

⚠️ Numbers wrong enough to change the work

issue	register says	measured
#408	governance / * stale on ~190 repos, cause = job rename	1 real defect (affinescript requires lowercase baseline). The job name never changed. 27 of the 36 non-emitting repos are startup_failure — a different programme. Classic branch protection exists on 1 of 420 repos
#306	48 repos need Pages	40. And "failing every push" is now mostly a mass cancellation event of 2026-08-24; some of the 40 fail before Setup Pages, so enabling Pages will not fix them
#342	8 repos, 3 findings	42 repos. 2 of the 3 findings are FALSE (echidna and burble are both fixed). Four schema versions exist, not two
#493	"META.scm, ECOSYSTEM.scm, state.scm are files, not projects"	FALSE — each is a full project tree; ECOSYSTEM.scm carries an IETF draft and IANA considerations. "Delete the entry" is the wrong fix. Also: .gitmodules declares 12, disk holds 10
#403	"manifesto root carries no doctrine — only RSR scaffold"	FALSE — README.adoc carries the manifesto in the first person, plus an assurance-turn section via merged PR #67. The register is staler than the issue it summarises
#646	899 files carry the AGPL line	646. And 253 of them name PMPL-1.0-or-later, a retired licence — a second defect the issue does not mention
#443	SonarCloud rollout ~unstarted	partially done, but not by the campaign — SonarCloud is auto-analysing repos with no properties file and no workflow. This register never mentions metadatastician at all
#399	~40 unbaselined secret findings gate it	0 open / 52 fixed. What actually keeps main red is governance.yml at startup_failure — now diagnosed (wiped Actions allowlist)
#95	blocked on burble#35	that issue and its three successors are all CLOSED; all three stragglers now call governance-reusable.yml. Closeable
#438	protection via a do-not-migrate label	no such label exists anywhere in the estate; protection was a Hypatia rule allowlist. Both named protected zones are undamaged

🔴 Genuinely yours, premise intact

• audit: dotfile drift across estate (2026-06-02 sample) — .editorconfig / .gitignore / .gitattributes #343 dotfile canon — premise holds and drift has worsened (.gitattributes 6 → 14 variants since June). Three options posted on the issue.
• Proposal: an honest, adoptable readiness/insight badge framework — grade our own work, and challenge the evaluators #446 badge framework — real decision, but its stated precondition is false: the Glama badge is still live in BoJ's README. Four options posted.

⛔ Blocked on execution, not on you

• Implement the DYADT production verifier (hyperpolymath/did-you-actually-do-that) #462 DYADT — target repo 404s; a 1,168-blob implementation exists in patallm-gallery with a different verifier taxonomy. Already retargeted by [umbrella] Surface the buried projects — DYADT is the worked example, the detector is the deliverable #635; suggest parking behind it.
• Build the Estate Manifesto & Atlas (front-door, owner-gated) #403 manifesto — Part I done, Part II (the Atlas) must be written; the cited source draft does not exist, so it cannot be recovered.

Suggested reshape of this register

Of the ~15 rows here, 1 closes, 10 need their numbers replaced, 2 are real decisions, 2 are execution-blocked. That is a materially smaller decision queue than it currently looks.