| REQ-SYS-01 | The study shall state the total pre-revenue capital requirement in Australian dollars for each candidate service model, or record a decision excluding that model from financial modelling together with the reason. | Capital at risk is the first thing a go/no-go decision turns on. | Zaid intake 2026-08-20 | PASS if every candidate model carries either a capital figure whose cost lines all cite a SRC-### or ASM-### reference, or a DEC-### excluding it from modelling. FAIL on any unreferenced line or any silently omitted model. | Inspection | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-SYS-02 | The study shall state the elapsed time in months from company incorporation to first NDIS revenue for each candidate service model. | Time to revenue sets the size of the cash runway Zaid must fund from savings. | Zaid intake 2026-08-20 | PASS if each candidate model carries a month-count and a cited source for its registration lead time. FAIL if any model states a range with no source. | Inspection | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-SYS-03 | The study shall present a registration-strategy trade study containing at least three alternatives, each of which a competent practitioner in the sector could adopt without breaching a legal obligation, with criteria weights recorded before scoring. | GOV-B6.1 forbids a single-option decision record; registration timing is the highest-consequence choice in the project. | GOV-B6.1, GOV-B6.3 | PASS if the trade study shows >=3 alternatives, none of which breaches a legal obligation stated in the study, a weights block whose recorded timestamp precedes the scores block timestamp, a sensitivity statement naming the weight change that flips the result, and a rationale. FAIL on any missing element. | Inspection | Master Brain | Verified | quality | High | v2.0.0 |
| REQ-SYS-04 | The study shall present a service-model trade study covering core supports, support coordination and supported accommodation to the same evidential standard as REQ-SYS-03. | Zaid selected all three models at intake and asked for a recommendation. | Zaid intake 2026-08-20 | PASS if all three models are scored against identical weighted criteria with a sensitivity statement. FAIL if any model is omitted or scored on different criteria. | Inspection | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-SYS-05 | The financial model shall compute gross margin per billable hour from a sourced NDIS price limit and a sourced fully-loaded labour cost. | Margin per hour is the unit economic on which every volume calculation rests. | GOV-F8.8 | PASS if the margin cell resolves to a price input traced to SRC-002 and a labour input traced to SRC-009, SRC-011, SRC-012 and SRC-013. FAIL on any hard-coded margin. | Analysis | Master Brain | Verified | performance | High | v2.0.0 |
| REQ-SYS-06 | The study shall list every Victorian and Commonwealth licence, clearance, registration and insurance obligation with its issuing authority, current cost and lead time. | An unlisted obligation is an unbudgeted cost and a compliance exposure. | Zaid intake 2026-08-20; GOV-B2.8 | PASS if every row carries authority, cost, frequency, lead time and a SRC-### reference, with zero empty cells. FAIL on any empty cell. | Inspection | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-SYS-07 | The financial model shall compute the cash runway required under a downside case in which no participant is onboarded for six months. | The most common failure mode is running out of cash during the registration and referral lag. | GOV-E4.1 | PASS if a downside scenario exists with a stated month-by-month cash position and a single runway figure. FAIL if only a base case exists. | Analysis | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-SYS-08 | The study shall state a go/no-go recommendation together with at least three named, independently checkable items of evidence that would reverse or materially change it. | GOV-A1.12 requires that Zaid be made able to overrule the AI. | GOV-A1.12 | PASS if the recommendation section lists >=3 items, each naming a specific register ID and the party who could check it. FAIL otherwise. | Inspection | Master Brain | Verified | quality | High | v2.0.0 |
| REQ-SYS-09 | The study shall contain a conditional execution roadmap that activates only on a GO decision. | Zaid asked for go/no-go first and the roadmap second. | Zaid intake 2026-08-20 | PASS if the roadmap is present, is explicitly labelled conditional, and sequences every dependency named in REQ-SYS-06. FAIL otherwise. | Inspection | Master Brain | Verified | functional | Medium | v2.0.0 |
| REQ-SYS-10 | The financial model shall compute break-even billable hours per month for every candidate service model that is not excluded from financial modelling by a recorded decision naming the reason. | Break-even volume converts an abstract margin into a number of clients Zaid must win. | OBJ-2 | PASS if a break-even figure is computed from input cells for every model not excluded by a DEC-###, and every excluded model is named in the study. FAIL on any figure not derived from input cells, or on any silently omitted model. | Analysis | Master Brain | Verified | performance | High | v2.0.0 |
| REQ-SYS-11 | The study shall state the working capital required to fund one full wage cycle ahead of payment receipt. | Wages fall due before claims are paid; the prior working files omitted this entirely. | DEF-001 | PASS if a working-capital figure exists and is derived from the wage rate, the billable-hours driver and a stated payment-lag assumption. FAIL if the figure is asserted. | Analysis | Master Brain | Verified | performance | High | v2.0.0 |
| REQ-SYS-12 | The study shall state the effect of the mandatory registration expansion commencing July 2027 on each candidate service model. | The expansion has a confirmed date and it terminates the unregistered pathway for personal care. | SRC-015 | PASS if each candidate model carries a stated July-2027 consequence and a cost to be ready. FAIL if any model is silent. | Inspection | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-SYS-13 | The deliverable set shall include a diagram for every business process and dependency structure it describes. | GOV-F5.1 forbids a text-only description of a process. | GOV-F5.1 | PASS if every process narrative in the study has a corresponding image artefact. FAIL on any process described in prose alone. | Inspection | Master Brain | Verified | quality | Medium | v2.0.0 |
| REQ-SYS-14 | The Help Hub shall contain a troubleshooting entry for each of the nine failure modes named in GOV-G3.4. | GOV-G3.4 sets the minimum troubleshooting coverage. | GOV-G3.4 | PASS if the count of the nine named symptoms present in the Help Hub equals nine. FAIL on any lower count. | Test | Master Brain | Verified | quality | Medium | v2.0.0 |
| REQ-SYS-15 | The project shall carry exactly one dashboard generated as a projection of the registers. | GOV-F2.1 and GOV-F2.2 forbid a hand-maintained or duplicated status surface. | GOV-F2.2 | PASS if the dashboard is produced by a build script and every value on it maps to a register row. FAIL if any value is hand-entered. | Demonstration | Master Brain | Verified | interface | High | v2.0.0 |
| REQ-SYS-16 | Every metric shown on the dashboard shall carry a lineage entry stating its definition, formula in words, source data and as-of date. | GOV-G3.7 requires plain-language data lineage for every displayed number. | GOV-G3.7 | PASS if metrics-without-a-lineage-entry equals zero. FAIL on any count above zero. | Test | Master Brain | Verified | data | Medium | v2.0.0 |
| REQ-SYS-17 | The transfer pack shall re-derive from its sources under an automated check that fails when any file is missing, extra or altered. | GOV-F9.14 requires the pack be proven current by a script, not by assertion. | GOV-F9.14 | PASS if checker C22 runs, reports zero discrepancies, and its negative test demonstrates it can fail. FAIL otherwise. | Test | Master Brain | Verified | quality | High | v2.0.0 |
| REQ-SYS-18 | The study shall state, for every cost and rate input, its confidence level and what would raise that confidence. | GOV-F8.2 requires confidence wherever it is below high, with the reason. | GOV-F8.2 | PASS if every SRC row below High confidence carries a non-empty 'why below high' field. FAIL on any empty field. | Inspection | Master Brain | Verified | quality | High | v2.0.0 |
| REQ-CON-01 | The project shall cite a resolvable source for every external fact, figure, price, rate, date and legal position. | GOV-F8.8 names fabrication the highest-consequence failure in the Standard. | GOV-F8.8 | PASS if the source checker reports zero material external claims without a SRC-### reference. FAIL on any count above zero. | Test | Master Brain | Verified | constraint | High | v2.0.0 |
| REQ-CON-02 | The project shall scope its jurisdiction to Victoria, Australia. | Worker screening, workers compensation, portable long service leave and child-safety instruments are all state-specific. | Zaid intake 2026-08-20 | PASS if every jurisdiction-dependent obligation names Victoria. FAIL on any obligation stated for another state. | Inspection | Master Brain | Verified | constraint | High | v2.0.0 |
| REQ-CON-03 | The project shall carry a currency date and hold no source accessed after it or more than ninety days before it. | GOV-F1.16 makes a citation stale after twelve months. The original wording of this requirement demanded that every source share ONE accessed date, which was satisfiable only while the project had a single research wave. The aged care research is dated 7 September 2026 and the NDIS research 20 August 2026; under the old wording the project could only stay green by pretending the older sources had been re-read, which is exactly the falsification the requirement existed to prevent. Amended under CR-007 from equality to a staleness window, which is the test GOV-F1.16 actually describes. | GOV-F1.16 | PASS if no SRC row carries an accessed date later than the currency date, and none is more than STALENESS_DAYS days earlier than it. FAIL on any source outside that window, naming the source and its age. | Inspection | Master Brain | Verified | constraint | Medium | v2.0.0 |
| REQ-CON-04 | Every delivered Office artefact shall carry no text below 10 points, no document table above six columns, and a frozen header row and label column on every worksheet grid wider than six columns. | GOV-F4.5a requires artefacts that cannot carry a mode toggle to be readable on a phone without horizontal scrolling and without six-point text. The original wording of this requirement operationalised that as a flat six-column limit on every table, INCLUDING worksheet grids — and a twelve-month cash projection is fourteen columns across or eight transposed, so the criterion made the requirement unsatisfiable rather than demanding. The project first tried to interpret its way around that (DEC-008) and independent verification pass V9 was right to reject the interpretation: a spreadsheet is scrolled by nature, so what GOV-F4.5a actually asks of one is that scrolling does not lose the labels. Amended under CR-007 to test that instead, which is both satisfiable and closer to the rule. DEC-008 is superseded. | GOV-F4.5a | PASS if, across every delivered .docx, .pptx and .xlsx: text runs and cells below 10 point equal zero; document tables with more than six columns equal zero; and every worksheet using more than six columns has both a frozen header row and a frozen label column. FAIL on any count above zero or any wide sheet without frozen panes. | Inspection | Master Brain | Open | quality | Medium | v2.0.0 |
| REQ-CON-05 | The project shall record every assumption in the Assumptions Register with a confidence level and a stated failure consequence. | GOV-F1.9 requires that a spoken assumption also be a recorded one. | GOV-F1.9 | PASS if assumptions-without-a-confidence-level equals zero and assumptions-without-a-failure-consequence equals zero. FAIL otherwise. | Test | Master Brain | Verified | constraint | High | v2.0.0 |
| REQ-CON-06 | The project shall exclude any payment credential, government identifier and third-party personal datum from every file it produces. | INV-2 and INV-5 are inviolable and cannot be waived. | GOV-A3 INV-2, INV-5 | PASS if the secrets and personal-data scan returns zero hits across the controlled set. FAIL on any hit. | Test | Master Brain | Verified | constraint | High | v2.0.0 |
| REQ-MOE-01 | The delivered study shall leave zero open questions that block the go/no-go decision without an owned action. | GOV-C5.7 forbids declaring a cycle done while any item is open. | GOV-C5.7 | PASS if every open decision-blocking question carries an ACT-### with an owner and a due date. FAIL otherwise. | Inspection | Master Brain | Verified | MOE | High | v2.0.0 |
| REQ-MOP-01 | The proportion of paragraphs and table cells in the delivered study that contain a currency amount and also cite a SRC or ASM identifier shall be at least 90 percent. | A measurable anti-fabrication threshold makes GOV-F8.8 checkable rather than aspirational. | GOV-E7.1 | PASS if checker C27 reports an attribution ratio >= 0.90 over the delivered study. FAIL below 0.90. | Test | Master Brain | Verified | MOP | High | v2.0.0 |
| REQ-AC-01 | The study shall state the contribution per client per month, after direct labour, care management and administration, for the aged care business and for the NDIS business on the same basis. | The two sectors have different binding constraints — hours in the NDIS, budget in aged care — so a per-hour comparison misleads. Per client per month is the unit both businesses actually acquire and serve. | Zaid intake 2026-09-07 (RQ-004, RQ-005) | PASS if both figures appear in the delivered study, both trace to a computed model function, and a checker compares the workbook's cached values against the model to the cent. FAIL if either figure is stated without a computed source or the two artefacts disagree. | Test | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-AC-02 | The study shall state what an aged care participant's fixed classification budget buys in service hours across the whole classification range, not only at the modelled average. | The classification budget, not the hourly price, is the ceiling on revenue per client. A single average conceals that the bottom classification buys under two hours a week. | V4 independent verification finding | PASS if the delivered study carries a table spanning at least the bottom and top ongoing classifications with hours, revenue and contribution for each, every row computed rather than typed. FAIL on a single-point figure. | Inspection | Master Brain | Verified | functional | High | v2.0.0 |
| REQ-AC-03 | The study shall price the owner's choice of two separate entities against one combined entity, in dollars per month, per year and over three years. | GOV-B6.2 reserves the decision to the owner but does not excuse the project from putting the number in front of it. | Zaid intake 2026-09-07 (RQ-005) | PASS if all three figures appear in the delivered study and in the business plan, and all three derive from one function. FAIL if any is stated without a computed source. | Inspection | Master Brain | Verified | constraint | Medium | v2.0.0 |
| REQ-AC-04 | The study shall reconcile the aged care sector's published negative financial result against its own finding that the margin is workable, or record that it cannot. | A study that quotes a favourable margin while the sector reports losses, without addressing the gap, is not decision support. | SRC-068, SRC-069 | PASS if the delivered study states the mechanism, cites the independent verification pass that established it, and names the early warning indicator that would falsify it. FAIL if the contradiction is unaddressed or addressed by assertion alone. | Inspection | Independent verifier | Verified | functional | High | v2.0.0 |
| REQ-BUS-01 | The project shall carry a business_plan.json twin and a generated single-sheet A3 business plan page, both validated mechanically before any build is called complete. | Standard v4.0 REQ-BUS-001 makes the twin mandatory: no plan, no approval to build. | GOV v4.0 PIL-BZ-1 | PASS if the twin exists, all twelve of the generator's checks pass, the A3 page is no staler than the twin, and the rendered page fits one A3 landscape sheet as measured in a browser. FAIL on any of those. | Test | Master Brain | Verified | constraint | High | v2.0.0 |
| REQ-WEB-01 | The project shall carry the web-application function as a declared design slot: a deploy-anywhere self-contained container, a named-user login with passwords hashed at rest, and a pluggable AI-engine connector, with engine credentials held in server environment variables only. | GOV-B7.1 makes the slot mandatory from framing and applies it retroactively to every existing project. GOV-B7.2 leaves the build timing to the owner. | GOV-B7.1, GOV-B7.2, GOV-B7.3 | PASS if the slot is declared in the system breakdown with its status, carried in the backlog, and the credential scan over every generated artefact returns zero hits. FAIL if the slot is absent, or if any credential material appears in a client-served or generated file. | Inspection | Zaid | Verified | constraint | Medium | v2.0.0 |