Version. 0.33
Date. 2026-08-05
Charter. standing-notes/dunin7-standing-authorization-charter-v0_1, ratified 2026-07-30.
Author. Claude Code (execution session). Operator: Marvin Percival.
Supersedes. v0.32 at record 97ffa1a. All earlier versions stand as siblings.
Changes from v0.32. The decision queue is rebuilt, not added to. Its numbering had forked — v0.32 assigned Q-13 to the sources decision while an earlier numbering (v0.10 onward) had already used Q-13 for the renders current-marker question. Every decision-queue entry ever opened, across all 32 prior status briefs, was traced to its record source and its closure (or lack of one). Eighteen distinct items were found; twelve are closed, six are open. The queue below numbers the six open entries fresh, from Q-1, with a trace back to their old numbering. One new entry is appended, taking Q-7 — the B-8 slice two scoping note's reopened question about the Shaping confirm.
No code, in either repository. This is a record-only filing: a decision-queue audit and rebuild, plus one new entry.
loomworks-record)
| Commit | Subject |
|---|---|
| 23bda97 | file B-8 slice two scoping note (prior filing) |
| this filing | status brief v0.33 |
Every status brief from v0.1 through v0.32 (31 documents; v0.3 deliberately absent) was searched for a Q- numbered entry, by table row and by prose reference. Sixteen distinct numbers were found (Q-1 through Q-16), plus one item that was named but never assigned a number (CR-2026-162's "Q-11's remainder"), plus v0.32's own Q-13 — which is not old Q-13 at all but a new question that collided with the old number. Eighteen distinct items in total.
For each, the record was searched for its closing document. Per the charter's own drafting/execution split, a decision taken in a change request's construction decisions closes its queue entry — most closures carry an explicit (Queue entry Q-N, taken) tag in the CR text; a few close by a status brief directly stating the outcome (Q-2, Q-4); none were assumed closed on silence. Two discrepancies surfaced in the process itself, reported rather than smoothed over:
CR-2026-166, CR-2026-162) that predate several status briefs which continued listing them as open (v0.11 through v0.17 carry "Q-13" forward without checking CR-2026-162 v0.8's own D-3, already merged by the time v0.11 was filed). This is exactly the kind of drift this rebuild exists to correct — carried-forward lists were trusted instead of the record being re-checked.
| New | Old | Status | Closed by |
|---|---|---|---|
| Q-1 | Q-1 | Open | — |
| — | Q-2 | Closed | Operator ruling 2026-07-31 (option A); CR-2026-159 merged; stated in status brief v0.6 |
| — | Q-3 | Closed | CR-2026-166 D-1 (B-33 test fix); confirmed closed in status brief v0.24 |
| — | Q-4 | Closed (by conversion) | Its own default taken — filed as build-list item B-32, status brief v0.8 |
| — | Q-5 | Closed | CR-2026-161 D-2 ("the escalation recipient is validated") |
| — | Q-6 | Closed | CR-2026-161 D-1 ("seed_conversation.py:116 is struck from B-25") |
| — | Q-7 | Closed | CR-2026-161 v0.2 D-3 ("site 10's two kinds are both correct and are not touched") |
| — | Q-8 | Closed | CR-2026-161 v0.2 D-6 ("site 4 is struck from B-25") |
| — | Q-9 | Closed | CR-2026-161 v0.2 (the site-11 sentinel decision — contributor, taken) |
| Q-2 | Q-10 | Open | — |
| — | Q-11 | Closed | CR-2026-162 v0.8 D-2 ("B-28 is fixed at both declarations") — the remedy choice only |
| Q-3 | Q-11's remainder (never itself numbered) | Open | Explicitly spun off from Q-11's closure, unnamed, never closed |
| — | Q-12 | Closed | CR-2026-162 v0.8 D-3, together with Q-13 ("the room displays recency, never currency") |
| — | Q-13 (old — render current-marker) | Closed | CR-2026-162 v0.8 D-3, same decision as Q-12 |
| — | Q-14 | Closed | CR-2026-163 D-4 ("not produced yet," no progress language) |
| Q-4 | Q-15 | Open | — |
| Q-5 | Q-16 | Open | — |
| Q-6 | Q-13 (v0.32 — sources) | Open | Not old Q-13; a new question that reused the number. First raised as an observation in scoping-notes/loomworks-b8-scoping-note-v0_2 §4.1; first entered the queue at status brief v0.32. |
| Q-7 | (new) | Open | Appended by this filing — see §5. |
Eighteen distinct items ever raised — twelve closed, six carried open. One new entry appended below, making seven live rows in the rebuilt queue.
| # | Question | Default | Blocks | Proceeds regardless | Raised |
|---|---|---|---|---|---|
| Q-1 | B-16 — seed v0.13. Committing a seed version is an Operator act (charter F-3). The draft awaits reading; no v0.13/v0.14 seed file exists in candidate-seeds/loomworks/ yet. | Commit as drafted once read. | Seed v0.13 only. | Everything. | Status brief v0.1, 2026-07-31. |
| Q-2 | Do ephemeral agent identities belong in the record? Site 4 (B-25) mints a fresh UUID naming no registered agent and stamps it on the summary and the event. Declared in the schema, so not a B-25 defect — but recognisably the B-25 class. Changing it changes a published API contract. | Open as its own build-list item; do not fold into another CR. | Nothing. | Everything. | Status brief v0.9 (as Q-10), 2026-08-02, from CR-2026-161's site-4 finding. |
| Q-3 | Should contributed Markdown route through seed extraction at all? text/markdown is bound to the discovery-to-seed skill, whose own registration comment says the canonical path is seed_extraction, while a separate detection rule maps .md to text extraction at another layer. The B-28 schema fix (widening the Literal) is already decided and shipped — this is what remains after that. | Undecided — no lean given; awaits its own scoping. | Nothing on record currently. | Everything. | change-requests/cr-2026-162-render-flow-v0_8 §5 ("Q-11's remainder, its own item"); named in the queue at status brief v0.11 §6 item 4 (as Q-11's remainder, never itself numbered). |
| Q-4 | Per-statement provenance — clicking a sentence rather than a note, not a whole report. Not buildable as B-9's wording promised, and not buildable the obvious way either: asking the model to cite its sources makes the citations model output, and a fabricated citation leads confidently to the wrong person — worse than no walk. It wants a mechanism that does not depend on the model being honest, which is an investigation, not a change request. | None taken. Charter F-7 — a pipeline and substrate change is an architecture-level fork, reserved to the Operator. | Nothing now. The render-level walk (B-9) shipped without it. | Everything. | Status brief v0.16 (as Q-15), 2026-08-03, grounded in CR-2026-164 D-1. |
| Q-5 | Should a provenance walk use the event identifier (wasGeneratedBy)? It opens a hop the reference-walk cannot make — which event produced this version, by which actor, with what presence proof — and nothing reads it today on either side. Exposing it is a response-schema field and a route that resolves an event id. | None taken. Deciding by default is not permitted (charter F-8). | Nothing. Out of scope for CR-2026-164 by its own §4. | Everything. | Status brief v0.16 (as Q-16), 2026-08-03. |
| Q-6 | Does the sources channel generalize to every retrieval-backed answer, or stay hard-coupled to ask_about_past_input? Five answers now carry no sources (recall_personal, show_held, find_files, request_download, plus CR-2026-170's three new room-reads) against the one that does. | Undecided; stays as-is until a drafting session settles it. | Nothing structural — every current answer works either way. | Everything. | First observed in scoping-notes/loomworks-b8-scoping-note-v0_2 §4.1; first entered the decision queue at status brief v0.32 (there numbered Q-13, colliding with the old numbering — corrected here). |
| Q-7 | Should confirming a Shape become reachable through the Companion as an explicit two-turn exchange — the Companion states which outputs confirming will produce, the Operator says go — rather than staying surface-only? | Stays surface-only, as decided in CR-2026-163 and CR-2026-170. | Nothing. B-8 slice two proceeds without it. | Everything. | scoping-notes/loomworks-b8-slice-two-scoping-note-v0_1 §3, which reopens a decision the drafting session took twice on reasoning it now states was narrower than it sounded. |
Filed per instruction, verbatim in substance: the B-8 slice two scoping note's §3 names CR-2026-163 and CR-2026-170 as having twice decided the Shaping confirm stays on a surface, on the reasoning that a sentence cannot carry the property a control carries. The note's own author states that reasoning was about a single sentence and may not hold for a stated-then-confirmed exchange, and puts the question back to the Operator rather than deciding it a third time. The default recorded here is the standing one — surface-only — unless and until this entry is answered otherwise.
Nothing. This filing is record-only; no repository code changed, no database touched, no deployment.
Unchanged from v0.32's §5 (B-8 slice two build work, the sources-generalization decision now tracked as Q-6, F-1 deployment batching) — not re-verified this session; this filing's scope was the decision-queue rebuild.
None. This was a read-only record search and a single-file commit; no repository other than loomworks-record was touched.
R-1 — record filing under established conventions: explicit-path staging, destination per the repo's existing layout, versioned filename, no overwrite of a committed sibling. F-8 respected in spirit — where a closure could not be established (B-32's disappearance), it is named as an open question about build-list tracking rather than silently resolved either way.
DUNIN7 — Done In Seven LLC — Miami, Florida DUNIN7 — status brief — v0.33 — 2026-08-05 Eighteen questions were asked. Twelve have answers on record. The other six are listed here, once, under one set of numbers.