Wiki QA Decision Register — July 2026

Full Q&A review of the wiki (13 Jul 2026): every page in concepts/, entities/, sources/, decisions/, analysis/, projects/, _audit/ plus the meta files was read in full and cross-checked against the rest of the vault. This register lists everything that needs a ruling, a verification, or a canonical-fact decision before the wiki can be trusted as source of truth again. Straight factual corrections (no decision needed) are listed at the end as a fix list.

Headline: the wiki’s April 2026 snapshot was high quality, but maintenance largely stopped mid-April (log.md is empty 14 Apr → 7 Jul). The 6 Jul Xero audit and the July staffing departures have overtaken a large share of the finance and operations pages, and several April “decisions” were never executed or never re-examined. The vault currently asserts, in different places, opposite states for: the books’ reliability, PL insurance, Friday trading, the primary bank, the licence pathway, and the kitchen’s approval status.

Section	Items
A — Business decisions awaiting a ruling	14
B — Facts needing verification (reality check, not opinion)	12
C — Canonical-fact conflicts (pick one, propagate)	8
D — Wiki process & housekeeping decisions	8
E — Straight corrections (fix list, no decision needed)	18

---

A — Business decisions awaiting a ruling

A1. Post the Xero correction pack — and who approves the write-off? Approved April 2026 ([[Xero Corrections Approved]]), but the 6 Jul audit proved none of the four journals ever posted; drift has compounded (account 901 grew $11,575 → $18,520). decided_by is “unrecorded” on a decision that includes a $3,251.25 P&L write-off + 11 account archivals — write-offs are meant to go through Mat and there is no record he approved. Decide: post as-is (re-derived to current balances) or re-approve first; get Mat’s sign-off on record. Owner: Shae → Mat.

A2. Amaka repair or replacement (Square Balance bleed). Amaka’s bank-leg posting stopped ~Dec 2025; Square Balance bleeding ~$55–67k/month (−$156k). [[Daily Sales Sync]] and [[Settlement Sync]] still describe the pipeline as “operational”. Decide: fix Amaka config/subscription vs switch to LinkMyBooks (evaluated Apr, 4.98/5, ~$62/mo, AU GST support — never ruled on). Owner: Shae.

A3. Superannuation funding plan. $49,012.53 super payable unfunded, holding account empty; directors personally liable for SGC. Audit calls it “the single most dangerous liability”. No recorded plan or owner since the 6 Jul audit. Decide: funding schedule + who takes it to Mat. Owner: Shae → Mat.

A4. May/June COGS catch-up. June COGS collapsed to $305 (bills unentered); FY26 P&L “NOT usable for decisions” until fixed. No recorded owner. Owner: Shae.

A5. Corporate structure reform — Mat’s ruling + legal advice. [[Corporate Structure Reform]] has sat at “pending Mat review + legal advice” since the review scheduled for 21 Apr. The s 113(1) breach carries criminal penalties and blocks the entire capital raise and NFP-only grants ($100k+/yr). Highest-consequence open decision in the vault. Gate on B1 (entity identity) first — it could change the premise. Owner: Mat (+ legal).

A6. Public liability insurance — deferred or urgent, and at what real cost? The decision page says “URGENT — complete within 7 days” at $2,000–$8,000 p.a.; the concept pages say the real quote was $142,890–$157,179 (18 of 19 insurers declined) and procurement was “deprioritised per Shae’s direction”. Both can’t stand. Decide: record the deferral formally (with a revisit trigger) or reactivate procurement; correct the decision page’s cost by ~20–70×. Owner: Shae → Mat.

A7. Food Safety Supervisor — who, now Monique has resigned? Monique was the nominated FSS replacing Lorelei Gardner. Her resignation silently reverts the nomination to unfilled; Class 2 registration → kitchen legality → licence pathway all queue behind it. Options: Mat certifies (SITXFSA005/006, he flagged willingness), Charlene Poole (relationship undocumented — see B8), or a certified casual. Owner: Mat.

A8. Licence reclassification — submit the On-Premises application? The R&C → On-Premises correction was made in April but never captured as a decision page, and there is no evidence of a submission. It is the single largest cost lever ($104k–$307k/yr security) and On-Premises does not require the kitchen first. Decide: submit now vs wait; record it as a decision page. Owner: Mat + Shae.

A9. Debt restructuring pathway — never ruled on. [[Debt Restructuring Options]]: business “structurally cash-flow negative while the Lumi facility is in place” ($357k debt, ~$111k/yr service); pathways save $40–68k/yr; no decisions/ page, Judo Bank never engaged. Blocks capital raise per the page’s own relationships. Owner: Mat + Shae.

A10. Rostering/payroll platform — urgent post-Tom. Deputy→Xero manual bridge lost its only operator (Tom, undocumented spreadsheet). Evaluations complete (Tanda / KeyPay / RosterElf / Microkeeper) but no choice made. This is now an operational continuity risk, not a nice-to-have. Owner: Shae.

A11. Loyalty platform — PinTuna’s fate. Stated three incompatible ways across the vault: live system / “eliminated” / “under review for gift cards” — and July research reframes paid memberships as a live co-op-story option. ~$2,328/yr; loyalty component never activated (0 customers). Decide: keep-First-tier / drop for Square Loyalty / expand to memberships. Owner: Shae → Mat.

A12. Remaining tooling calls (batch). Dangling evaluations with no ruling: Fathom financial reporting ($65/mo); Hiver vs native Google Groups shared inbox; Mailchimp for email marketing (confirm not moot — 7 Jul found a recommended email structure “already existed”); Plausible ($19/mo) for website analytics (from a never-ingested source — see D4). One sitting, four small calls. Owner: Shae.

A13. Accounting firm tender — trigger check. Decision is “tender after lodgement completes”. Whether lodgement has completed is recorded nowhere; outstanding bill drifted $8,000 (Apr) → $5,695.39 (6 Jul audit). Decide: confirm trigger status with Collins & Co and set a tender start date. Owner: Shae.

A14. Google Business Profile + QuickSuper access. GBP: Tom was primary manager, no documented successor — reassign. QuickSuper: confirm an active clearing-house user exists before Owen’s OWENPH1 login is disabled (open loose end). Owner: Shae/Mat.

---

B — Facts needing verification (check reality, then record it)

B1. Entity identity — Pty Ltd or public Ltd? Constitution names “Pride of our Footscray Ltd (ACN 618 613 095), unlisted public company”; the operating entity everywhere else is ”…Community Bar Pty Ltd (ABN 33 621 811 372)”. If the shareholding entity is already a public company, the s 113 breach premise (and the co-op recommendation built on it) may be wrong. One ASIC lookup — cheap, highest leverage in this register.

B2. Kitchen status as of July 2026. The 27 Mar visit was an inspection with an action-items list, not an approval (the decision page overstates it as “approved”). Final inspection passed? Registered? Operating? ~12 pages describe a four-month-old state.

B3. Theatre Restaurant Model — launched? Confirmed 11 Apr, implementation “upon Mat’s return 21 Apr”; no recorded launch or Saturday revenue result anywhere.

B4. Friday trading — active or cancelled? [[Revenue Model]] and [[Cash Forecasting]] say “Friday nights currently cancelled”; [[Drag Bingo]], [[Programming Calendar]], [[Profitability Analysis]] treat Friday as the reliable $6,000 flagship. Likely “club nights cancelled, Bingo continues” — confirm and fix whichever side is wrong.

B5. Monique’s director / Pillar Member status post-resignation. The draft constitution ties Pillar protection to shareholding, not employment — resignation may not remove her directorship. Affects board composition, quorum, and the still-live “confirm Monique as director” agenda item. Ask Mat.

B6. PL insurance history. Was the $157k policy ever bound, did it lapse, or was it never taken? [[Insurance Crisis Timeline and Status]] builds sustainability maths on paying it; [[Insurance Gap - Public Liability]] says uninsured. One authoritative line, all insurance pages point at it (pairs with A6).

B7. FY24/FY25 returns + income-tax due date. [[Tax Lodgement]] contradicts itself (30 Jun vs 30 Oct 2026) and still flags prior-year returns “verify filed”. Confirm with Collins & Co.

B8. Charlene Poole’s formal relationship (staff/contractor/volunteer) — undocumented; she’s a candidate FSS and was co-leading kitchen setup.

B9. Website DNS cutover — recorded 10 Jun as “pending launch-day action”; no later record. Did it happen?

B10. Mat wiki viewer adoption — deployed, but no record Mat uses it; sync drift (10 days) flagged 24 Apr, unresolved.

B11. Security cost baseline — $2,000/wk vs $5,900/wk (Mat Input Q10); needs invoice verification before any saving figure is standardised (pairs with C5).

B12. Mat Input Required — remaining open questions. Of the 20 questions (Apr 2026), only Q6 (staffing restructure) is resolved. Still open: supplier register (Q4), Fernando Cardoso terms (Q8), banned-persons list formalisation (Q9), bartender pay rates vs HIGA (Q11), board formalisation (Q12), constitution ratification + original prospectus (Q13), CEO time allocation (Q14), event-level P&L data (Q15), TryBooking consent / Spam Act (Q16), POOF Ltd ASIC status + loan terms (Q17), private-function process (Q18), revenue recognition basis (Q19), incident logging (Q20). Schedule a working session with Mat.

---

C — Canonical-fact conflicts (pick one value, propagate everywhere)

C1. Primary bank of record — Westpac or CBA 2203? Entity layer says Westpac primary and even labels account 2203 as Westpac; projects/finance.md, [[Bank Reconciliation]]’s own routing, the 6 Jul audit ($5.42M of Square payouts into 2203 across 1,535 transfers), and workspace ground truth all say CBA 2203. Note: pride-turnaround CLAUDE.md still says “Westpac (primary)” — resolve at source, then correct wiki + CLAUDE.md together.

C2. Books-reliability narrative. ~10 pages still assert “phantom balance resolved / Xero reliable / clean statements” ([[Bookkeeper Role and Financial Management Capacity]], [[Daily Sales Sync]], [[Settlement Sync]], [[Capital Raise Blockers]], [[Compliance Obligations]], entities Square POS/Xero/Shaemus Corcoran, [[Mat Meeting 11 April 2026]]). The 6 Jul audit is the truth. Decide the correction mechanism: audit-aware banners vs body rewrites.

C3. Shareholder numbers. 200 vs 206 vs 207 holders; “~80” vs “~189” external; “764 issued” $500-class shares impossible against 662.02 total; certificate counts (231/215/105) don’t reconcile. Pin canonical figures from the registry tabs; everything (s 113, capital raise, co-op) cites them.

C4. Saturday baseline + root cause. Baseline: $20,000+ (narrative pages) vs $16,630 peak / $12,263 typical (data-derived [[Saturday Trading Pattern]]) vs $15–18k ([[Saturday Turnaround]]). Root cause: timing mismatch (crowd exists at 1am+) vs footfall shortfall (r = 0.05 tickets↔bar). The turnaround maths and the Performer Scheduling $260k/yr thesis depend on which reading wins. Recommend the data page as canonical.

C5. Security cost/saving figure. Current cost quoted $62k–$104k/yr; post-reclassification outcome ranges from “reduce to $15–30k/yr” to “eliminate entirely ($104k–$307k/yr saving)”. Pick the On-Premises figure set after B11 verification.

C6. SQ-600000 / Zeller magnitudes. Zeller misclassification is ~$208,965 (FY25) — but Forensic Audit Run3 and two entity pages mis-state $65,829 (the net balance) as the finding; [[Other Payment Clearing Issue]]’s components don’t sum ($573k ≠ $396k) and are superseded by the audit’s $45,007.35. Reconcile before anything feeds BAS/ITR.

C7. Highest-attendance event category. Three pages each claim “highest per-event average”: Live Music 143.8 (correct of the full categories), Themed Nights 114.1 (wrong — it’s highest volume), Fundraisers 149 (2-event sample). Settle the wording convention.

C8. Roles and titles. Mat: “Chairman” (one page only) vs Founder/CEO. Shae: “pro bono bookkeeper ~1hr/wk” vs “~10 hrs/wk” (same page!) vs project record “Director, Business Manager”. Merge [[Board Composition]] into [[Board Composition and Meetings]] (duplicate pages, divergent details) and settle both titles there.

---

D — Wiki process & housekeeping decisions

D1. Index rebuild + count source of truth. Master index missing 6 pages (Lumi Refinancing Loan, Single-Point-of-Failure Leadership, Website DNS Cutover, DT’s Hotel, Wiki Sync Bot, Events Page Discovery Section Removal); page count stated 6 different ways (262/259/265/263/127/131), all wrong vs 329 files on disk. Options: manual rebuild now, and/or generate counts from disk in the audit script; drop hard-coded counts from prose.

D2. log.md — restart or retire. Empty 14 Apr → 7 Jul; the whole window in which April’s decisions were meant to execute has no record either way — the root cause of most staleness found here. Options: recommit per-session logging (hook it), derive a changelog from git, or formally demote log.md.

D3. Audit script scope. The 3 Jul run silently dropped the frontmatter + orphan checks (June’s run had them); and the script enforces decided_on/decided_by that WIKI-SCHEMA never documents. Restore full checks; document the decision fields in the schema.

D4. sources/inbox-2026-07/ triage. 29 raw files, zero frontmatter, un-logged drop. 25 are retained raws of already-promoted summaries — but they are now the only surviving raw copies (uploads/ is empty, research/ mostly emptied), so 20+ summaries’ sources: citations dangle. 4 files were never ingested: Plausible-vs-Fathom web analytics, venue-intelligence dashboard spec (757 lines, relevant to Pride-Dashboard), investor-facing website design, website tech-stack rec. Decide: (a) re-point summary citations at a permanent raw archive location and move the inbox there; (b) ingest or route the 4 orphans (the two website ones likely belong to pride-website/pride-intelligence repos).

D5. Google Drive review program — live or parked? Designed 7-phase pipeline over 80 shared drives (incl. Board, Legal, Finances, Capital Raise 2026); stalled since 23 Apr on a Drive-MCP recursive-enumeration blocker; catalogue/gap-map/skip-list are empty templates. Fix the access path, switch approach, or mark it formally paused.

D6. Page merges (one fact, one place). Board Composition → Board Composition and Meetings (see C8); [[Square Cash Clearing Issue]] (near-empty stub) → [[Other Payment Clearing Issue]]; decide the canonical home for Lumi facts (entity vs concept both carry full tables) and check Deputy the same way.

D7. Superseded-analysis handling. [[Saturday Revenue Recovery]], [[Licence Reclassification Financial Case]], [[Strategic Sequencing — Kitchen Before Capital Raise]] all carry correct April banners but bodies still argue the dead R&C pathway / stale 48%-no-email figure. Also [[Saturday Turnaround]] and [[Strategic Plan]] still recommend R&C with no banner at all, and QA-PASS-12-APR-2026 reads as current. Decide: rework bodies vs mark-superseded-and-freeze.

D8. Delete empty stray files. Untitled.canvas, Untitled.base, Untitled 1.base, 2026-06-12.md — all empty Obsidian cruft (three created the day of the 12 Jun git-reset incident). Destructive, so listed here for explicit go-ahead rather than done.

---

E — Straight corrections (no decision needed — fix list)

Status update (13 Jul 2026, later same day): items 1–14 and 17–18 APPLIED (see log.md). Exceptions: item 15’s “missing sources: fields” was NOT done — provenance can’t be back-filled without knowing the actual sources (the schema-format parts of 15 were applied); item 16’s Humphrey-vs-Pride-Dashboard naming was NOT standardised — on inspection this is ambiguous, not mechanical (the workspace has separate pride-intelligence and Pride-Dashboard repos, so they may be two different apps) — treat it as an additional Section C question: C9. Are “Humphrey Intelligence App” and “Pride-Dashboard” the same product or two? The Friday-cancelled text in Revenue Model/Cash Forecasting was deliberately left pending B4.

1. Landlord Relationship and Lease Terms L23: address “321 Hopkins Street” → Level 1, 86–88 Hopkins St (only file with the wrong address).
2. Humphrey Intelligence App L42: naming rationale is fiction (Bogart was not LGBTQIA+/Australian/in Breakfast at Tiffany’s) — syncs to Mat’s public viewer; fix or delete the line.
3. Fundraiser Events: Frocktober is ovarian (OCRF), not breast cancer (×2).
4. Pride Comedy: Thursday references wrong — it’s the Week-3 Wednesday slot per Programming Calendar.
5. Insurance Crisis Timeline L277: inquiry report due 27 Oct 2026 (not 30 Oct).
6. Tax Lodgement L44: “FY26 return due 30 June 2026” — wrong on its face; L104’s 30 Oct is the operative claim pending B7.
7. GST Treatment: BAS owner “mat” → Shae; delete “filed by 30 June deadline”; verify monthly-vs-quarterly BAS.
8. Stakeholder Trust and Credibility + Strategic Sequencing: purge residual “48% no email” (actual ~96% coverage).
9. Cash Forecasting: Saturday 50% vs 60% of weekly revenue — internal contradiction; and “bad weeks ~$20,000” vs “$9k worst week” elsewhere.
10. Bar Operations L87: “$15,000 Saturday vs $1,500 Wednesday” are revenue targets, not wages.
11. TryBooking Integration: claims attendance is captured — contradicts TryBooking Attendance Data Loss; also predates the working API (2 Apr).
12. wiki-sustainability-roadmap L112: “Shae’s viewer” → Mat’s viewer.
13. Financial Reporting L94: “kakaoke” typo; owner tables still name Tom.
14. index.md: duplicate ”### Governance Sources” header; two title-resolved links ([[WEB-AG-5 QA Findings]], [[WEB-META-1 CLAUDE.md Level-Up]]) worth normalising to filenames.
15. Frontmatter/schema sweep: ~12 entity pages missing H1; missing sources: on ~7 concept pages; malformed multi-link relationship values (Sunday Market, Succession Planning, Community-Owned Venue Economics, etc.); Nightclub-Only Model still lists R&C as a live contradicts: alternative.
16. Notation: standardise Square account codes (SQ-600000 style) and kill “000-000-00000000”; standardise “Humphrey Intelligence App” vs “Pride-Dashboard” naming.
17. Grant and Funding Eligibility: strip expired “Urgent: Apply Now” deadlines (all pre-13 Jul); re-date the calendar.
18. Stale hard-coded page counts in WORKFLOW.md (“127”), INGEST-PROMPT.md (“131”), projects/index.md (“263”) — remove per D1.

---

Method note: 8 parallel read-only agents, each covering one vault section, all 329 md files read in full, cross-checked by grep against the rest of the vault; high-severity claims spot-verified by the coordinating session before filing. No wiki content was changed by this review other than creating this register and its index/log links (per WIKI-SCHEMA lint rule: report, don’t auto-fix).