The Curriculum Audit — needs list, rulings, build order

Owner order: “do a run through the curriculum and audit that for needs list we need more education / deeper education on.” Three adversarial Opus audits ran in parallel — one per school — against the three finished books and every school pointer in them; this document is the synthesis and the rulings. The three full reports sit beside it (AUDIT-automation-school.md, AUDIT-re-school.md, AUDIT-rei-school.md): 88 needs in all, every one with the unit to build, its learning objectives, its hands-on build, a length, and a video note. Nothing below asks a question the audit could answer itself; the few facts only the owner has are listed in one place at the end.

The state of the three schools in three sentences

The automation school is the only one with lessons — 24 files written 2026-08-17 against the old Book 1 draft — and Book 1 v5 has since moved under it: five v5 chapters have no module at all, one module (P7) teaches the inverse of its chapter, the book’s own appendix assigns school codes that collide with the registry seven times out of eleven, and the modules still print a retired cast and locked numbers the current book disowns. The real estate school (Tracks A/B/C) is design only — 12 modules, 38 units, zero drafted — and Book 3 now carries 59 school pointers into it, 24 of them from Appendix B into a Track A that, as designed, answers none of what Appendix B promises. The REI school is 20 charter rows and zero lessons, while Book 2 points at it five times and makes three present-tense membership promises, one of which (the term-sheet automation) the school’s own charter marks GATED.

Totals: 88 needs — 57 P0 (a book sends the reader somewhere nothing exists), 19 P1 (exists, too thin to deliver the promised outcome), 8 P2, 4 P3. Built out, the three schools go from 24 lessons to roughly 24 → 45 automation-school units, 0 → 81 real-estate-school units, 0 → 68 REI-school units — about 190 units, each a 900–1,500-word module in the existing nine-section format, each one video-ready by construction (see Part 3).

Part 1 — Rulings that cut across all three schools (apply before any lesson is written)

X-1 · One living state-law map, shared. Book 2 promises it three times (wholesaling, Ch9/13/18), Book 3 twelve times (from the copyright page on), and the RE design defers it to a section that doesn’t contain it. Ruling: build it ONCE as a cross-school asset — a per-state, per-rule record with a “verified as of” date, a reviewer, and a review cadence — and let Track A/B units, Track C’s compliance module, and the REI wholesaling map all read from it. It is the cheapest P0 on the list and the one that makes fifteen printed sentences true. (RE01, R-19, N09.)

X-2 · Rules never go in the audio. Anything the living map owns renders as dated on-page text; narration says “the current rule for your state is on this page.” Law units are narration-light on purpose. A stale sentence in a cloned voice sounds authoritative and cannot be diffed. (Standing rule for Part 3.)

X-3 · The Cost Defense lesson moves to the REI school. It was filed as a sixth Track C module in the agent school, against the same-day R42 ruling (“no sixth module”), and everything it teaches is investor-side. Ruling: it becomes the REI school’s RA-level cost-defense unit, split into three units (the stack, the disclosed margin, the seller-facing frame), re-formatted to the nine-section spec; its PropFlow P9 proposal does not travel with it. (R-23, RE31, N26.)

X-4 · The school keeps its module codes; the books adjust. Book 1 v5’s appendix assigned codes that collide with the registry (its P02 = verification, ours = scheduling; its A3 = the approve law, ours = ops elective; seven collisions). Eleven lines in one appendix vs 25 lesson files, the gallery and the assistant knowledge base. Ruling: the appendix’s code lines get corrected in the next book pull cycle, after the owner’s current review round — not during it. The same cycle fixes Book 3’s routing defect: Appendix B’s unlicensed reader is sent eight times to B1/B3, which are the Broker upgrade track; those pointers move to the new Track A modules below. (N06, RE33.)

X-5 · Roadmap stays roadmap in the school too. Three addendum units (C1-U3, C2-U2, C2-U3) assign as student builds the three mechanisms the book labels “roadmap, not running” (the call list, the anniversary check, the time-value view). Ruling: those units teach the shape and the approval discipline, with no build step and a plain “not running yet” line — R-H applies to lessons exactly as to chapters. (RE06.)

X-6 · The retired-cast sweep. Maya, Dan, Lena, Rosa, Theo and their locked numbers appear in nine modules and all six demos; none of them exists in Book 1 v5. Ruling: one mechanical sweep replaces them with the v5 book’s cast and numbers before any module is recorded — a recorded lesson with a dead character is the expensive kind of wrong. (N13.)

X-7 · A build gets a screen recording, not a slide. The secrecy line forbids invented UI. Every unit whose center is a platform build names its recording; the slide pipeline carries everything else. (Part 3.)

Part 2 — The needs list, ranked (the top of each school; full lists in the three reports)

2a · Automation school (27 needs: 11 P0 · 9 P1 · 5 P2 · 2 P3)

(N08–N11 in that report are REI-school needs and are carried in 2c.)

2b · Real estate school, Tracks A/B/C (33 needs: 26 P0 · 5 P1 · 1 P2 · 1 P3)

Track C holds at five modules (R42) — 14 → 27 units. Program: 12 modules/38 units → 17 modules/81 units; bank ≈945 → ~1,200 items.

2c · REI school (28 needs: 20 P0 · 5 P1 · 2 P2 · 1 P3)

Doctrine note carried from that report: Ch15’s eviction-filing gate (“no path up”) vs Ch17 (“I expect that to change”) — the lessons teach it as an owner-set policy gate with a review date; the licensed/consent ceiling remains the only permanent one.

2d · If we could only build nine (three per school, in order)

1. X-1 the living state-law map — one asset, fifteen printed promises, three schools.
2. N06 + N13 — codes and cast reconciled before anything is recorded (one week; everything later inherits it).
3. N02 P12 The Approve Queue — the law the whole automation school stands on.
4. RE16 C5-U3 Reading the Ladder — one twenty-minute unit that makes 26 others teachable.
5. RE07 F5 Entering the Profession — where the RE school becomes worth money to Appendix B’s reader.
6. RE26 C4-U2 The Team Intake Standard — Ch14 printed “walked step by step.”
7. R-21 RP4 The Offer Engine — spec already at verification depth; answers Book 2’s villain.
8. R-02 RP3 Feeding the Hunter — a sweep on garbage data is worse than none, and it carries the terms-of-use rule.
9. R-19 — the wholesaling half of the living map, the honest membership driver the ledger designed the hand-off around.

Part 3 — Production: cloned voice, timelined slides, avatar clips — 100% automated

The owner’s constraint is the design: no filmed lectures; a cloned voice; slideshows timelined to the topics of a course so the result looks intentionally designed; his avatar only for a few clips in between. A working prototype now exists at school/_video/ (see its README): one command turns one module file into a finished lesson package — the module’s nine sections become a segment list (cold open · concept · why · prerequisites · one segment per build step · “you’ll know it worked when” · one per failure mode · the words · show your build · one per check question · outro); each segment gets a designed slide (seven layouts, the school palette and type), a narration text in spoken register, a duration from spoken pace, and a place on the timeline; the pipeline renders a keyboard-navigable deck, 1080p frames, a captioned MP4 with crossfades, a narration file for the voice clone, and a manifest with a checksum of the source markdown so an edited lesson re-renders itself. The two sample lessons (U2 and P3) came out at 8 and 9 minutes, 22–23 segments each, three avatar slots each. A second script turns the curriculum map into a course page with chapter navigation deep-linked to the segment.

What only the owner supplies: the voice clone (an API key and a voice id, never stored in a file), avatar clips for the three slots per lesson (drop a file in avatar/<segment>.mp4 and the next run uses it), and a final watch. What the pipeline never does: put a rule in the audio (X-2), invent a screen (X-7 — builds are screen recordings, captured once, re-used by the timeline), or add a fact the module doesn’t contain.

Per-lesson production cost once the pipeline is wired to the voice: minutes of compute, one owner watch. The 190 units above therefore cost what the writing costs, which is the same fleet that wrote the books; the video is free after the words are right — which is why the reconciliation items (X-4, X-6) come first: a recorded lesson is the expensive place to be wrong.

Gaps in the prototype, honestly: the deck fonts fall back to a local serif inside the render container (the font files install in one step on the device); the voice backend is implemented against one vendor’s API and tested only in its silent fallback; screen recordings are not generated — they are captured by a person once per build and dropped in by filename like avatar clips.

Facts only you have (the one-line list — nothing else on this list waits on you)

1. Is the member term-sheet automation live? Book 2 Ch10/12/17 say it is “there for any reader of this book, free, inside the membership section”; the REI charter marks RP9 GATED “until the funding funnel is live + compliance cleared.” If it is live, RP9 gets a module and the compliance workstream closes first; if not, three sentences soften to direction in the next pull cycle. The default if you say nothing: the sentences stay as printed through your review round, and no RP9 lesson is recorded until you confirm.
2. The voice clone and avatar — the key/voice id, and clips for the three slots per lesson when you want them; the pipeline runs silent until then.

Automation school (U · P · A · demos) — 27 needs

lens: the ProcessAutomater School (Universal U1–U5 · Platform P01–P10 · Applied A1–A3 · Demos D1–D6 · SCHOOL-MODULE-SPEC · DEMO-LIBRARY-SPEC · CURRICULUM-MAP) · audited 2026-09-02 against Book 1 v5 (B1v5-00…16), Book 2 draft, Book 3 draft, the ledger, and the governing briefs.

STATE OF THIS SCHOOL

Twenty-five lesson files, 41,419 words, all written on 2026-08-17 against a Book 1 draft that no longer exists. That single fact is the audit. The old book ran 00-intro → 18-appendix and the module registry in SCHOOL-MODULE-SPEC maps every module to one of those files by name; Book 1 is now v5 — fifteen chapters plus an appendix, a different spine (SEE IT → SHOW IT → TRUST IT → MULTIPLY IT → BECOME IT), a different cast, and an entire third movement about trust that did not exist when these modules were written. The damage is not cosmetic. Five of the fifteen v5 chapters have no module at all — Ch5 (what to teach first / the judgment audit), Ch7 (verified is done), Ch8 (the machine prepares, you approve), Ch9 (when it breaks), Ch14 (what survives) — and a sixth, Ch12, has a module (P7) that teaches close to the opposite thing: P7 builds a step where the machine exercises judgment; the chapter is about finding the one decision that stays yours and building the whole chain to protect it. Meanwhile the book’s own appendix, “The Builds,” has already assigned school module codes to every chapter’s build — and seven of those codes collide with modules that already exist under different titles (the appendix’s “Module P02” is verification; the school’s P02 is scheduling). The appendix flags this itself: “A few are provisional… nothing below is final until that mapping is confirmed.” Nobody has confirmed it.

Underneath the mapping problem sits a truth problem. The v5 storycraft directive rules the old cast dead — “Banned: the retired cast” — and Maya, Dan, Lena, Rosa and Theo appear in exactly zero v5 chapters. The entire demo library is built on them, and every “locked number” the DEMO-LIBRARY-SPEC pins (6 hrs/wk, 340 rows, 90 min/day, 45 min/night, ~3 saved mistakes a month) traces to a continuity ledger the current book does not carry. Those numbers are still printing in nine files’ “Why it matters” paragraphs as fact.

Where modules do exist, depth is genuinely uneven. P1, P3, P4, A3 and D2 are real: a build a learner can perform, a “You’ll know it worked when…” that is falsifiable, and failure modes with usage-level fixes. U1, U2, U5 and P2 are thin by comparison — U1’s build is three sentences on paper and a contract with yourself; P2’s build is “find the question, pick one of three answers, walk away.” They are not wrong, they are just short of teaching. And across all 25 files there is not one word about the things a business owner hits in week one that no chapter covers either: signing up, connecting an account, cleaning the records the jobs will read, triaging errors when four jobs fail overnight, what any of this costs to run, who on the team can see what, or how to actually count the hours you got back — despite the narrator being canonically an “obsessive tracker of hours saved” and every module’s “Show your build” asking for a before/after number the school never teaches anyone to produce.

Finally, Book 2 sends readers to a school five times, and every one of those five lands in the REI school — chartered on 2026-08-18, curriculum-mapped, never built. Not one of those pointers is answerable by anything in this corpus today.

RANKED NEEDS

THE NEEDS IN FULL

N01 · P11 — Verified Is Done

1. Where it sits. PLATFORM track, new module P11, taught immediately after P02 (scheduling). Prerequisite: P01 + P02. The book’s appendix calls this “Module P02”; see N06 for why the school should keep its own code and have the appendix updated.

2. What promises it. B1v5-07, the whole chapter: “scheduled is not done; verified is done.” The appendix build entry is explicit — “The walkthrough — a single verification follow-up… The walkthrough — a full trigger chain… School module. Configuring the verification trigger for a single scheduled task, and the full escalation ladder for a multi-step legal or compliance chain, is Module P02.” Ch13 makes it structural: “Scheduled is not done. Verified is done. A Routine doesn’t have to remember that rule. It’s built into when it runs.” Ch12 extends it: “verified is payable.”

3. What exists today. Nothing. The words “verify,” “verified” and “verification” do not appear in any of the 25 lesson files. P02 (1,015 words) teaches scheduling and stops there; its “when it goes sideways” even normalises the gap — “Results are piling up and you’ve stopped checking them. That’s not a breakage, it’s a housekeeping habit.” That is precisely the belief Ch7 exists to break.

4. The gap in plain words. The school teaches a learner to schedule a job and walk away, and never teaches them that a job marked done and a job actually finished are two different objects. Every downstream module inherits the hole: P9 chains jobs with no completion proof between links, P10’s morning stack reports what ran without confirming what landed, A3 posts an invoice into a review queue with no check that the work it bills for happened.

5. What “deeper” means. Three units.

• P11-U1 — A calendar entry is a promise, not a receipt (12 min). Objectives: (a) state the difference between a scheduled step and a verified one, in your own business’s words; (b) find the three places in your own week where a status flips to “done” the moment a task is assigned; (c) explain why silence is the hardest failure for any system to notice. Build: pen and paper — walk your own job list and mark every step whose “done” is really “sent.” You’ll know it worked when you can point at one status field in your own tools and say out loud what it actually proves, and it isn’t what you assumed.
• P11-U2 — The one question, asked automatically (25 min). Objectives: (a) set a follow-up that fires a set stretch after a scheduled step and asks the person who would know whether it happened; (b) set the mirror prompt to whoever was supposed to do it; (c) build the mismatch case — what happens when the two answers disagree or neither comes back inside your window. Build: on the learner’s own account, against a real scheduled thing (a delivery, a visit, a call-back). You’ll know it worked when a scheduled thing in your business comes back with a plain yes-it-happened or a plain nobody-answered — and you didn’t ask for either one.
• P11-U3 — The chain, link by link (30 min). Objectives: (a) map a multi-step sequence where each step is supposed to fire when the one before it completes; (b) require proof of the fact at each link, never the intent; (c) set an escalation window per link that surfaces to a person before an outside calendar moves on without you; (d) run the arithmetic of one slipped link in your own numbers. Build: the learner’s own escalation chain — an onboarding sequence, a claims chain, a renewal chain. You’ll know it worked when you can name, for every link in one real chain, what evidence would prove it happened, and one link now refuses to advance without that evidence.

6. VIDEO NOTE. U1 and U3’s arithmetic are pure slide+narration — the per-diem model in Ch7 (a door, a day-rate, a docket cycle) is a five-slide build that lands hard and needs no screen. U2 is the one screen recording in this module: setting a trigger relative to a completed step, and the mismatch branch. Avatar moment: one, and it is the contractor no-show. “My team looked at that ticket, saw the word scheduled attached to a contractor’s name and a date that had already passed, and believed it was handled.” That is a confession, it is the reason the module exists, and it needs a face.

N02 · P12 — The Approve Queue: The Machine Prepares, You Approve

1. Where it sits. PLATFORM track, new module P12, taught after P11. A3 (Ops & Finance) stays as the applied sibling that points at it rather than duplicating it.

2. What promises it. B1v5-08 is a whole chapter and states the law twice: “the machine prepares. You approve.” The taxonomy is the payload — “Reversible and cheap, let it run and forget about it. Reversible and expensive, let it run inside a ceiling you set, and read the report after. Irreversible, it waits for you.” The appendix assigns it: “Setting a specific ceiling for a specific vendor or payment category, and the full de-entry extraction and audit-queue configuration, is Module A3.” Ch13 applies it (“the Routine can move it forward on its own right up to a ceiling I set myself”) and Ch12 names it as the pattern the fence follows.

3. What exists today. A3 — Ops & Finance, Drained (1,676 words), the strongest applied module in the corpus. It teaches “it prepares, you approve,” puts a review gate wherever money or a customer is touched, and carries the money-honesty rule well. But it is an elective requiring P1–P4, it teaches one bucket (gate everything that touches money), and it contains none of: the three-bucket reversible/expensive/irreversible taxonomy, the ceiling as a settable number, the de-entry queue build, the extraction-is-not-judgment distinction, or the counterweight.

4. The gap in plain words. The single most load-bearing design rule in the book is currently reachable only by a learner who has finished four platform modules and then chosen the finance elective. And the version they get teaches “gate it” where the book teaches “sort it into one of three buckets, and the bucket — not your mood that day — decides.” The missing counterweight is the dangerous part: “an approval queue you never actually read is worse than no queue at all… it manufactures the feeling of control without the fact of it.” A school that teaches people to build queues and never teaches them to keep queues narrow is manufacturing rubber stamps at scale.

There is also a doctrine conflict to settle inside this module, not around it. The Trust Doctrine binds the school: “Never write ‘permanently / every time / no exceptions’ about review gates; write ‘until promoted / until it earns it.’” Ch8 writes: “Irreversible, it waits for a person. Not usually. Always.” Both are true if stated precisely — the category never graduates, the ceiling on the category moves. P12 must say that in one sentence and hold the line; A3 already gets close (“that’s not the same thing as an eternal wall”).

5. What “deeper” means. Four units.

• P12-U1 — Three buckets, not one rule (15 min). Objectives: (a) sort any task by asking whether a wrong outcome is reversible, and whether reversing it is cheap; (b) name why it’s the presence of a decision, not the presence of money, that decides; (c) place each of your own running jobs in one of the three buckets. Build: sort your own job list. You’ll know it worked when at least one job you’d been approving by hand every time moves to bucket one and stops asking you.
• P12-U2 — Setting a ceiling on purpose (20 min). Objectives: (a) pick a dollar or scope line where a mistake underneath it would be genuinely cheap; (b) set it so everything under it runs and reports afterward, and everything above it waits; (c) say why you chose that number. Build: one ceiling, on one real category. You’ll know it worked when something under your ceiling ran without asking you and told you afterward, and something above it stopped.
• P12-U3 — The de-entry queue (30 min). Objectives: (a) accept a document however it actually arrives; (b) read fixed fields the same literal way every time and know why that’s extraction, not discretion; (c) compose the finished entry; (d) route it to a short audit queue before it posts, and name the graduation path. Build: invoices or receipts into whatever the learner keeps their books in. You’ll know it worked when a finished, correctly-composed entry is sitting in a queue waiting for a glance, and you didn’t type any of it.
• P12-U4 — Keeping the queue narrow enough to read (12 min). Objectives: (a) count what actually reaches you in a week; (b) identify the reversible items that shouldn’t be there and retire them; (c) recognise rubber-stamping in yourself before it costs you. Build: a one-week queue audit. You’ll know it worked when your queue is short enough that you read every line in it, and you can name what you removed to get there.

6. VIDEO NOTE. U1 is the strongest pure slide sequence in the whole proposed curriculum — three buckets, three consequences, one rule, animate the sort. U2 and U3 are screen recordings (setting a ceiling; watching an entry compose and land in a queue). U4 is slide+narration. Avatar: the counterweight, and only that. “A gate you rubber-stamp without reading isn’t a gate — it’s a delay with a signature attached.” Said to camera, it is a warning; on a slide, it’s a bullet.

N03 · P13 — When It Breaks: Notice, Stop, Report, Wait

1. Where it sits. PLATFORM track, new module P13, after P12. Buildable on paper if the learner has nothing running yet, the way U4 handles the pre-platform case.

2. What promises it. B1v5-09 in full. “That’s what a well-designed system does when something breaks: it notices, it stops, it reports, and it waits.” And the sentence the whole module turns on: “a system that stops to ask a question is a success, not a failure… A system that never stops isn’t a system you can trust; it’s a system you haven’t found the edge of yet.” The appendix specifies the build: “Setting expected-structure baselines for scraped or portal-driven steps, heartbeat windows for standing sources, and recheck intervals for time-bound assumptions is Module U5.”

3. What exists today. Nothing, at any depth. The closest the corpus comes is a reassurance in the opposite direction: P6 and U4 teach “it figured it out” as “the ordinary outcome, not the exception.” That phrase is the vocabulary ceiling and it is correct, but taught alone it trains a learner to expect adaptation and never to design for the case where adaptation is the wrong answer.

4. The gap in plain words. The school currently teaches learners to build things and never teaches them what a well-built thing does at its own edge. Three specific failure shapes are untaught: a source that changes shape (fields move, a login flow grows a step), a source that goes quiet (the absence is the event, and “an absence is the hardest thing for anyone — person or system — to notice”), and a rule that quietly stops being true (a fee, a grace period, a required document). The third is the one nobody builds for: “The fix isn’t remembering to check it — you won’t, and neither will I. The fix is building the recheck into the automation itself.”

There is a fourth thing here the school has to carry and currently does not: Ch9’s honesty about fluency. “AI is fluent. Fluency is not the same thing as truthful.” A1 has the verify-sources rule for drafting, which is good but is scoped to content. This belongs in the platform spine.

5. What “deeper” means. Three units.

• P13-U1 — A system that stops is the one that’s working (12 min). Objectives: (a) unlearn the instinct that a pause is a failure; (b) name the four moves — notice, stop, report, wait; (c) describe what a system that never stops is actually doing instead. Build: look at one running job and write down what it would do if the thing it reads changed overnight. If you can’t picture the answer, write that down as a gap. You’ll know it worked when you can say, for one of your own jobs, exactly how you’d find out it hit something it didn’t recognise.
• P13-U2 — Baselines, heartbeats, and rechecks (30 min). Objectives: (a) set an expected-structure baseline on a step that reads something you don’t control; (b) set a heartbeat window on a standing source so the silence itself becomes the alert; (c) put one time-bound assumption on its own recheck schedule. Build: all three, on real jobs. You’ll know it worked when you’ve made a source go quiet on purpose and the silence reached you before you went looking.
• P13-U3 — Human on exception (15 min). Objectives: (a) distinguish being in the loop from being the exception handler; (b) catch the sentence “it’s safer if I just check it myself” as the old bottleneck in new clothes; (c) decide what actually reaches you and what doesn’t. Build: write your own exception list — the things that should find you, and nothing else. You’ll know it worked when your list is short and you can defend every line on it.

6. VIDEO NOTE. U1 is slide+narration and carries the strongest metaphor the book owns — the loom that stops itself when a thread breaks, and the cord any worker could pull. Animate it. U2 is the module’s one screen recording, and it needs to show a failure on screen: a changed page, a stop, and the specific difference reported. U3 is slide. Avatar: the ten-thousand-dollar crash. “I had rented something and never asked what was holding it up.” One clip, thirty seconds, no slide behind it.

N04 · P07 — Teams, the Relay, and the One Fenced Decision (rewrite)

1. Where it sits. PLATFORM track, existing P07, rewritten and expanded from 1 unit to 4. The current judgment-step build survives as one of them.

2. What promises it. B1v5-12, whole chapter. “You protect a relay by finding the one decision in the whole chain that genuinely cannot be reduced to a rule… and you build everything else around delivering that one decision to you cleanly, rarely, and only when it’s real.” The appendix: “Standing vendor-discovery monitoring, materials-mismatch propagation, and the full confirmation-cadence and fallback configuration for a vendor-dependent chain are Module P07.” Also the prerequisite nobody warns about: “you cannot automate a paper process… nothing automates until the paper is gone.”

3. What exists today. P07 — Teams of Workers: Judgment Steps (1,620 words), a well-written module about adding one step where the machine reads a pile and returns a shortlist with reasons. It carries a genuinely good device (the judgment-vs-rules test) that survives intact.

4. The gap in plain words. The chapter and the module point in opposite directions. The book’s “fence” is where a human decision waits, protected by everything around it; the module’s “judgment step” is where a machine decision happens, feeding a review queue. Both are real, but the school has built the second and called it the chapter’s name, so a learner who reads Ch12 and opens P7 gets a different lesson than the one they came for. Missing entirely: the relay itself (one verified confirmation triggering the next step across people who never speak to each other), the vendor bench as a compounding asset, materials-mismatch propagation, the ghost protocol (a hard cutoff, an automatic fallback, exactly one message reaching a person), the arithmetic that makes a fenced decision obvious ($380 against $2,200), and the paper prerequisite.

5. What “deeper” means. Four units.

• P07-U1 — You cannot automate a paper process (10 min). Objectives: (a) find the paper still sitting inside a process you want to hand over; (b) distinguish digitising from automating; (c) sequence the two. Build: one process, paper inventory, digitisation plan. You’ll know it worked when the last clipboard in one process has a digital equivalent something else can read.
• P07-U2 — The relay: verified handoffs (30 min). Objectives: (a) chain two steps so the second fires on a verified confirmation, never on a guess at when the first “should” be done; (b) propagate a change downstream automatically the moment it’s confirmed; (c) explain why each link knows what it’s waiting for. Build: a two-link, then three-link relay from the learner’s own work. You’ll know it worked when a confirmation you didn’t send released a step you didn’t start.
• P07-U3 — Finding the fence (20 min). Objectives: (a) run the judgment-vs-rules test across the whole chain (this is the current P7 content, kept); (b) identify the one call that genuinely refuses to compress; (c) build everything upstream to keep bad information away from it and everything downstream to move the moment it’s answered; (d) price the decision so it’s obvious why it’s yours. Build: one fence, deliberately placed, with its arithmetic written down. You’ll know it worked when exactly one question reaches you out of a full chain run, and you can say in one sentence why nothing else did.
• P07-U4 — When a link goes quiet (25 min). Objectives: (a) set a check-in cadence and a hard cutoff; (b) re-notify downstream with a conditional timeline instead of waiting; (c) hold a fallback automatically and route exactly one decision to a person. Build: the ghost protocol on a real vendor-dependent step. You’ll know it worked when a missed check-in produced a protected schedule and one clean question, instead of a discovery on somebody’s dock.

6. VIDEO NOTE. U2 and U4 are diagram-driven slide+narration — a relay is a picture, and animating one confirmation releasing the next crew is worth more than any screen. U3 needs a short screen recording of a review-queue item arriving with its reason attached. Avatar: the week with four jobs running. “Notice what I was doing across all four of those jobs, in that one week: nothing.” And the one thing that did need him — the repair that passed every rule and still didn’t sit right. That is the fence, said by the person who found it.

N05 · U6 — What to Teach First: The Judgment Audit

1. Where it sits. UNIVERSAL track, new module U6, sequenced U1 → U2 → U6 → U3 → P01. Pen-and-paper, no account required — which makes it free-tier acquisition content, the same job U1/U2 already do.

2. What promises it. B1v5-05 in full. “If you can write down what the judgment actually consists of — the specific things checked, in a specific order — it isn’t judgment. It’s a rule you never got around to writing.” And the ordering rule: “Start where a wrong guess costs you an afternoon, not a client.” The appendix names all four builds: the audit, the cheap-mistake ranking, the print-and-mail sub-build, and the trade-thread watch, and assigns them a module.

3. What exists today. U2 (1,347 words) gets a learner to a top-five list ranked by frequency × duration, and its “when it goes sideways” gestures at the problem — “you’re listing your judgment calls, not your repetition — the tasks only you can do, dressed up as busywork.” But it gives no method for testing that, and it explicitly ranks by cost of the task, not by cost of a mistake — the opposite of what Ch5 teaches.

4. The gap in plain words. The school’s answer to “where do I start” is “whatever eats the most hours.” The book’s answer is “wherever a wrong guess is cheapest, and don’t believe yourself when you call something judgment.” Those produce different first jobs, and the book’s produces a learner who succeeds. The whole photo-payout story — an eye that turned out to be “four or five plain questions run in the same order every time” — is the best teaching artefact in Book 1 and appears in no module.

5. What “deeper” means. Three units.

• U6-U1 — Write down what the judgment actually is (20 min). Objectives: (a) take a task you protect with the word “judgment” and try to write out what you check, in order; (b) recognise a rule wearing an instinct’s clothes; (c) find the one that genuinely refuses to compress and say why. Build: the implicit-rule worksheet, run against three tasks. You’ll know it worked when at least one thing you’d have sworn was judgment is sitting on your page as a numbered list.
• U6-U2 — Start where a mistake is cheap (12 min). Objectives: (a) rank your candidates by what a wrong guess costs, not by what the task costs; (b) explain why trust compounds the same way with software as with a new hire; (c) pick your first job on that basis. Build: re-rank the U2 top five by cost-of-error. You’ll know it worked when your first job has changed, and you can say what it would cost you to be wrong about it.
• U6-U3 — Two rooms where nobody thought it reached (15 min). Objectives: (a) design a pipeline that expects a delay instead of treating it as failure (the records-request case); (b) point a standing watch at a public conversation instead of searching it when you’re already stuck (the vendor case); (c) name the “reservoir” idea — anything that fills quietly while you do something else. Build: identify one of each in your own business. You’ll know it worked when you’ve named one wait you can start and forget, and one search you can stop repeating.

6. VIDEO NOTE. Entirely slide+narration — this module has no platform surface at all, which makes it the cheapest deep module in the list to produce. Avatar: the upside-down faucet handle. Two seconds of looking, a stopped payout, a job that wasn’t finished. It is the book’s best small story and it earns a face; the four-questions reveal that follows can be slides.

N06 · Module-code reconciliation (CURRICULUM-MAP + SCHOOL-MODULE-SPEC)

1. Where it sits. The registry itself — SCHOOL-MODULE-SPEC’s module tables and CURRICULUM-MAP’s three track tables, plus a corresponding correction pass on B1v5-16’s eleven “School module” lines.

2. What promises it. B1v5-16: “Some codes are settled. A few are provisional, marked as such, because the school’s own module map is still being finalized against this book’s final chapter order; nothing below is final until that mapping is confirmed.” Two entries carry <!--PROVISIONAL CODE — school-mapping owner confirms-->. The confirmation has not happened, and the collision is worse than the appendix realises.

3. What exists today. Eleven appendix entries. Two agree with the school (P01, P06). Two are marked provisional (P04, U5). Seven collide silently — the appendix’s P02 is verification while the school’s P02 is scheduling; its P03 is risk-phased widening while the school’s P03 is dataset column-mapping; its P04 is the judgment audit while the school’s P04 is the results drawer; its P05 is say-it while the school’s P05 is the phone assistant; its A3 is the approve law while the school’s A3 is the ops elective; its U4 is a four-rung ladder while the school’s U4 is a two-rung one; its U5 is failure discipline while the school’s U5 is the relief test; its P07 is the fenced decision while the school’s P07 is a machine judgment step.

4. The gap in plain words. A reader finishes Chapter 7, turns to the appendix, is told the full version is “Module P02,” goes to the school, and gets a lesson on scheduling. That happens seven times. The school corpus is also the source for the website pages and the site assistant’s knowledge base — a wrong code is a wrong answer served three ways.

5. What “deeper” means. Not a lesson — a decision and a sweep, and it must land before any new module ships or the new modules inherit the collision.

• The recommendation, stated plainly: the school keeps its existing codes; the book’s appendix is edited to match. Reason: eleven appendix lines are cheaper to change than 25 published lesson files, a gallery whose posts reference module IDs, and an assistant KB keyed to them — and the book is not printed yet.
• The deliverable: one reconciliation table, published in CURRICULUM-MAP, with a column for every v5 chapter, the module that mirrors it, and a plain “no module — argument only” for Ch1, Ch2 and Ch14. Ch14’s row should read “U7” once N16 is built, not “no module”; the appendix is wrong to say Ch14 leaves nothing behind, because a Universal-track concept lesson is exactly what it leaves behind.
• Also fix in the same pass: SCHOOL-MODULE-SPEC’s registry still cites 03-first-job.md, 05-whole-list.md, 10-assistant-in-your-pocket.md, 17-conclusion.md and eight other filenames from a book that no longer exists. Every row needs a v5 chapter number.

6. VIDEO NOTE. None — this is registry work, not a lesson. It does, however, gate every other video in this document: a slideshow narrated with a module code that turns out to be wrong is more expensive to fix than a markdown file.

N07 · D7, D8, D9 — the three builds the book names by hand

1. Where it sits. DEMO LIBRARY, three new demos under DEMO-LIBRARY-SPEC’s existing 9-section format.

2. What promises it. B1v5-15, the invitation, naming the library’s contents literally: “a growing collection of builds other people have already taught, once, and put back into the community… a vendor approve-gate, a follow-up sequence, a verification check, a whole chained morning like the one I described to you earlier in this chapter, already built, ready to be shown a second time, in your business, in an afternoon instead of a month. You can name what’s in it because it’s real.” Corroborated by REI-EXPANSION §8’s main-school launch set: #13 vendor self-onboarding, #16 morning briefing, #17 review-queue follow-up chain.

3. What exists today. Six demos. D5 (Invoice Chaser) is the follow-up sequence and is genuinely good. The other three named items do not exist. D1–D4 and D6 are excellent builds but none of them is a vendor approve-gate, a verification check, or a chained morning.

4. The gap in plain words. The book tells a reader they can name what’s in the room, then names four things, and three of them aren’t there. That is the single most checkable promise in Book 1 and it currently fails 3-for-4.

5. What “deeper” means. Three demos, 700–1,200 words each, built to the existing spec.

• D7 — The Vendor Approve-Gate. The Ch8 cleaning-company loop, generalised: find the companies worth pursuing, fill out each one’s vendor application in its own format, draft the request for whatever requirement isn’t met yet, gather what it costs — then stop, with one line: this is what it costs. Approve? Practices P01, P03, P12. “You’ll know it worked when a dozen applications are finished, a dozen requirements are priced, and exactly one question is waiting for you.”
• D8 — The Verification Check. The Ch7 build, persona-neutral: something got scheduled; the same day or the next, ask the person who’d know whether it happened, ask the person who was supposed to do it, and escalate the mismatch. Practices P02, P11. “You’ll know it worked when a scheduled thing in your business came back marked done by two people who don’t work for each other — or came back flagged because it didn’t.”
• D9 — The Chained Morning. The Ch13/Ch15 capstone as an installable build rather than a module project: the check, the message, the approval gate, the exception, chained with one order, one branch, one schedule. Practices P09, P10, P11, P12. “You’ll know it worked when tomorrow’s coffee comes with a note you never asked for, and exactly one thing in it is waiting on you.”

Note for the spec: DEMO-LIBRARY-SPEC’s “Numbers: ONLY the locked cast numbers” rule cannot be satisfied by these three, because the v5 book retired the cast and kept almost no numbers (Ch15 is explicit: “I can’t tell you the dollar value of that morning, and I’m not going to invent one”). These demos should carry no performance claim at all — describe the shape, not the saving. That is a spec amendment, and it is the honest one.

6. VIDEO NOTE. Demos are the best avatar-free content in the school: build sheets read well as slide+narration with a screen recording under the numbered steps. D7’s one screen moment is the approval line arriving. D8’s is the mismatch branch. D9 is the one place in the whole corpus that earns a time-lapse rather than a recording — an empty morning filling itself, ending on one waiting question.

N08 · REI school RP1–RP10 — “full builds, module by module”

1. Where it sits. REI PROCESS AUTOMATER SCHOOL, RE-PLATFORM track RP1–RP10, chartered in REI-SCHOOL-SIBLING-CHARTER §2. Outside the automation school’s walls; inside its format law (the charter inherits SCHOOL-MODULE-SPEC’s 9 sections wholesale, plus §5a “Install the worker” and §10 “The line”).

2. What promises it. B2-CH20 ¶49: “The school exists past that door, and it’s real, and it goes as deep as anyone wants to take it — full builds, module by module, for every structure and every stage this book has walked you through, taught by people who’d rather show you the whole thing working than gate it behind a homework assignment you have to finish before you’re allowed to see what you paid for.” That sentence is the closing argument of Book 2 and it is positioned as the antidote to the Chapter 2 program.

3. What exists today. A charter with a ten-row module table, a five-row applied table, community rules, a beta bundle recommendation, and an offer-engine spec good enough to build RP4 from. Zero lesson files. The automation school covers none of it: RP1–RP10 are all real-estate pipeline stages, and the nearest automation-school analogues (P3 for lists, P4 for results, P6 for portals) teach the mechanic without the domain.

4. The gap in plain words. Book 2’s entire ethical position — we’re not the program that gated you — is cashed by a school that has nothing in it. If Book 2 ships before RP1–RP10 exist, the book repeats the exact failure it spends twenty chapters itemising, and a reader who buys on that sentence finds a charter.

5. What “deeper” means. This is not a “deepen an existing unit” need; it is a build order. Recommended first three, matching the charter’s own beta bundle and the ledger’s priorities:

• RP1 — Lists: Where Deals Hide (3 units: getting on lists · parsing inbound · staleness and immediacy timers). The daily-habit moat; touches every member every day.
• RP4 — The Offer Engine (4 units: the inputs that are actually true · one deterministic answer per structure · the receipt as the interface · the stress toggle). Fully specced already in OFFER-ENGINE-SPEC §§0–2 — this is the least-blocked deep build in the entire corpus, and it is the differentiator no coaching program has.
• RP7 — Rehab at a Distance (3 units: the photo/doc gate · verified-is-payable · the confirmation loop). Shares its spine with the automation school’s N01 and N04, so it can be written second-cheapest.

Each unit needs 2–4 objectives in the school’s own voice, an install-first opening per charter §5a, a hands-on build with a “You’ll know it worked when…”, and the compliance line per §10.

6. VIDEO NOTE. RP4 is the one place in either school where a screen recording is the entire product: watching a defensible offer compute from facts, with every number expanding into its own receipt, is the demo that makes Book 2’s argument without a word of narration. RP1 and RP7 are slide+narration with short screen inserts. Avatar: one clip, on RP4’s opening — the blank cell, the hidden sheet named “.”, and the sentence that a tool nobody is allowed to open rots exactly where nobody’s looking.

N09 · REI school — the living state-law map

1. Where it sits. REI school, RU5 (The Covenant) extended plus a new standing reference module — proposed RU6 — The Line That Moves, with a maintained state table rather than prose.

2. What promises it. B2-CH09 ¶11: “a wave of state laws passed in 2025 and 2026 now regulates the transfer of the equitable interest a contract like this creates, with disclosure duties and buyer-cancellation windows attached… The school keeps the state-by-state version of that current; what belongs here is the shape and the caution, not the map.” Backed by ledger card L-008 (ruled 2026-08-24): “School treatment: the living state-by-state module, kept current, with citations, revised as legislatures move.” And by the Depth Test (L-013), which classifies exactly this as SCHOOL-DEEP: “VOLATILE (laws in motion…) — anything with an expiration date.”

3. What exists today. Nothing. RU5 is a one-line charter row. The automation school has no legal content and correctly shouldn’t.

4. The gap in plain words. This is the pointer the whole “books fossilize, our school doesn’t” business argument rests on (L-013: “every ‘the living version lives in the school’ pointer is an honest membership driver”). It is also the highest-liability item in either school. A book that says “the school keeps it current” and a school that keeps nothing is worse than a book that stayed silent.

5. What “deeper” means. One module, three units, plus a maintenance contract.

• RU6-U1 — Why this page has a date on it (10 min). Objectives: (a) name the categories that move (licensed activity, equitable-interest transfer, disclosure duties, cancellation windows); (b) read a page’s revision date as part of the content; (c) know when to stop reading and call an attorney. Hard disclaimer, per L-012.
• RU6-U2 — The map (reference, not timed). Not a lesson — a maintained table, per state, with citation and last-verified date, and an explicit “no entry” state for jurisdictions not yet reviewed. Per the RE school’s §0 rule and L-008: written fresh, cited, never reproduced from anyone’s materials.
• RU6-U3 — The permanent principle (8 min). Objectives: (a) know the rules where you do business; (b) build the recheck into your own calendar the way P13-U2 builds it into a job; (c) recognise when a structure’s name changed but its regulated substance didn’t. Build: find your own state’s entry, read it, and set your own recheck. You’ll know it worked when you can say what’s true in your state today and what date that answer expires.

6. VIDEO NOTE. U1 and U3 are slide+narration. U2 must never be video — a recorded map is a map that lies within a quarter. It ships as a page with a revision date, and the video points at it. No avatar anywhere in this module: a face attached to a legal boundary reads as advice, and the compliance posture (L-012) forbids it.

N10 · REI school — capital: credit strategy and private money

1. Where it sits. REI school. Two homes: credit strategy into a new RU7 — Where the Money Comes From; private lending into RP9’s non-gated half (RP9 itself is gated on the funding funnel being live).

2. What promises it. Two pointers. B2-APP-F ¶193: “Credit strategy — which cards, which lenders, how to sequence financing across a growing portfolio — is exactly the kind of fast-moving, situation-specific territory this book treats as the school’s job to keep current, not the book’s job to freeze in print.” And B2-CH09 ¶39, on private money: “it’s worth a paragraph here even though the real depth of it belongs to the school… the school carries the rest of this at the depth it actually deserves.” Ledger card L-001 holds the owner’s own material for the first; the naming rule there is binding (third-party services stay unnamed).

3. What exists today. Nothing in either school. L-001 and L-002 are captured and assigned; no unit drafted.

4. The gap in plain words. Two of Book 2’s five school pointers are about money the reader doesn’t have yet, which is the reader’s actual first question. Both are also the two highest-disclaimer topics in the corpus (L-012 flags both). The private-money one carries a securities line the book states carefully and the school must state at least as carefully: pooling several people’s money looks different under securities law than one private note.

5. What “deeper” means. One new module, four units, all carrying the hard disclaimer and attorney/CPA pointers per L-012.

• RU7-U1 — Credit is a track record (12 min). Objectives: (a) understand credit as something tended before you need it, not the week of; (b) name what a lender is actually reading. Build: the reader’s own honest position, written down.
• RU7-U2 — Sequencing, not stacking (15 min). Objectives: (a) understand why card selection is the skill and terms-of-service differ; (b) model the cost of money honestly across a hold period; (c) know what disqualifies a route before you’re standing in it. Per L-001’s naming rule, routes are taught structurally; no third-party service is named.
• RU7-U3 — A single private note (15 min). Objectives: (a) name the four channels that actually work (local investor rooms, self-directed retirement accounts, the county recorder’s own records, referral networks); (b) state the line between a quiet one-off note and an offering; (c) know which side of it you’re on before you advertise anything. You’ll know it worked when you can say in one sentence what would move your own plan from the first category to the second.
• RU7-U4 — What this actually costs (12 min). Objectives: (a) price money end to end, including both sets of closing costs; (b) run the same deal at two rate environments. Ties into RP4’s engine rather than duplicating its math.

6. VIDEO NOTE. All four are slide+narration; the arithmetic in U2 and U4 animates well and needs no screen. No avatar — same reason as N09, plus rule R-C (zero brag): a face attached to “$215k at 0% APR” is exactly the tone the addendum bans. If any avatar clip is used here at all, it is the disclaimer, said plainly, once.

N11 · REI school — every way to buy, at depth

1. Where it sits. REI school, RU2 (The Offer Family) extended from one module to a module plus four short reference units.

2. What promises it. B2-CH09 ¶41, after naming land, notes, mobile and manufactured homes, and auctions and tax liens: “Each of those is a real path with real rules, and each one gets the depth it needs in the school rather than a rushed paragraph pretending to cover it here.”

3. What exists today. RU2 is a charter row covering cash / seller-finance / hybrid. The four named paths have no coverage anywhere.

4. The gap in plain words. The book names four doors and says the school opens them. It doesn’t. This is the least urgent of the four Book 2 pointers — a reader is unlikely to buy on it — but it is the most explicitly enumerated, which makes it the easiest to be caught on.

5. What “deeper” means. Four short units, 10–12 min each, deliberately shallow-but-honest rather than falsely complete. Each: what it is, what makes it different from buying a house, the one thing that loses people money fastest, and where the real depth lives.

• RU2-U2 — Land (no structure to inspect; financing unlike a residential mortgage; value driven by use and access).
• RU2-U3 — Notes (buying the debt instead of the property; stepping into a lender’s position).
• RU2-U4 — Mobile and manufactured (real property vs personal property by title, and how that changes financing, taxes, and what “buying” means).
• RU2-U5 — Auctions and tax liens (a compressed rule-bound process; redemption periods; how homework-skipping loses money fast).

Each ends: **“You’ll know it worked when** you can say, in one sentence, why this path is not the path you’re on right now — or why it is.”

6. VIDEO NOTE. Pure slide+narration, four short decks, ideal for the automated pipeline: these are the cheapest units in this entire audit to produce and among the most linkable. No screen, no avatar.

N12 · U4 — The Trust Ladder, rebuilt to four rungs

1. Where it sits. UNIVERSAL track, existing U4, rewritten and expanded from 1 unit to 3.

2. What promises it. B1v5-06: “Four rungs. Nobody skips one and calls it caution.” The appendix is specific about what’s missing: “Setting the specific look-only, fill-only, and submit-under-ceiling boundaries for a given task, and the habit of reading a first run properly, is Module U4.” Ch10 reuses the ladder across breadth (“look only, then fill, then submit — except here you’re climbing it across breadth instead of depth”), and Ch11’s appendix entry applies all four to a portal.

3. What exists today. U4 (1,503 words) — a good module teaching a two-rung ladder: bottom rung (look only) and top rung (irreversible waits), plus receipts, the quiet contract, a promotion rule and a promotion step. It is the module the rest of the corpus leans on hardest (P6, A2, A3, D3, D4 all cite it).

4. The gap in plain words. The two rungs U4 teaches are the two ends. The book’s middle rungs are where the actual work happens: fill but don’t submit (do the whole job, stop one step short of the irreversible act) and submit under a ceiling (act freely inside a fence you drew, report afterward). Without them, a learner has only two settings — free, or gated — which is exactly the binary Ch8 spends a chapter dismantling. U4 also lacks the first-run reading discipline, which is short and load-bearing: watch whether it did what you meant rather than what you said, and watch what it does when it isn’t sure.

5. What “deeper” means. Three units. The existing 1,503 words become U4-U1 and U4-U3 with light edits; U4-U2 is new.

• U4-U1 — Four rungs, one at a time (15 min). Objectives: (a) name all four rungs and what each one lets you stop doing; (b) place each of your own jobs on one; (c) say why skipping to the fourth is the opening sentence of this book in a new coat. Build: the existing audit, re-sorted across four rungs instead of two.
• U4-U2 — The two middle rungs (20 min). Objectives: (a) set a job to prepare completely and stop one step short of the act you can’t take back; (b) set a job to act on its own inside a boundary you drew on purpose, and read the report after; (c) explain what evidence moves a job from one rung to the next. Build: move one job to rung two and one to rung three. You’ll know it worked when one job hands you a finished thing to release, and another one acts without you and tells you afterward.
• U4-U3 — Reading a first run (12 min). Objectives: (a) watch for what you meant rather than what you said; (b) watch what it does at the edge of what it knows; (c) treat a stop as evidence you can trust the next rung. Build: read one first run against both questions and write down what it taught you. You’ll know it worked when you can name the gap between what you asked for and what you meant, on a real run.

6. VIDEO NOTE. U4-U1 is the school’s signature slide — four rungs, climbed, one at a time, and it will get reused in six other modules; build it once, properly. U4-U2 needs a screen recording per rung (the finished-but-unsent artefact; the ceiling being crossed and reported). Avatar: the first lockbox code. “I remember the specific discomfort of sending the first one… I sent the code. I put the phone down.” That is the emotional beat the entire trust movement rests on, and it is wasted on a slide.

N13 · The retired-cast and locked-numbers remediation sweep

1. Where it sits. All 25 lesson files, all 6 demos, DEMO-LIBRARY-SPEC’s registry table, SCHOOL-MODULE-SPEC’s cross-module continuity note, and CURRICULUM-MAP’s demo table.

2. What promises it. STORYCRAFT-DIRECTIVE-v5 §DOCTRINE: “Old cast dead… Banned: the retired cast.” And “Missing numbers stay missing.” SCHOOL-MODULE-SPEC currently instructs the opposite: “cast may be referenced by first name only as gallery examples (Maya’s rent-check build, Dan’s price-check build, Theo’s spoken restock) with the book’s locked numbers — never new numbers.” DEMO-LIBRARY-SPEC repeats it: “Numbers: ONLY the locked cast numbers above/in the book’s continuity ledger.”

3. What exists today. Maya, Dan, Lena, Rosa and Theo appear in zero v5 chapters. They appear across the school in: P02 (Maya, six hours a week), P04 (Dan’s Monday, two hours), U3 (Theo, forty-five minutes), U4 and P06 (the recruiter, seven and a half hours a week / ninety minutes a day), A1 (fifteen hours a week → three to four), A2 (eighteen hours → about four), A3 (two hours forty minutes), D1–D6 (all of them, all their numbers).

4. The gap in plain words. The school is publishing named people and specific performance numbers that the current book does not contain and that the current doctrine forbids. Book 1 v5 is conspicuously careful about this — “I’m not going to invent one to make this paragraph land harder” appears twice in the closing chapter — and the school’s numbers, printed alongside, read as the thing the book just refused to do. This is the fastest-to-fix, highest-embarrassment item in the audit.

5. What “deeper” means. Not deeper — a correction sweep, done once, in one pass, before any new writing:

1. Decide the persona policy. Recommended: personas become unnamed and generic (“a property manager with forty doors”, “a two-person shop”) in modules, and the demo library keeps role-based personas with no first names. The v5 book’s own Lane-2b practice — a character framed once, fresh name, never reused across chapters — is the alternative if the owner wants names back, but it requires new names and one framing line per demo.
2. Strip every performance number that isn’t the learner’s own. Nine modules and six demos. Replace with the shape of the saving, not its size: “a full Monday” instead of “340 rows”; “most of an evening, every month” instead of “two hours and forty minutes.”
3. Amend both specs. DEMO-LIBRARY-SPEC’s “locked cast numbers” rule and SCHOOL-MODULE-SPEC’s cross-module continuity note both cite a ledger that no longer governs. Replace with: no performance claim in any module or demo except one the learner produces themselves.
4. Keep the two numbers that survive. The book’s own arithmetic that is labelled as an example — Ch7’s per-diem model, Ch12’s $380/$2,200 — may be reused, labelled the same way.

6. VIDEO NOTE. This is the gate on all video production. Every number spoken in a cloned voice is a number that has to be re-recorded when it changes; a module scripted around “six hours a week” is a module that gets re-cut. Do this sweep before a single slide is generated, and write the corrected modules so no spoken sentence contains a performance figure at all.

N14 · P00 — Your First Ten Minutes

1. Where it sits. PLATFORM track, new module P00, before P01. First platform module; no prerequisites beyond an account.

2. What promises it. B1v5-15’s invitation sends readers into the school with nothing but a book behind them: “The school is where the parts of this book that don’t fit in a chapter actually live — the click-by-click version of everything I’ve described to you in plain words.” And B1v5-14’s room objection defines who arrives: “I’m not going to learn how to deal with code and all that stuff — even if the AI knows how to write it, setting up websites and programs and figuring out where and how to use it is more than I’ll ever get into.” That person is the target reader, and the first thing they meet in this school today is P01 step 1, which assumes an account already exists.

3. What exists today. Nothing. P01’s “Before you start” says “All you need is your ProcessAutomater account” — the account is a prerequisite with no module behind it. U1–U3 are pen and paper and explicitly pre-platform.

4. The gap in plain words. The school has a beginning for the reader who is thinking about automation and a beginning for the reader who already has an account, and nothing for the reader in between — who is the majority. Note the framing constraint: Ch4’s whole argument is that there is no setup, so this module cannot be called “setup” or it contradicts the book. It is a tour, not an installation.

5. What “deeper” means. Two units, deliberately short.

• P00-U1 — What’s actually here (10 min). Objectives: (a) name the four places anything lives (your jobs, your results, your Routines, your assistant); (b) know what a free account gets you; (c) find the one question the platform asks about everything you build. Build: open each of the four and say what’s in it. You’ll know it worked when you can find your own results drawer without hunting for it.
• P00-U2 — Connecting the first account, safely (12 min). Objectives: (a) understand that a login arrives by a texted link and is never typed into an automation; (b) connect exactly one account — the one your first job needs, nothing wider; (c) know where to see it and how to take it back. Build: one connection, made and inspected. You’ll know it worked when one account is connected, you can see when it was last used, and you know where the button is that takes it away.

6. VIDEO NOTE. Almost entirely screen recording — this is the one module where the screen is the content, and slides would be a lie. Keep it under ten minutes total and re-record it whenever the interface moves; script it so no spoken sentence names a button, only a destination, so the narration survives a redesign. Avatar: the opening thirty seconds. The room objection, answered directly, by the person who heard it: “the whole barrier he’d just described was never actually required.” That is the only sales moment in the entire school and it belongs at the front door.

N15 · P08 — the correction pass

1. Where it sits. PLATFORM track, existing P08, extended from 1 unit to 2.

2. What promises it. B1v5-04 and the appendix’s P05 walkthrough, which is entirely the correction pass: “Expect it to be close, not perfect… Say the missing rule once, in the same plain words as the first description. Not a rewrite — an addition.” And what it hands back: “A working Routine built from a conversation instead of a demonstration, with the correction habit that keeps ‘close enough’ from quietly becoming ‘good enough’ for good.” The appendix also names what the module owes: “How a plain-language description gets matched against something already performed, proven, vetted, and stored — instead of assembled fresh from a guess.”

3. What exists today. P08 (1,500 words) is one of the better modules. It teaches saying a job into existence, the check-what-it-knows step, watching the live build, reading the receipt, confirming it saved, scheduling it. Its “when it goes sideways” gets close — “If it’s off, say the task again, more specifically” — but that is a re-try, not a correction pass.

4. The gap in plain words. Three things are missing and all three are the chapter’s actual argument. First, the correction pass as a taught habit rather than a failure recovery: the first description will leave something out, always, “because that’s what happens every single time a person who knows a job in their bones tries to hand it to someone else.” Second, “performed, proven, vetted, stored” — the reason what comes back isn’t a guess, and the school’s only honest answer to “how do I know this works before I’ve watched it.” Third, the agentic-versus-automation distinction: improvising is right the first time and wrong the six-hundredth, “the task graduates into something closer to a recipe with the guesswork already taken out of it.” U4 gestures at this distinction; P08, which owns it, doesn’t state it.

5. What “deeper” means. Two units. Existing content becomes P08-U1 with the “performed, proven, vetted, stored” paragraph added.

• P08-U1 — Say it (25 min; existing build, lightly extended). Add two objectives: explain why what comes back has already been performed, proven, vetted and stored rather than assembled fresh; distinguish improvising a task from running a proven one, and say which you want on the six-hundredth run.
• P08-U2 — Correct it once (20 min, new). Objectives: (a) expect the first run to be close and know why; (b) find the gap by comparing the result against your own instinct, not by guessing at what might be wrong; (c) say the missing rule as an addition, not a rewrite; (d) stop watching once it matches what you meant. Build: deliberately under-describe a job, run it, find the unstated rule, add it. You’ll know it worked when the second run does the thing you meant and not the thing you said, and you only had to say one more sentence to get there.

6. VIDEO NOTE. U2 is a screen recording and it is the most persuasive thirty seconds available to this school: a flag list of eleven names, one spoken sentence, a flag list of three. Record it as one unbroken take. U1’s “performed, proven, vetted, stored” is four slides. Avatar: none needed — the screen carries it, and the module already has a strong non-authorial story in the grooming-shop owner at her kitchen counter.

N16 · U7 — What Survives

1. Where it sits. UNIVERSAL track, new module U7, taught last, replacing U5’s position in the recommended path. U5 (relief test) survives as an elective — see the note below.

2. What promises it. B1v5-14 in full: “Name the process. Sharpen the judgment. Own the list. Everything else in this book — every automation, every Routine, every system you’ve built chapter by chapter — is scaffolding around those three things, and scaffolding is supposed to come down eventually and leave something standing.” The appendix declines to give Ch14 a build entry because it “doesn’t leave a step-by-step behind” — which is true of a platform build and false of a Universal-track concept lesson, whose builds are pen and paper by design.

3. What exists today. U5 — What Stays Human (1,163 words), a good module built on old ch16 (future-proof): the relief test, the relief-test list, the quarterly hour. Book 1 v5 contains no relief test, no birthday card, no quarterly ritual — the words “relief,” “birthday” and “quarterly” appear in zero v5 chapters.

4. The gap in plain words. The book’s final argument — the answer to “if the ground keeps moving, what’s actually mine?” — has no lesson. It is also the argument most likely to close a sale, because it is the objection every capable owner raises. Meanwhile the school’s closing module teaches a concept the current book dropped. U5 shouldn’t be deleted: the Trust Doctrine still names it as canon (“whatever the reader keeps by CHOICE (the relief test)”), so it stays, re-labelled as an elective and re-anchored to the doctrine rather than to a chapter that no longer exists.

5. What “deeper” means. Three units, all pen and paper.

• U7-U1 — Name the process (15 min). Objectives: (a) understand that a described process doesn’t care what runs it; (b) write one of your own running jobs out as a description a stranger could follow; (c) explain why the tool is a rental and the naming is the purchase. Build: one job, described in plain words on paper, independent of any tool. You’ll know it worked when you could hand your description to someone using entirely different software and they’d build the same thing.
• U7-U2 — Own the list (15 min). Objectives: (a) distinguish an audience from a relationship; (b) inventory where each of your own lists actually lives and who controls it; (c) name what you’d lose overnight if any one of those platforms changed its terms. Build: the inventory, plus one export you actually perform. You’ll know it worked when the most valuable list in your business exists somewhere you can carry it out of.
• U7-U3 — Sharpen the judgment (12 min). Objectives: (a) name the failure mode that never goes away — fluency mistaken for truth; (b) describe the muscle that gets stronger every time you verify something; (c) say why nobody can revoke your access to it. Build: write down the standard you’re actually applying when you check a finished thing. You’ll know it worked when you can state your own standard in a sentence, without reaching for a rule you read somewhere.

6. VIDEO NOTE. All slide+narration, and this module has the best archival slide material in the corpus: the algorithmic-timeline change, the API price change, the Dynabook described before the parts existed. Avatar: the Twitter following. “It felt owned. It was never owned. It was borrowed shelf space in someone else’s store.” Said to camera by someone who ran a business on that shelf. One clip; the historical material carries the rest.

N17 · P16 — The Review Queue at Volume, and Error Triage

1. Where it sits. PLATFORM track, new module P16, after P12. This is the module a learner needs in month two, and it is the one that decides whether they’re still here in month four.

2. What promises it. B1v5-08’s counterweight: “A queue with three items a week gets read. A queue with three hundred gets rubber-stamped… The discipline isn’t ‘make a person approve everything that touches money.’ It’s ‘make sure the only things that reach a person are the things a person’s approval actually changes the outcome of.’” And B1v5-09’s posture: “You’re the exception handler… Real oversight is being the person who shows up exactly when the system has hit the edge of what it knows — and nowhere else.” Plus the trap: “‘It’s faster if I just do it’ doesn’t disappear once you’ve built a system that stops and asks — it comes back dressed as ‘it’s safer if I just check it myself.’”

3. What exists today. Warnings, in three modules, all correct and none actionable. A1: “You catch yourself approving a batch without really reading it.” A2: “You’re tempted to turn the review queue off before you’ve actually earned it.” A3: “A long streak of correct drafts is exactly what makes skimming feel safe.” No module teaches what to do when the queue has grown, when four jobs failed overnight, or how to tell a job that broke from a source that changed from a row that was always bad.

4. The gap in plain words. The school teaches learners to build queues and never teaches them to run one. The failure it produces is predictable and terminal: a learner reaches thirty items a day, starts skimming, gets burned once, and turns everything off. Error triage has the same shape — P3 teaches “check the row, not the job” for one failure mode, and nothing teaches the general skill of reading a morning’s worth of failures and knowing which three actually mean something.

5. What “deeper” means. Three units.

• P16-U1 — Why your queue got long (15 min). Objectives: (a) count what reaches you and sort it by whether your approval changed the outcome; (b) retire the reversible items; (c) set the threshold that lets the routine ones through. Build: a one-week queue census and one retirement. You’ll know it worked when your queue is shorter and you can name what you removed and why removing it was safer, not riskier.
• P16-U2 — Reading a bad morning (20 min). Objectives: (a) sort overnight failures into the three shapes — the source changed, the source went quiet, the row was bad; (b) fix the one that’s actually yours and leave the two that aren’t; (c) know when a repeated failure is a rule you haven’t written. Build: triage a real batch of failures, or a supplied one. You’ll know it worked when you handled a morning of failures in the time it used to take to panic about one.
• P16-U3 — Not showing up (10 min). Objectives: (a) catch “it’s safer if I just check it myself” in your own week; (b) name the things that genuinely still need you; (c) stop reviewing everything else on purpose. Build: your exception list, written and honoured for a week. You’ll know it worked when you went a full week without opening something that ran correctly.

6. VIDEO NOTE. U2 is the screen recording — a real queue with real failures, sorted live. U1 and U3 are slide+narration. Avatar: the line about the lanyard. “That’s not diligence. That’s the old bottleneck wearing a lanyard that says quality control.” It is the funniest sentence in Book 1 and the one most likely to change a behaviour, and it needs a delivery, not a caption.

N18 · P17 — Access, Logins, and Taking Keys Back

1. Where it sits. PLATFORM track, new module P17, after P06 (workers with hands), where access first becomes a real question.

2. What promises it. B1v5-16’s P06 entry names the walkthrough the school owes: “Grant access the way you’d grant a new employee a login: scoped to the one system it needs, nothing wider. Keep it visible… Know how to pull it back at any time, the same way you’d take back a key… the access-scoping walkthrough [is] Module P06.” Reinforced by Ch11’s build markers and, from the other direction, by REI-SCHOOL-CHARTER Guardrail 4 — a real incident of plain-text credentials sitting in a shared document.

3. What exists today. One vocabulary line, repeated in five modules: “Your logins (kept safe) — access that arrives by a texted link, never typed into the automation.” That is a reassurance, not a lesson. P06 mentions the texted link and stops.

4. The gap in plain words. A learner is taught to hand credentials to something that will use them unattended, and is never taught scoping, visibility, revocation, or what to do when a person leaves. The reassurance actually makes it worse: told the logins are safe, nobody asks the next question. Every business-owner reader has this question by week two, and the school’s silence on it is the single most likely reason a cautious owner stops.

5. What “deeper” means. Three units.

• P17-U1 — Scoped to one thing (12 min). Objectives: (a) grant access the way you’d grant a new employee a login — the one system it needs, nothing wider; (b) recognise over-granting as the default mistake; (c) name what each of your running jobs actually needs to reach. Build: audit every connection you’ve made and narrow one. You’ll know it worked when at least one connection reaches less than it did yesterday and everything still runs.
• P17-U2 — Seeing what was used, and when (12 min). Objectives: (a) find the record of when an access was used and for what; (b) build the habit of reading it on a rhythm you’ll keep; (c) know what an unexpected entry looks like. Build: read one week’s access record. You’ll know it worked when you can say what touched which account this week without asking anyone.
• P17-U3 — Taking the key back (12 min). Objectives: (a) revoke one access and confirm it’s gone; (b) write your own offboarding step for when a person or a vendor leaves; (c) know which of your jobs would stop, and decide whether that’s the right outcome. Build: revoke something, on purpose, and watch what breaks. You’ll know it worked when you’ve taken a key back and know exactly what stopped working when you did.

6. VIDEO NOTE. Almost entirely screen recording, and it should be shot deliberately dull — this is the module where reassurance is worth less than watching someone do it. One slide sequence: the new-employee analogy, four frames. Avatar: none. A face on a security module reads as a promise; the screen should make the promise instead.

N19 · U8 — Getting Your Records Ready

1. Where it sits. UNIVERSAL track, new module U8, between U3 and P01. Pen-and-paper plus one spreadsheet; no account required.

2. What promises it. B1v5-12, stated as a prerequisite nobody warns about: “you cannot automate a paper process. I didn’t make real headway on any of this until paper was gone completely — inspection sheets, work orders, invoices, all of it moved off clipboards and into something a system could read. Digitizing isn’t automating. But nothing automates until the paper is gone, and that’s a real, unglamorous prerequisite nobody warns you about.” Ch14 adds the other half: “the actual list, sitting in a system you can export from, back up, and carry with you.” The school’s own P3 knows the consequence — “Bad rows in, bad runs out” — and the Book 3 addendum’s C2-U2 states the test exactly: check whether the underlying dates are even populated, because “the answer determines whether their build is real or a design.”

3. What exists today. One bullet in P3’s failure list. Nothing else.

4. The gap in plain words. Half the platform track assumes the learner has a list with real columns, a tracker with real statuses, and records something can read. Many won’t, and the school lets them find out at step 4 of P3, after they’ve already invested in a build. Worse, the failure is silent: a job that runs against half-populated records produces results that look fine. This is a fifteen-minute lesson that prevents a whole class of month-two abandonment.

5. What “deeper” means. Three units.

• U8-U1 — Where your paper still is (12 min). Objectives: (a) find the steps in one process that still live on paper or in someone’s head; (b) distinguish digitising from automating; (c) sequence them honestly. Build: one process, walked, paper marked. You’ll know it worked when you can name the exact step where a process you want to automate stops being readable by anything.
• U8-U2 — Is your list actually a list (15 min). Objectives: (a) check that the fields your future job will read are actually populated, not just present; (b) split a vague column into the pieces a job needs; (c) recognise the difference between a build and a plan. Build: open your real list, count the blanks in the columns that matter. You’ll know it worked when you know what percentage of your own rows would work today, and you fixed the column that was worst.
• U8-U3 — The list you can carry out (12 min). Objectives: (a) find where each list lives and who controls it; (b) export one, for real; (c) explain why the export is the asset. Build: one export, performed. You’ll know it worked when your most valuable list exists in a file you hold.

6. VIDEO NOTE. Slide+narration with one short screen insert for the export. This module produces the best “before” imagery in the school — a clipboard, a half-filled column — and stock-free abstract slides will do it fine. Avatar: one line, on U8-U1. “Digitizing isn’t automating. But nothing automates until the paper is gone.” Said flatly, it inoculates a learner against a month of wasted effort.

N20 · P05, A1 and A2 — re-anchoring the orphaned modules

1. Where it sits. Existing P05 (Assistant in Your Pocket), A1 (Content Factory), A2 (Sales & Leads). No new IDs; this is a source-anchor decision plus targeted edits.

2. What promises it. SCHOOL-MODULE-SPEC’s registry rows cite 10-assistant-in-your-pocket.md, 07-content-factory.md and 08-sales-and-leads.md. None of those files exists in Book 1 v5. What does exist: the vocabulary law still defines “your AI assistant — the phone interface… it EXECUTES” and the texted link, so the capability is canon even without a chapter; Ch5 contains a full content-pipeline paragraph (“record yourself once, saying the thing you’d say anyway, and let a Routine push it through editing, formatting, captions, and a schedule across every platform at once”) plus the cloned-voice-and-likeness note and the one judgment that survives (“whether the message is actually something you’d stand behind”); and Ch9 supplies the verify-before-print discipline A1 already teaches.

3. What exists today. Three complete, competent modules with no book behind them. A2 is the most exposed: the anti-spam covenant — four rules stated as “encoded, not something sitting in a queue” — has no anchor in any of the three books. Its nearest home is the REI charter’s RU5 (The Covenant), which is a charter row, and the Book 3 school’s C5-U2, which is a different school’s compliance unit.

4. The gap in plain words. Three modules are teaching things the current corpus doesn’t back. That is survivable for P05 and A1 (both re-anchorable in an afternoon) and serious for A2, because A2 makes conduct promises about how outreach behaves and there is nothing above it saying so.

5. What “deeper” means. Three decisions, then edits — no new units except where noted.

• P05. Re-anchor to the vocabulary law rather than a chapter, and add one unit, P05-U2 — The one sentence that isn’t a job yet, connecting the assistant explicitly to P08’s say-it door (the two are currently taught as unrelated). Also add the Ch11 access-scoping point: the assistant’s connections are the ones most likely to be over-granted. 15 min.
• A1. Re-anchor to B1v5-05’s content-pipeline paragraph. Add the two things the chapter has that the module doesn’t: the honest limit on a cloned voice and likeness (“no clone of your own voice can make that check for you”), and the fact that the pipeline itself needed more work before it was trusted unattended — which is the school’s own product story and should be told, not hidden. Strip A1’s fifteen-hours figure per N13.
• A2. Escalate the covenant. It needs to live somewhere above a single elective — recommended: a short U-track covenant lesson (or a numbered section of U4) that A2, the REI school’s RU5, and any future outreach module all point at. Until that decision lands, A2 should carry a line naming where the rules come from, because “encoded, not something sitting in a queue” is a product claim and needs an owner.

6. VIDEO NOTE. A1’s cloned-voice honesty beat is the one place where the production method should appear in the lesson — the school is narrated by a cloned voice and illustrated by an avatar, and saying so inside the module about cloned voices converts a possible objection into the school’s most credible demonstration. That is the single strongest avatar moment available anywhere in this corpus: the avatar disclosing that it is one, in the module about using one honestly. P05 is screen recording (a spoken sentence, a phone, a result). A2 is slide+narration only.

N21 · U9 — Measuring What You Got Back

1. Where it sits. UNIVERSAL track, new module U9, taught alongside P02 and revisited at P10.

2. What promises it. VOICE-AND-RULES §2 makes it the narrator’s defining habit: “Obsessive tracker of hours saved.” Every one of the 25 modules asks for it — “Post the before number and the after number” (P03), “one real before-and-after” (A3), “how long ten cold outreach messages used to take you… versus how long it took to review ten” (A2). U2 asks the learner to estimate their own grind. And Book 1 models the method twice: Ch7’s per-diem model (“look up the real rate where your doors actually sit, run this same model with your own number”) and Ch12’s $380-against-$2,200 framing.

3. What exists today. Nothing teaches it. Fifteen modules ask for a number the school never teaches anyone to produce, which means the gallery fills with guesses.

4. The gap in plain words. The school sells time back and never teaches anyone to count it. Two consequences: learners can’t tell whether a build was worth it, so they can’t choose the next one well; and the gallery — “the textbook the next learner reads” — fills with numbers nobody measured, which is the exact failure the truth doctrine exists to prevent.

5. What “deeper” means. Two units.

• U9-U1 — Measure the before, before (12 min). Objectives: (a) time a task honestly, including the part where you switch back into it; (b) count frequency across a month, not a week; (c) write it down before you build anything. Build: one baseline, measured, not estimated. You’ll know it worked when you have a real before-number for the task you’re about to teach, recorded before you taught it.
• U9-U2 — What it’s actually worth (15 min). Objectives: (a) build a per-unit model in your own numbers, the way Ch7 builds a per-diem; (b) count the checking time honestly as part of the after-number; (c) name the saving you can’t count and refuse to invent a number for it. Build: one before/after, honestly stated, ready to post. You’ll know it worked when your posted number is one you could defend to someone who wanted to argue with it.

6. VIDEO NOTE. Pure slide+narration; the per-diem model animates in five frames. Avatar: one line, and it’s a refusal — “I can’t tell you the dollar value of that morning, and I’m not going to invent one to make it sound bigger than it is.” Said to camera in a school full of before/after claims, that sentence is worth more than any of them.

N22 · P18 — Working With a VA or a Team

1. Where it sits. PLATFORM track, new module P18, after P09.

2. What promises it. B1v5-12’s closing move: “more than once I’ve handed a contractor the same recorder I used to build my own system, so he could stop losing hours to invoices he didn’t know how to write. Helping the people in your own relay automate their piece of it isn’t charity so much as maintenance — a stronger link makes a stronger chain.” And Ch15’s culture line: “giving someone else the thing you figured out is one of the most quietly powerful moves available to you.” REI-EXPANSION §8 build #14 names the artefact: the invoice generator for contractors, “the friend-hero automation — give it away.”

3. What exists today. Nothing. Every module is written to a single operator (“you,” singular), which is correct for the voice and wrong for the reader — the target reader has an admin, a VA, a bookkeeper, or a crew.

4. The gap in plain words. A learner who builds ten jobs and then hires help has no lesson on what to hand over, how, or what happens to a job when the person who taught it leaves. The gallery culture (publish a build, install someone else’s) already models the mechanic; nobody has applied it inside one business.

5. What “deeper” means. Two units.

• P18-U1 — Handing a job to a person (18 min). Objectives: (a) decide which jobs a teammate runs and which stay yours; (b) hand over a job with its receipts, so they can see what it did without asking you; (c) keep the fence where it belongs when someone else is standing at the chain. Build: hand one job to one person and watch a week. You’ll know it worked when somebody else ran a job you taught, and you found out from the receipt.
• P18-U2 — Building something for someone else (20 min). Objectives: (a) find the task that costs a person in your relay the most and isn’t yours at all; (b) build it for them; (c) name why a stronger link is worth the afternoon. Build: one job, built for someone else, handed over. You’ll know it worked when someone you work with is running something you built for them, and it isn’t your business it’s saving.

6. VIDEO NOTE. Slide+narration with one screen insert (handing a job over, seeing the receipt). Avatar: the contractor and the recorder. “I’d rather be the reason a good contractor stays in business than watch him quit the trade the way the last plumber did.” That line is the school’s culture in one sentence and it should be said by a person.

N23 · P19 — What It Costs to Run

1. Where it sits. PLATFORM track, new module P19, late — after P10, as part of the “you own this now” arc.

2. What promises it. B1v5-03’s arithmetic, which the whole book leans on: “everybody pays a little more the first time… you pay with your time, or you pay by having a quality AI step in and perform it once, and then it’s programmatically reproducible forever.” B1v5-14’s cautionary half — small operators “had to either start paying a bill that hadn’t existed the year before, or lose the automated piece of their marketing outright.” And REI-EXPANSION §6, which makes usage pricing an explicit business model: “user price = multiple of our cost + pass-through of their AI operating costs.”

3. What exists today. Nothing. Cost appears nowhere in the school except as the thing automation saves.

4. The gap in plain words. A learner scales from one job to twenty and has no framework for what that costs, no way to compare a build’s cost against the subscription it replaces, and no early warning when a job is expensive for the value it returns. Book 1 argues “pay once” as a doctrine and the school never teaches anyone to check whether their own case actually obeys it.

5. What “deeper” means. Two units.

• P19-U1 — Pay once, in your own numbers (15 min). Objectives: (a) separate the one-time cost of teaching from the running cost of running; (b) compare one build against whatever it replaced, subscription included; (c) recognise the case where pay-once doesn’t hold and say so. Build: one honest comparison. You’ll know it worked when you can say what one of your builds cost to make and what it costs to keep, and which of those numbers is bigger.
• P19-U2 — Watching your own usage (12 min). Objectives: (a) find what your jobs actually consume; (b) spot the job that’s expensive relative to what it returns; (c) decide whether to narrow it, re-schedule it, or retire it. Build: a usage review, plus one change. You’ll know it worked when you retired or narrowed one job on cost grounds and didn’t miss it.

6. VIDEO NOTE. Slide+narration with a short screen insert for usage. Script it with zero absolute figures — per addendum rule R-G, any printed cost dates the material, and a cloned-voice narration containing a price is a re-record waiting to happen. Avatar: none.

N24 · P03 — the risk framing, throughput, and the door

1. Where it sits. Existing P03, one new unit and one reframe. No renumbering.

2. What promises it. B1v5-10 and its appendix entry. The ramp is framed by risk, not familiarity: “a slice small enough that a mistake costs you an afternoon of checking, not a real position.” The reframe is the chapter’s thesis: “What was actually scarce, the entire time, was never opportunity. It was throughput.” And the prerequisite the chapter closes on: “Scale, at the volume this chapter has been describing, only works where a door is standing open for it.”

3. What exists today. P03 (1,433 words), one of the two or three best modules in the corpus. The 10 → 50 → whole-list ramp is there, spot-check culture is named, the courtesy pacing rule is there, and the “check the row, not the job” instinct is taught well. It frames the ramp as testing your own demonstration — true and useful, but a different reason than the book’s.

4. The gap in plain words. Three additions, none large. The ramp should be sized by what a mistake costs, not by row count alone. The throughput reframe is the chapter’s actual argument and the thing that changes what a learner does next (widen the funnel instead of protecting yourself from it). And the door prerequisite tells a learner before they build why some lists scale and some need P06 instead — which currently they discover by failing.

5. What “deeper” means. One reframe plus one new unit.

• P03-U1 (existing build, edited): add one objective — size your first slice by what a wrong run would cost, not by how many rows feel safe.
• P03-U2 — Throughput, not opportunity (12 min, new). Objectives: (a) name the constraint that’s actually scarce in your business; (b) recognise when you’ve been shrinking to fit a capacity problem instead of fixing it; (c) check whether a door is standing open for the list you want to run, and know that P06 is the answer when it isn’t. Build: find one thing you’ve narrowed, capped, or unsubscribed from to keep up, and decide whether to reopen it. You’ll know it worked when you can name one opportunity you turned down for capacity reasons that you could now take.

6. VIDEO NOTE. U2 is slide+narration and carries Ford’s line and the twelve-and-a-half-hours-to-ninety-three-minutes figure, which is the best single number in the book. Avatar: the unsubscribe. “I did the thing you’re not supposed to do. I started unsubscribing.” One clip; the confession is the lesson.

N25 · P06, P09 and P10 — three named sub-builds and one reframe

1. Where it sits. Existing P06, P09, P10. Depth passes, no renumbering.

2. What promises it. Three specific appendix lines. P06: “Teaching a specific portal field by field, setting the trust-ladder gates for a portal, and the access-scoping walkthrough are Module P06” — the three gates on a portal specifically (read the screen back; complete the form and stop; submit only once the first two have held). P10: “The walkthrough — a personalized-output Routine, for anything that feels too close to hand to automate” — the letters-on-the-printer build, five steps, ending on “keep the one piece that has to stay a person’s — the signature, the ink.” And Ch13’s reframe: “Most owners spend their whole career reacting to a day that arrives already full… A Routine reverses that relationship.”

3. What exists today. P06 (1,321 w), P09 (1,515 w), P10 (1,373 w). All three are adequate and all three ship a learner to a working build. P06 stays entirely on the bottom rung and never shows the middle gates. P09 teaches order, branch and schedule cleanly but frames a Routine as tidying a pile rather than designing a day. P10 has the morning stack but not the personalized-output build.

4. The gap in plain words. Nothing here is broken; three named things are missing. The P10 one matters most: the letters build is the book’s answer to “personal and scaled were never actually opposites” — the objection every service business raises — and it is currently taught nowhere.

5. What “deeper” means. Three additions.

• P06-U2 — The three gates on a portal (18 min). Objectives: (a) have it read the screen back — which field it believes is which — before it touches anything; (b) let it fill and stop, and check its work against what you’d have typed; (c) move to submit only after the first two hold, run after run, because a submitted request is a record with a real party attached. You’ll know it worked when you’ve watched it fill a real form correctly and not submit it.
• P09-U2 — Designing the day (12 min). Objectives: (a) decide in advance what a day contains instead of reacting to it; (b) retire a recurring decision by writing it into a branch on a clear-headed morning; (c) name the difference between a chain that runs when you start it and one that runs because the day arrived. You’ll know it worked when a decision you used to make weekly is written down and hasn’t needed you since.
• P10-U2 — The thing that felt too personal (25 min). Objectives: (a) pull a real segment, not a generic list; (b) write the real version yourself enough times, across enough situations, that it learns your shape of “personal” rather than a template’s; (c) let the branches form per situation; (d) keep the one piece that has to stay a person’s. You’ll know it worked when something waiting for you is personal enough that you’d sign it, and you only have to sign it.

6. VIDEO NOTE. P06-U2 is screen recording — a filled, unsent form is a picture worth more than a paragraph. P09-U2 is slide+narration and should use Franklin’s table and its two questions verbatim; it is the best-looking slide in the school. P10-U2’s avatar moment is the stack of letters on the printer: “needing nothing from me but my signature and my own handwriting on the envelope, because that’s the part that has to stay a person’s: the ink, the stamp, the proof somebody actually touched it.”

N26 · The cost-defense unit (C6) — noted, not owned

1. Where it sits. Track C of the RE certification school (Book 3’s companion), as a sixth module — COST-DEFENSE-LESSON-2026-09-02 Part 4, “C6 — The Cost Defense: Automated Seller Presentations,” with unit C6-U1. Not the automation school, and it should not be moved here.

2. What promises it. The lesson document itself, which states the placement: “Fits the Track C module table in SCHOOL-RE-CURRICULUM-DESIGN-2026-09-01.md §4 as a sixth module.”

3. What exists today. More than exists for most needs in this audit: a title, four learning objectives, a five-point lesson outline, a worked exercise with real inputs, a falsifiable “You’ll know it worked when,” and five original assessment items with rationales. It is drafting-ready.

4. The gap in plain words. One structural conflict to flag, and one thing worth borrowing. The conflict: SCHOOL-RE-CURRICULUM-ADDENDUM-2026-09-02 §1 rules “No sixth module (integration doc R42). Track C stays five modules” — written the same day. C6 and R42 are in direct opposition and somebody has to choose. The obvious resolution: C6-U1 becomes a unit inside C1 (it is a seller-facing outreach artefact) or inside C5 (it is bounded by what a hosted link may claim), which honours R42 without losing the content. That is a Book 3 school decision, not mine.

5. What “deeper” means — for this school. One borrowable kernel, and it belongs in P12 (N02) as a single unit rather than a module: when a machine prepares a document a customer will read, disclosing what’s in it beats hiding it. Proposed P12-U5 — Showing your work to the other side (12 min). Objectives: (a) let the machine assemble a customer-facing explanation from the same facts it used to compute the answer; (b) disclose every line rather than folding any of them into another; (c) say plainly what the delivery method can and cannot protect. Build: one prepared, customer-facing explanation, assembled from a job’s own outputs, reviewed before it goes. You’ll know it worked when every number a customer can see traces back to something your job actually read.

6. VIDEO NOTE. For the automation school’s borrowed unit: slide+narration, one worked stack, no domain specifics — the general lesson is about disclosure, not about houses. No avatar. (Track C’s own C6, if built there, would want a screen recording of the hosted page and no avatar at all, given its honesty constraint about what the page can’t prevent.)

N27 · P01 — the SOP byproduct

1. Where it sits. Existing P01, one short unit appended.

2. What promises it. REI-EXPANSION §5, logged as the owner’s own product insight: “SOP documents are a byproduct of our process recording… One recording → (a) an executable automation, (b) a written SOP, (c) a training asset for the school, (d) assistant-KB content. Four outputs, one capture.” And B1v5-03’s teaching frame, which already tells the learner to review what it learned like a new hire’s first morning — one step short of noticing they now hold a written procedure.

3. What exists today. P01 ends at “Confirm the result landed.” The readout — the plain-English list of steps it learned — is used as a quality-control artefact and then discarded.

4. The gap in plain words. Every learner produces a written procedure for a task that previously lived only in their head, and nobody tells them it’s valuable. For a business owner this is often worth more on day one than the automation: it is the thing you hand a new hire, the thing a buyer’s diligence asks for, and the thing that makes N22’s handover possible.

5. What “deeper” means. One unit.

• P01-U2 — You just wrote the procedure (10 min). Objectives: (a) recognise the readout as a written procedure you now own; (b) keep it somewhere a person could read it; (c) name the three other things it’s good for — training, handover, and the answer to “how do we do this.” Build: save one readout as a document and give it to one person. You’ll know it worked when somebody who has never done the task could follow your readout and do it by hand.

6. VIDEO NOTE. Slide+narration, one screen insert showing a readout being saved. Avatar: none — but this unit is worth a single strong line in the P01 narration, because “four outputs, one capture” is the kind of sentence that makes someone re-tell the school to a friend.

COUNTS

Shape of the work. Nine new modules (P00, P11, P12, P13, P16, P17, P18, P19, plus U6, U7, U8, U9 on the universal side — twelve counting those), three new demos, four rewrites (U4, P07, P05/A1/A2 re-anchor, P03), six depth passes, one registry reconciliation, one corpus-wide correction sweep, and ten REI-school units that belong to a school nobody has started. Of the 27, exactly two (N06, N13) must happen before anything else is written, and neither of them is a lesson.

IF WE COULD ONLY BUILD THREE

Build N06 (the module-code reconciliation), N13 (the cast-and-numbers sweep), and N02 (P12 — the approve queue) — in that order, and treat the first two as one week of work rather than two projects. The reason is not that they are the biggest gaps; N01, N03 and N08 are all larger holes. It is that N06 and N13 are load-bearing: every module written from today forward inherits either the right module codes or seven collisions, and either an honest number policy or a corpus that keeps printing people the book has disowned — and a school narrated in a cloned voice pays for a wrong number twice, once in markdown and once in audio. They cost days, not weeks, and everything downstream gets cheaper. Then build P12, because it is the only P0 that is half-built already (A3 is genuinely good and can be split rather than started), because it is the law every other trust module points at, and because it is the single lesson most likely to keep a cautious business owner from turning everything off in month two — which is the failure mode this school currently has no defence against at all. N01 (verification) is the fourth build and it is a close fourth; if a fourth slot exists, take it, because “scheduled is not done, verified is done” is the most portable sentence the book owns and the school currently cannot say it.

Real estate school (Tracks A · B · C) — 33 needs

lens: Tracks A (Foundations) · B (Broker) · C (Automated Agency), as designed in SCHOOL-RE-CURRICULUM-DESIGN-2026-09-01.md (“the design”) and extended by SCHOOL-RE-CURRICULUM-ADDENDUM-2026-09-02.md (“the addendum”) · audited 2026-09-02 against Book 3 final draft (B3-FRONTMATTER, HOW-TO-USE, CH01–CH19, App A, App B, References), BOOK3-PREFLIGHT-DESIGN-2026-09-01.md, COST-DEFENSE-LESSON-2026-09-02.md, VOICE-AND-RULES + addendum, SCHOOL-MODULE-SPEC, LESSON-LEDGER.

STATE OF THIS SCHOOL

Nothing is drafted. The entire corpus is two design documents and a preflight — no unit page, no practice item, no glossary, no landing page. Every need below is therefore a build, and the only question worth asking is whether the design, taken at its word, would produce a school that makes Book 3’s sentences true. Read against the finished book, it would not.

Book 3 carries 59 SCHOOL markers across 20 files (18 chapters, plus How to Use This Book and Appendix B): Track A ×24, C5 ×7, C1 ×6, B1 ×5, B3 ×4, C2 ×3, C3 ×3, C4 ×3, “all tracks” ×2, “Track C” ×1, and one single unit-level pointer, C1-U4. All twenty-four Track A markers sit in Appendix B — and Appendix B is not a pre-licensing syllabus. It is a practice on-ramp: the MLS and the keybox, choosing a brokerage, splits and caps, the sphere database on day one, the first thirty days’ cadence, the CMA anatomized, the listing consultation, showings, agent safety, wire fraud on closing day, 1099-NEC and self-employment tax and the reserve and E&O, seven compliance terms, ten scripts, a ninety-day map, and forty vocabulary words. Track A as designed teaches none of it. Its four modules — estates, land use, contracts, agency, license law, disclosures, fair housing, finance, valuation, math, closing — are a licensing-exam topic map, drawn honestly from the standard topic universe, and they do not contain a single unit that answers a single one of those twenty-four markers. The book’s own sentence at App B ¶7 — “Wherever this appendix would need to hand you a number — hours required, exam format, fees, deadlines — it points you instead to the Foundations track” — is answered today by nothing at all, because the design explicitly writes Track A “at the national/generic level” and defers state specifics to a supplement that its own §2 cross-references to a §6 that doesn’t contain it.

That dangling reference is the first symptom of the largest structural failure. The book promises a living state-law map in at least twelve places, starting on the copyright page (“the companion certification school carries the current law for where you practice”) and running through Chapter 10 (“the school’s living map of licensed activity, B1, updated as legislatures move”), five separate entries of Appendix B’s compliance vocabulary, and the References’ closing note (“the honest place to keep a rule current”). The design has no living map. It has a Virginia broker track built from Moseley’s topic lists, a national-generic Foundations track, and an open question to the owner (§8, item 1) asking whether the school is Virginia-only or multi-state — a question the book has already answered on behalf of every reader in the other forty-nine states. There is no data model for a rule, no update cadence, no reviewer, no “verified as of” stamp, and no binding between a rule and the ~935 practice items that will assert it. A school whose central promise is currency has designed no mechanism for currency.

Underneath that sits a routing defect that makes several book sentences false as designed. Appendix B’s reader is, by definition, not yet licensed. Appendix B sends that reader to B1 for disclosure content and referral-fee rules, to B3 for agency and contracts, and to C5 for advertising, fair housing, and contact-and-consent. B1 and B3 are the Broker track — a license upgrade the design gates behind an existing salesperson license and years of practice. Chapter 8 sends a working agent to B3 for the same reason. So the school’s living-law answer, as designed, lives in a track most of the people the book points at cannot enroll in, and the generic Track A units that could carry it (F2-U2 agency, F2-U3 license law, F3-U1 disclosures) are written as national-generic exam content with no state layer at all.

Track C is the differentiator and the most under-built thing in the corpus. Appendix A names seventeen builds the book hands the reader. Five have a Track C unit that builds them (A-2, A-3, A-5, A-6, A-16). Four are half-covered (A-7’s stage tracker only, A-8’s quiet-file flag only, A-9’s boundary but not its build, A-14 at solo scale only). Eight have no unit at all — the rented-edge audit, the watch layer, the listing marketing set, the dual pipeline, the TC handoff, the team intake standard, the broker rollout, and the authority review. Chapter 14 says the routing rules and the intake floor’s exact fields “belong to the school’s team module, walked step by step”; the team module is one unit about hiring workflows and contains none of it. Chapter 11 is an entire chapter about listing marketing sets marked C1; Track C has no marketing unit. Worst of all, the authority ladder itself — prepare, propose, propose-with-track-record, authorized, human-on-exception — is taught nowhere, despite every one of the seventeen builds closing with a paragraph naming where it sits on that ladder, and despite Chapter 18 and A-17 making the review of that ladder the capstone of the whole book. The addendum’s nine new units are good units, correctly sourced and honestly bounded; they extend the relationship layer and leave the book’s machine layer where they found it.

Three smaller defects that cost real money if they ship. The design’s §1 says Track A has 10 units; §2 lists 11 (2+3+2+4), and §5 computes the quiz bank off 10 — so the advertised “≈935 original items” is ten items short of its own table, and the addendum’s “12 modules / 37 units” should read 38. Three Track C builds are taught as running when the book plainly labels them roadmap: the week’s call list (A-6, “roadmap, not running today”), the anniversary client-value packet (A-3 step 5, “that piece is roadmap”), and the time-value dashboard view (A-14 step 5) — and C1-U3, C2-U2 and C2-U3 respectively assign each as a student build, which owner rule R-H forbids. And the design never restates the secrecy line or the no-invented-UI rule for a track whose every unit is a platform walkthrough.

Finally, COST-DEFENSE-LESSON-2026-09-02.md Part 4 proposes C6 — The Cost Defense as “a sixth module” of Track C, complete with a unit, an exercise and five drafted assessment items. Integration-doc R42 rules there is no sixth module. The conflict is real and is resolved below (RE31) — the lesson is good, the module number is not, and the content is investor-side anyway.

RANKED NEEDS

A STANDING NOTE ON VIDEO, BEFORE THE NEEDS

Three rules govern every VIDEO NOTE below, and they follow from the owner’s own constraint (cloned voice, timelined slideshows, an avatar for a few clips) plus the school’s content law.

1. A rule never goes in the audio. Anything the living map owns — a state’s hour requirement, a protected class list, a referral-fee rule, a licensed-activity boundary — renders as on-page text carrying its own “verified as of” date, and the narration says “the current rule for your state is on this page.” Audio is the most expensive asset to re-cut and the one nobody re-checks; a stale sentence in a cloned voice is worse than a stale sentence in markdown, because it sounds authoritative and nobody can diff it. Track A and Track B law units are therefore narration-light on purpose.
2. A build gets a screen recording, not a slide. VOICE-AND-RULES §7 forbids invented UI — button labels, menu paths. A real screen capture is the only way to show a build without inventing one, and it is also the only asset a slide pipeline genuinely cannot fake. Every Track C unit below names its recording.
3. The avatar is rationed to the honesty moments. One clip per module at most, and it is spent where the book itself slows down: the limit, the boundary, the thing that stays yours. An avatar reading a syllabus is a waste of the one asset that reads as a person.

RE01 · LM — the living state-law map

1. Where it sits. A cross-track asset, LM, plus two thin usage units that teach a student to read it: F5-U1 (Track A) and B1-U0 (Track B). Not a module — a versioned rule store the unit pages render from.

2. What promises it. The copyright page: “Real estate law, agency rules, and disclosure requirements differ by state and change over time — the companion certification school carries the current law for where you practice; this book does not.” HOW-TO-USE ¶15: “the school carries what a printed page can’t responsibly carry — the law as it currently stands in your state.” CH10 ¶70 <!--SCHOOL: B1-->: “Your state’s rule on where that line actually sits lives in the school’s living map of licensed activity, B1, updated as legislatures move.” App B ¶390, identical claim. App B ¶384, ¶386, ¶388, ¶392, ¶394 — five more “the school’s X module carries it, current.” CH10 ¶35, ¶45, ¶51. References ¶44: “the honest place to keep a rule current.” App B ¶7: the Foundations track carries “hours required, exam format, fees, deadlines.”

3. What exists today. Nothing. Design §2’s note — “A state-specific license-law supplement (Virginia salesperson statute, DPOR-specific rules) is a separate, later build — see §6” — points at a §6 that is about static HTML delivery and says nothing about it. §8 item 1 still asks the owner whether the school is Virginia-only or multi-state.

4. The gap. Twelve book sentences assert a currency mechanism that has not been specified, let alone built. Worse, the design’s shape actively fights the promise: the only state-specific content is Virginia, and it sits in the Broker track, so a reader in Ohio holding Book 3 and turning to Appendix B is pointed six times at a module that contains another state’s law and that he is not eligible to take. There is no rule schema, no per-rule effective date, no verification owner, no review cadence, no citation requirement, no mechanism to mark a rule “unverified for your state” honestly rather than defaulting to Virginia’s answer.

5. What “deeper” means. One asset spec and two short units.

• LM — the rule store (spec, not a lesson). Every rule is a record: topic (one of the school’s fixed topic ids — agency attachment, disclosure content, disclosure timing, advertising, fair-housing classes, contact and consent, licensed activity, referral fees, unlicensed-assistant scope, escrow handling, licensing hours/fees/exam, post-licensing/CE), jurisdiction (state, plus federal for the floor), plain-words statement, citation, verified_on date, verified_by, and status (verified · unverified · under legislative review). A unit page renders the rules for the topic it teaches, filtered to the state the reader selects (RE29), and prints the date. Where a state is unverified, the page says so in those words and points at that state’s commission — the school never guesses and never substitutes Virginia. Launch scope: the federal floor plus Virginia verified; every other state ships as unverified with the honest sentence, and states are promoted one at a time.
• F5-U1 — Where the Rules Actually Live (and how to read them) (18 min). Objectives: (a) name the four sources that can each answer a licensing question differently — federal law, your state’s statute and commission rules, your MLS/association, your brokerage’s own policy — and say which one wins when they disagree (the strictest applicable one, and brokerage policy over your preference); (b) read a rule card on this site: what its date means, what “unverified for your state” means, and what to do when you see it; (c) state the two questions that must be answered by your own state and never by a book: what requires a license here, and what has to be disclosed, on what form, by when. Build: pick your state, open the licensing card, and write down four things — your hours requirement, your exam’s structure, your fee schedule, and the date the card was last verified — then verify one of the four against your state commission’s own page and record whether they matched. You’ll know it worked when you can name the one place you’d go to check whether the number on our page is still true, and you have actually gone there once.
• B1-U0 — The Broker’s Version of the Same Question (15 min). Objectives: (a) identify the rules that change when you carry supervisory responsibility rather than a salesperson’s license; (b) read the map for the six topics a broker is personally accountable for — escrow handling, supervision, advertising review, disclosure timing, referral fees, unlicensed-assistant scope. Build: produce a one-page “what my state requires of me specifically” sheet from the map, dated. You’ll know it worked when your sheet has a date on it and a line for every topic where your state’s card says unverified.

6. VIDEO NOTE. Slide+narration for the four-sources concept and the “strictest applicable rule wins” idea — that part is stable and safe to voice. No narration ever reads a rule aloud; the rule cards are on-page text with dates. One screen recording: selecting a state and watching the page’s rule cards change, including one deliberately unverified state so the student sees the honest failure state before they hit it live. Avatar: yes, one clip, ~40 seconds — this is the highest-value avatar moment in the entire school. A person saying, on camera, “we will not guess your state’s rule; where it says unverified, it means we haven’t checked it yet” is the school’s whole credibility posture, and it is exactly the sentence that reads as evasive in text and as integrity on a face.

RE02 · The compliance flag as page furniture

1. Where it sits. Site-wide: every unit page footer, every track landing page, every marketing surface, and the wording of the “completion record” the addendum sells in nine units.

2. What promises it. Design §7: “Until that approval is in hand, every page of the school’s site and every piece of marketing must describe the offering as exam PREPARATION, not certification or licensing credit… it cannot claim to satisfy a statutory hour requirement or issue a credit-bearing completion certificate until DPOR (or the relevant state board) says so.” Addendum §4: “the school still describes itself as exam preparation and Track C as neither preparation nor credit.”

3. What exists today. One paragraph of intent in the design and one restatement in the addendum. No template, no wording, no check.

4. The gap. The school is named a certification school in Book 3’s own HOW-TO-USE (“a certification school built to sit alongside it”) and in the preflight. That is the exact word §7 forbids until DPOR says otherwise, and it is already printed in a finished book. Meanwhile the addendum promises a “completion record” as a paid deliverable in nine of nine new units, with no wording specified — and a completion record that looks like a certificate is precisely the artifact §7 bans. There is no per-page furniture, no lint, and no reconciliation between the book’s word and the site’s required word.

5. What “deeper” means. Not a lesson — three artifacts, before lesson one ships.

• The page footer block (one paragraph, identical on every unit page): what this is (exam preparation), what it is not (statutory credit, licensing hours, legal advice), where the actual authority sits (the state commission), and the date the page was last reviewed.
• The completion-record wording, fixed: it records that a named student completed a named unit on a date, states in its own body that it carries no licensing credit and satisfies no statutory hour requirement, and never uses the words certificate, certified, accredited, or credit. If DPOR approval later lands, the record gains a credit line; until then it does not.
• A pre-publish check over every page and every marketing string for the banned set (certification, certified, accredited, licensing credit, satisfies hours, approved provider), run the way the originality check is run in §5 — a pass, not a hope. Plus one reconciliation decision the owner owes: the book says “certification school” in print; either the school’s public name drops the word, or the §7 language is tightened to explain the mismatch on the landing page. It cannot be both silently.

6. VIDEO NOTE. None — this is furniture, not a lesson. But it constrains video: no narration anywhere may say “certified,” “accredited,” or “credit,” and the cloned-voice pipeline should carry that banned list as a hard filter, because a wrong word in audio is the most expensive kind to remove.

RE03 · Binding the item bank to the law it asserts

1. Where it sits. Design §5, production plan — a sixth step, plus a change to step 2’s tagging.

2. What promises it. Design §5 step 4: “A second pass verifies the law/fact stated is correct and current (VA statutes and DPOR rules move; an item correct in 2020 Moseley material may be stale in 2026).” App B ¶540 promises the student “the real practice exam waiting at the end of pre-licensing.” The blueprint promises ~935 original items, of which 70 are state-law items (Track A’s 30 + Track B’s 40).

3. What exists today. A five-step pipeline — draft, tag to unit, originality check, accuracy check, difficulty calibration, publish. Tagging is to the owning unit only.

4. The gap. Two failures. First, an item is tagged to a unit but not to the rule it asserts, so when the living map updates a rule, nothing in the bank identifies which of 935 items just became wrong. The design acknowledges staleness as a risk and then designs a one-time check against it — an accuracy pass at drafting time is a snapshot, not currency, and the school’s whole claim is currency. Second, Track A’s 30-item state section cannot be drafted at all: it “requires a VA salesperson-law layer” (§5’s own note) that does not exist and that §2 defers to a section that doesn’t contain it. Thirty of the four hundred forty items in Track A’s bank have no source.

5. What “deeper” means. Three changes to §5, none of them a lesson.

• Tag every item to a rule id, not only a unit id. A state-law item carries LM rule id + jurisdiction + the rule’s verified_on at drafting time. When the map updates that rule, every item bound to it flips to needs re-verification automatically and leaves the live bank until a person clears it. This is the only mechanism that makes “current” true for an exam bank rather than for a lesson page.
• Split the bank by volatility. National-concept items (contracts, agency theory, valuation approaches, math) are stable and re-verified annually. State-law items are volatile and re-verified on every map change to their rule. Publish those two review cadences, and show a bank-level “last verified” date to students, the same way the rule cards do.
• Resolve Track A’s state section before drafting starts. Either (a) the Foundations state section is built per-state off LM and ships only for verified states, with the honest message elsewhere, or (b) it ships as a Virginia salesperson section explicitly labeled as such, and the national section is what every other state’s student takes. Option (a) is what the book promised; option (b) is what the design can actually produce this quarter. Pick one in writing — the item count depends on it.

6. VIDEO NOTE. None. Worth one line in the school’s own landing narration: “every practice item carries the date the rule behind it was last checked” — a claim that is only speakable once this exists.

RE04 · Registry reconciliation — the unit counts and the code scheme

1. Where it sits. Design §1 table, §2 table, §5 blueprint, addendum §4, and the page-naming convention.

2. What promises it. Design §1: Track A = “4 modules · 10 units”; §5: “10/unit × 10 units = 100”, bank “~430”, program total “≈935 original items”. Addendum §4: “Program totals go from 12 modules / 28 units to 12 modules / 37 units.” Addendum §5.4: “Whether unit codes should continue within each module (C1-U2…) or renumber Track C’s units globally… needs revisiting before pages are generated.”

3. What exists today. §2 lists eleven Track A units — F1 (2) + F2 (3) + F3 (2) + F4 (4) — against §1’s ten. Every downstream number inherits the ten.

4. The gap. Track A’s per-unit quiz bank is 110, not 100; Track A’s total is ~440, not ~430; the licensing bank is ~945, not ~935; and the program is 12 modules / 38 units before a single addition below. The “≈935 original items” figure is the one number the design tells the owner not to promise publicly until resourcing is confirmed (§8 item 5) — it should at least be right. Separately, the page-naming scheme is undecided, and design §6’s example (F<NN>-slug.html) doesn’t accommodate a module-plus-unit code at all: F2-U3 has no F<NN> form.

5. What “deeper” means. Not a lesson — one reconciliation pass, done before any page is generated.

• Correct §1, §5 and addendum §4 to eleven Track A units and re-derive every total.
• Adopt within-module unit codes everywhere (F2-U3, B1-U7, C1-U6) — the addendum’s own choice, matching §5’s B1-U3 tagging example — and change §6’s page-naming convention to <module>-U<N>-slug.html. A global renumber breaks the tagging example already in use.
• Publish one registry table, in the pattern of the platform school’s CURRICULUM-MAP.md, that is the single source of truth for module ids, unit ids, unit titles, prerequisites and page names — because the platform school’s audit found seven code collisions from exactly this omission, and this school is about to generate three times as many pages.

6. VIDEO NOTE. None. But it gates video: unit ids appear in narration (“this is where C5-U2 picks it up”), and re-cutting audio after a renumber costs more than the reconciliation does.

RE05 · A unit-page format spec for the RE school

1. Where it sits. A new SCHOOL-RE-UNIT-SPEC, sibling to SCHOOL-MODULE-SPEC.md.

2. What promises it. SCHOOL-MODULE-SPEC v1.0 governs the platform school: nine sections in a fixed order, 900–1,500 words, “the learner finishes with something real RUNNING,” every build ending with “You’ll know it worked when…”, and §7’s rule that describes product actions “at a generic level… NO invented button labels or menu paths.” VOICE-AND-RULES §5: “We describe what capabilities make POSSIBLE and how to USE them. We NEVER describe how they are built,” plus the ban on internal names.

3. What exists today. The RE design’s §6 names a file-naming convention and nothing else. The addendum gives each unit objectives, a build, an assessment topic map, a compliance note and a free/paid line — a good five-part shape, but not the school’s nine-section format, and with no word budget, no failure-modes section and no “You’ll know it worked when.”

4. The gap. Track C’s units are entirely platform walkthroughs, and no document tells a drafter that the secrecy line applies, that invented UI is banned, or that a unit is 900–1,500 words. Left unstated, a drafting pass writing “the discovery sweep” will reach for mechanism, and the failure will be systematic rather than occasional. The addendum’s shape also drops the two sections that carry the most teaching weight in the platform school: When it goes sideways (3–5 usage-level failures with plain fixes) and Words you now own.

5. What “deeper” means. One spec, four clauses.

• The nine sections, adapted. Title + one-line promise · What you’re learning (one analogy) · Why it matters · Before you start (unit ids + what must already be running) · The build (numbered, generic UI, ends with “You’ll know it worked when…”) · When it goes sideways · Words you now own · Show your work · Check yourself (3 questions). Law units (Tracks A/B) substitute The rule, and where it lives for The build, and their “build” is a worked application or a fact-pattern, never a platform action.
• Word budget 900–1,500 per unit page. Depth is more units, never a longer unit — the design’s own §1 arithmetic depends on it.
• The secrecy clause, restated in the spec so no drafter has to go find it: usage yes, internals no; no internal or product project names; “it figured it out” is the ceiling; no invented button labels or menu paths.
• The three standing compliance lines every unit carries: the §7 exam-preparation footer (RE02), the education-not-advice line, and — for the units the addendum marks with a direct DPOR note — the routing line naming which module holds the current rule.

6. VIDEO NOTE. The spec should carry a tenth field, VIDEO PLAN, per unit: which slides come straight from the lesson text, which screen recording is required, and whether an avatar clip is claimed. Writing it at drafting time is what makes the “100% automated, looks intentionally designed” pipeline actually work — a lesson written without a video plan gets a generic slideshow bolted on afterward and looks it.

RE06 · The roadmap-labeling rule for three Track C builds

1. Where it sits. C1-U3 (the week’s call list), C2-U2 (the anniversary client-value packet), C2-U3 (the time-value dashboard view).

2. What promises it. App A-6, in the book: “This one is roadmap, not running today, and it’s labeled that way here the same way Chapter Seven labels it — plainly, in the future tense, because that’s what’s actually true.” A-3 step 5: “that piece is roadmap, not running yet, and it stays plainly labeled that way until it is.” A-14 step 5: “One category on this dashboard is roadmap, not running, and it’s labeled that way here.” Owner rule R-H: “No present-tense claim that an automation is running… unless confirmed.”

3. What exists today. C1-U3’s build instructs the student to “configure a routine that produces a ranked weekly call list” and “configure drafted personal outreach” — the first is A-6, labeled roadmap in the book. C2-U2’s build instructs the student to “set the anniversary trigger against their own records” and “build the packet assembly” — A-3 step 5, labeled roadmap. C2-U3’s build says “adds the time-value category to the dashboard built in C2-U1” — A-14 step 5, labeled roadmap. None of the three carries a roadmap label.

4. The gap. Three units would assign a student a build they cannot perform. The book was careful about this in three separate places; the school would undo that care on its first day, and it would do it in a cloned voice, which is the least correctable medium in the stack.

5. What “deeper” means. No new units — a rule and three rewrites.

• The rule: a unit may only assign a build that is available to a student today. Where a mechanism is roadmap, the unit teaches its design — what it will do, what inputs it needs, what decision it will surface — and the student’s deliverable is a written configuration and a stated readiness check, in the future tense, labeled on the page. When the mechanism ships, the unit’s build section is swapped and the page’s date changes; nothing else moves.
• C1-U3 rewrite: the drafted-touch half (A-5) stays a live build; the call-list half becomes a design exercise — write the ordering rule your list would use, name the three facts that must be attached to each name, and state the cadence you’d set and why. You’ll know it worked when your ordering rule produces a defensible top ten from your own C1-U2 queue by hand, in under fifteen minutes.
• C2-U2 rewrite: the quiet-file flag stays a live build; the anniversary packet becomes a data-readiness exercise — check whether purchase dates are even populated in your own records, and report the percentage. The addendum already half-says this (“checks honestly whether the underlying dates are even populated — the answer determines whether their build is real or a design”); make it the whole assignment until the packet ships.
• C2-U3 rewrite: teach the outcome-vs-process-health distinction and the categorization boundaries live; make the time-value view a design deliverable with an explicit blind-spot statement.

6. VIDEO NOTE. These three units are the reason the video pipeline needs a roadmap slide template — a visually distinct treatment (different frame, stated future tense) so a viewer can never mistake a designed mechanism for a running one at a glance. Avatar: one clip, shared across the three — “some of what this track teaches isn’t built yet, and we’re going to tell you which parts” is the honesty moment, and it protects the school’s credibility more than any feature demo will.

RE07 · F5 — Entering the Profession (new Track A module, 5 units)

1. Where it sits. Track A, new module F5, sitting before F1 in the recommended sequence.

2. What promises it. Five Track A markers. App B ¶31: “Nothing ahead of you in this section automates the studying, sits the exam for you, or shortens the state’s required hours by a single minute” + ¶7’s promise that the Foundations track carries “hours required, exam format, fees, deadlines.” ¶73 (choosing a brokerage): “No split, cap, or fee schedule tells you that on its own.” ¶372 (the money): the 1099-NEC, self-employment tax, quarterly estimates, E&O, the reserve. ¶269 (safety): “the profession has spent decades learning that this is a real risk rather than a paranoid one.” ¶450 (the ninety-day map).

3. What exists today. Nothing. F1–F4 are exam topics; none of the five subjects appears in any objective.

4. The gap. These are the five things a person actually does in the ninety days around getting licensed, and the book routes all five to Track A by name. A Foundations track that teaches estates and easements but cannot tell a student how many hours their state requires, what a 70/30 split with a cap means for their first check, or that they owe quarterly estimated tax, is not the track Appendix B describes.

5. What “deeper” means. Five units.

• F5-U1 — Where the Rules Actually Live (18 min) — specified in RE01.
• F5-U2 — Choosing a Brokerage: Splits, Caps, Fees, and the Agreement You Sign (22 min). Objectives: (a) read a compensation structure and compute what you actually keep on a given file under a percentage split with a cap, a flat-fee plan, and a per-file transaction fee; (b) name the recurring costs that exist regardless of production — association and MLS dues, E&O, desk or technology fees — and where to find each one in an agreement; (c) identify the non-financial terms that matter more than the split (training, supervision access, what happens to your database and your pending files if you leave); (d) state plainly that an independent-contractor agreement is a contract with financial consequences and that an attorney read is a reasonable thing to ask for. Build: take two real brokerage plans in your market, compute your net on three illustrative closings at different price points under each, and write the two questions you’d ask each managing broker that the fee schedule doesn’t answer. You’ll know it worked when the two plans produce different numbers on the same three files and you can say at what production level they cross over.
• F5-U3 — The Money: How You Get Paid, and What Isn’t Yours (25 min). Objectives: (a) trace a commission from sale price through the negotiated total, the side split, the brokerage split and cap, and any referral or transaction fee, to the number that lands in your account; (b) explain what independent-contractor status means for withholding, why self-employment tax exists, and why quarterly estimates are the most common avoidable mistake in a first year; (c) define what E&O insurance covers and, more importantly, what it does not; (d) build a reserve target from your own real monthly expenses rather than from a rule of thumb. Build: write your own twelve-month cash plan — fixed business costs, personal floor, a reserve target in months, and the number of closings at your market’s median that funds it. You’ll know it worked when you can state, in one sentence and with a number, how many months of no closings your plan survives. Carries the education-not-tax-advice line and a CPA pointer, per App B’s own wording.
• F5-U4 — Staying Safe on the Job (15 min). Objectives: (a) name the ordinary moments where the risk actually sits — first meetings at unknown addresses, vacant properties, open houses alone — and why they normalize by week three; (b) state the six habits that carry most of the protection (meet first at the office or a public place, verify identity before unlocking a door, tell one specific person where you are and when you’re back, let the client walk ahead, park at the curb, know your exit); (c) find and read your own brokerage’s written safety policy; (d) recognize that the decision to end an appointment is yours and needs no justification. Build: write your own showing protocol on one page, name the specific person who gets your itinerary, and run it once on a real or simulated showing. You’ll know it worked when somebody who isn’t you can say where you were, when you were expected back, and got the message that you were done.
• F5-U5 — The First Ninety Days, Week by Week (20 min). Objectives: (a) lay out a realistic first-ninety-day sequence — weeks one to four: mechanics, previewing property, the first awkward calls; weeks five to eight: the first real conversations and shadowing; weeks nine to twelve: a first file, a rhythm; (b) name what shadowing teaches that no page does, and how to ask for it; (c) set two honest leading indicators for a first ninety days that are not closings. Build: write your own ninety-day plan against your own calendar, with the two indicators and a weekly check. You’ll know it worked when week one has specific named actions on specific days and week eleven has a stated plan for the week it doesn’t feel like it’s working.

6. VIDEO NOTE. F5-U2 and F5-U3 are slide+narration with on-screen arithmetic — a split computation building line by line is the single most watchable thing in Track A, and it comes free from the lesson text. No screen recording anywhere in this module except one optional capture in F5-U1 (the state selector). Avatar: one clip in F5-U4, ~30 seconds, on the decision to leave a showing. That is a sentence that lands from a person and reads as boilerplate in text, and it is the one paragraph in Appendix B written with the most conviction.

RE08 · F6 — Day One in Practice (new Track A module, 3 units)

1. Where it sits. Track A, new module F6, after F5.

2. What promises it. App B ¶89 and ¶97 (two Track A markers on the database entry): “Sit down with an actual list of people in your life… resisting the urge to inflate the count” and “the queue begins at prepare, the same rung every new mechanism in this book begins at.” ¶121 (cadence and shadowing). ¶430 (“Original Scripts for the First Ten Conversations”). ¶544: “Ninety days from now, none of this will feel like vocabulary anymore.”

3. What exists today. Nothing. C1-U2 in Track C teaches working a database, but it assumes a licensee with a list and a platform account, and Track C carries no licensing content by design — a not-yet-licensed Appendix B reader is not its student.

4. The gap. The book gives a brand-new agent three concrete day-one assets — a segmented sphere list with one true note per name, ten scripts, and forty terms — and marks all three Track A. Track A has no unit for any of them. This is also the module that makes the school’s real differentiator legible early: CH19 promises the Foundations track adds “building the automated habit alongside the licensed one, from day one, instead of bolting it on five years later” — and Track A as designed contains zero automation content of any kind.

5. What “deeper” means. Three units.

• F6-U1 — Your Database on Day One (22 min). Objectives: (a) build a first contact list from real relationships and segment it three ways (personal sphere, professional contacts, past clients — the third one empty and that being correct); (b) attach one true, specific note to every name, and say why a note that could apply to anyone is worse than no note; (c) explain honestly what a scoring queue can and cannot do with a list this young — it begins at prepare, with no decision history to learn from — which is this track’s first, gentlest introduction to the ladder taught properly in C5-U3; (d) name what only you can know: which names on the list are real. Build: enter your own first list with a true note per name and label every entry with its segment. You’ll know it worked when you can open any three names at random and say what you’d actually say to that person first.
• F6-U2 — The First Thirty Days’ Cadence (18 min). Objectives: (a) describe what a professional week is actually made of before there is any business — previewing property, office hours, calls, shadowing; (b) ask for shadowing specifically and productively, and name what to watch for when you get it; (c) distinguish activity that compounds from activity that only feels productive. Build: block a real week, run it, and write down what actually happened against what you planned. You’ll know it worked when the gap between the planned week and the real one is written down rather than remembered.
• F6-U3 — The First Ten Conversations, and the Forty Words (25 min). Objectives: (a) run the ten conversations a first ninety days actually contains — the sphere announcement, the lender introduction, the “not now,” the price question you can’t answer yet — in your own words rather than a script’s; (b) recognize the two moments in those ten where a helpful answer becomes licensed advice, and stop; (c) define, plainly and out loud, the forty terms you’re expected to already know the first time someone uses one without stopping. Build: record yourself giving your version of three of the ten, and define ten glossary terms from memory before checking. You’ll know it worked when your “not now” script actually leaves the person alone afterward, and you can define contingency, proration and designated agency without looking.

6. VIDEO NOTE. F6-U1 and F6-U2 are pure slide+narration. F6-U3 is the exception in Track A and it is worth the exception: the ten conversations want recorded voice role-play, two voices, unhurried — the cloned voice can carry both sides, which is exactly the kind of asset a slideshow pipeline produces well and a human production budget never affords. The forty terms want a glossary card treatment: one term per card, spoken definition, four seconds each — automatable end to end and genuinely useful as a standalone review reel. Avatar: one clip, ~25 seconds, on the “not now” conversation and why the respect in it is in actually leaving someone alone.

RE09 · F7 — The Listing Side (new Track A module, 4 units)

1. Where it sits. Track A, new module F7.

2. What promises it. Four Track A markers. App B ¶49 (MLS, keybox, forms): “MLS rules, keybox terms, and required forms are set by your association, your brokerage, and your state — confirm all three.” ¶143 (the CMA): “What it hands you is a starting draft: the evidence, organized… What it doesn’t do… is decide the final number.” ¶167 (the listing consultation) and ¶153: “the actual craft of a listing consultation, practiced and refined — exactly the kind of thing the companion school teaches as its own subject.” ¶189 (the listing agreement).

3. What exists today. F4-U2 teaches the three appraisal approaches at exam depth. A CMA is a different artifact from an appraisal — the appendix says so explicitly, and the book’s own liability line depends on the distinction. Nothing else exists.

4. The gap. The listing half of the job — the MLS mechanics, the CMA, the consultation, the agreement — is four Track A markers and zero units. Note also a routing defect worth naming here: App B ¶153/¶167 promises the listing consultation as “its own subject” in the school, and the addendum builds that craft as C1-U5 in Track C, which a not-yet-licensed reader is not the student for. Both should exist: F7-U3 teaches the consultation as practice for a new licensee; C1-U5 teaches the same conversation against a machine-prepared packet for a licensee running the system. They are not duplicates — but as designed, only the second exists, and the marker points at the first.

5. What “deeper” means. Four units.

• F7-U1 — The MLS, the Keybox, and the Forms (20 min). Objectives: (a) name what an MLS actually is — a private cooperative with its own rules and its own penalties, separate from state law and from your brokerage’s policy; (b) list the three authorities that must all be satisfied before a listing goes live and what each one governs; (c) describe keybox access as a chain of custody with your name on it; (d) recognize the field-accuracy obligation: every number you enter is one you are responsible for. Build: obtain and read your own association’s listing-input requirements and write the five fields you would be most likely to get wrong. You’ll know it worked when you can name which of your five is governed by the MLS, which by your state, and which by your broker.
• F7-U2 — The CMA, Anatomized (28 min). Objectives: (a) select comparable sales and defend each selection — recency, proximity, similarity — and name the one comp you rejected and why; (b) adjust for size and condition transparently, showing the adjustment rather than burying it; (c) state the difference between a CMA and an appraisal in one sentence a seller understands, and never blur it; (d) explain what a machine-assembled comp set is (evidence, organized, current as of last night) and what it is not (the number). Build: build a full CMA on a real address in your market from raw comps, with every adjustment shown and one rejected comp explained. You’ll know it worked when a second person can follow your adjustments to your number without asking you a question. Carries the education-not-a-valuation line verbatim from App B.
• F7-U3 — The Listing Consultation (25 min). Objectives: (a) sequence a consultation — establish what the seller actually wants, present the evidence, arrive at price together, agree what happens next and by when; (b) present a number the seller doesn’t want to hear without either defending it aggressively or abandoning it; (c) name what may never be promised — a price, a sale, a timeline — and what may be promised, specifically, about your own work; (d) read the room after the number lands, and know when to let evidence sit. Build: run the consultation twice against the same CMA — once with a seller who agrees, once with one who is twenty percent above you — and write down which of your sentences changed. You’ll know it worked when your promise about what happens next has a date in it.
• F7-U4 — The Listing Agreement and the File It Opens (20 min). Objectives: (a) identify what a listing agreement generally establishes — parties, term, price, compensation, authority, and what happens at expiration — and where your state’s required form differs; (b) explain that the signature obligates a person, not a piece of software; (c) describe what happens the moment it executes: the file opens, the dates begin, and the promises made in the consultation become obligations. Build: read your state’s actual listing form and mark every blank that a conversation, not a system, has to fill. You’ll know it worked when you can name three terms in the form you’d want explained to you before signing if you were the seller.

6. VIDEO NOTE. F7-U2 is the module’s anchor and it is screen recording plus narration: a comp set on screen, adjustments applied one at a time, the number arriving last. That is genuinely instructional and cannot be faked with slides. F7-U1 and F7-U4 are slide+narration with document call-outs (a generic form, not any state’s actual form printed as fact — the specifics belong to the living map). F7-U3 wants recorded role-play, both sides, same as F6-U3. Avatar: one clip in F7-U3, ~35 seconds, on the moment the number disappoints a seller — the one place in this module where a face is doing work text can’t.

RE10 · F8 — The Buyer Side and the Transaction (new Track A module, 5 units)

1. Where it sits. Track A, new module F8.

2. What promises it. Six Track A markers. App B ¶221 (a buyer from first call to keys), ¶239 (the buyer consultation): “When a representation agreement has to be signed, what it must contain, and how compensation may be discussed are set by your state and your MLS — confirm all three with your broker before your first showing.” ¶288 (the paper an offer generates), ¶316 (the transaction anatomized — five deadline families), ¶334 (closing day): “What stays yours is the call you make before the money moves.” ¶352 (the two negotiations after the offer).

3. What exists today. F4-U4, “Transfer of Title & Closing Procedures,” at exam depth — deeds, conveyance, the sequence. Nothing on buyer representation, showings, offers, contingency mechanics, wire fraud, or the inspection and appraisal negotiations.

4. The gap. The buyer half of the job carries six Track A markers and no units, and one of the six — closing-day wire fraud — is the single highest-consequence practical topic in the appendix and appears nowhere in the standard exam topic universe either, so nothing else will catch it.

5. What “deeper” means. Five units.

• F8-U1 — The Buyer Consultation and the Representation Agreement (25 min). Objectives: (a) run a first buyer conversation that establishes ability, timeline and motivation before it establishes a wish list; (b) explain what a buyer representation agreement is and why the moment it must be signed is a state-and-MLS question, not a preference; (c) discuss compensation plainly and without guessing at a rule; (d) have the honest conversation when the range and the wish list don’t fit each other. Build: run a consultation against a supplied buyer profile and produce a written summary of ability, timeline, motivation and the gap. You’ll know it worked when your summary names the gap in the buyer’s own words rather than yours. Rules on timing and content route to the living map.
• F8-U2 — Showings and Open Houses (18 min). Objectives: (a) plan and run a showing sequence that ends with a decision rather than a tour; (b) run an open house as two jobs at once — selling that house, and meeting the neighbors and unrepresented buyers who walked in; (c) capture what happened while it’s still true, that evening; (d) recognize the moment a helpful answer at an open house creates an expectation, and what to do about it. Build: run or shadow one open house, capture every visitor with one real note, and follow up the same evening. You’ll know it worked when every sign-in has a note attached that isn’t the person’s name.
• F8-U3 — The Paper an Offer Generates (25 min). Objectives: (a) name what an offer package generally contains and what each piece is for — the offer, the addenda, the disclosures, the proof of funds or financing; (b) describe what constitutes acceptance and why “we agreed on the phone” is not it; (c) sequence a counteroffer and explain what happens to the original; (d) identify which terms are negotiable and which are set by the form. Build: assemble a complete offer package on a practice file, then mark every element whose content or deadline your state’s form dictates. You’ll know it worked when the package is complete without a single blank you filled by assumption.
• F8-U4 — The Transaction’s Five Deadline Families (28 min). Objectives: (a) sort every date in a contract into the five families — money, inspection, financing, title, closing — and say who has to act on each; (b) count backward from a deadline to the day the work has to start; (c) name what a missed date in each family actually costs; (d) distinguish a routine delay from the first sign of a file breaking. Build: take a real executed contract (or a supplied one), extract every date, sort it into families, and name the responsible party for each — by hand, before any system does it for you. You’ll know it worked when your timeline has a date the contract didn’t state outright but its language creates.
• F8-U5 — Closing Day, Wire Fraud, and the Two Negotiations After the Offer (25 min). Objectives: (a) sequence the final walkthrough and the closing, and say what each is actually for; (b) name wire-fraud’s exact mechanism — instructions that arrive by email and change at the last moment — and the one habit that defeats it: call a number you already had, before the money moves, every time; (c) run the inspection negotiation as a repair-or-walk decision rather than a line-item fight; (d) make the appraisal-gap call yourself, early, before the client hears it anywhere else. Build: write your own pre-wire verification protocol and the exact words of the low-appraisal call. You’ll know it worked when your protocol names the phone number’s source and does not accept one that arrived in the same email as the instructions.

6. VIDEO NOTE. F8-U4 is a diagram-led slide unit — five families on a timeline, dates populating one at a time — and it is the most reusable asset in Track A because it recurs verbatim in C2-U5. Build it once, use it twice. F8-U5’s wire-fraud segment is the strongest short-video candidate in the whole school: a sixty-second sequence showing an email arriving with changed instructions is more persuasive than any paragraph. Avatar: one clip in F8-U5, ~30 seconds, on the call before the money moves. No screen recording anywhere in this module; nothing here is a platform action.

RE11 · The tested topics Track A doesn’t cover

1. Where it sits. Four new units inside existing modules: F1-U3, F1-U4, F2-U4, F4-U5.

2. What promises it. Design §2: Track A is “outlined here from the standard topic universe every U.S. salesperson pre-licensing exam draws from — the same broad categories the Moseley Broker courses assume a student already has.” §5 allocates 80 national items “weighted across F1–F4’s 7 sub-topics” proportional to “standard licensing-exam blueprint weightings.”

3. What exists today. Eleven units covering estates and interests, land use and environment, contracts, agency, license law and ethics, disclosures, fair housing, financing, valuation, math, and transfer/closing.

4. The gap. Four categories that appear on standard salesperson content outlines have no unit, and a national section of 80 items cannot be weighted across a topic map that is missing them. Legal descriptions and surveys (metes and bounds, lot-and-block, the government survey system) — absent; F4-U3’s math covers area but not description. Co-ownership and common-interest communities (tenancy in common, joint tenancy, tenancy by the entirety, condominium/co-op/PUD, HOA governance and resale documents) — nominally inside “estates and interests,” which is a different topic and already carries freehold/leasehold, easements, liens and encroachments in one unit. Leases, landlord-tenant, and property management — leasehold is named in F1-U1’s objectives as an estate; lease types, landlord-tenant obligations, security deposits and the property-management relationship are absent, and property management is a licensed activity in most states. Property taxation and transfer taxes — absent entirely; assessment, exemptions, tax liens as a superior lien, transfer/recordation taxes, and the homeowner tax basics an agent gets asked about weekly.

5. What “deeper” means. Four units, each 18–22 min, each with a worked application rather than a build.

• F1-U3 — How Two or More People Own the Same Property (20 min). Objectives: (a) distinguish the common forms of concurrent ownership and what happens to each at death; (b) describe what a purchaser of a condominium, a co-op unit, or a PUD lot actually buys, and what the association controls; (c) identify the association documents that matter in a resale and why their delivery is often a deadline. Application: read a supplied fact pattern with three co-owners and state who may sell what, alone.
• F1-U4 — Describing Land So It Can Be Conveyed (18 min). Objectives: (a) read the three common description systems and say which is used where; (b) explain why a street address is not a legal description; (c) identify when a survey is warranted and what it settles. Application: match four descriptions to their systems and identify the one that could not convey.
• F2-U4 — Leases, Landlord-Tenant, and Property Management (22 min). Objectives: (a) distinguish the common lease types and what each shifts between the parties; (b) name the core landlord and tenant obligations and the usual security-deposit constraints; (c) describe the property-management relationship as an agency with its own duties and, in most states, its own licensing requirement. Application: given a fact pattern, identify the lease type and the party bearing a given cost.
• F4-U5 — Property Taxes, Transfer Taxes, and the Tax Side of Ownership (20 min). Objectives: (a) explain assessment, the rate, and how a tax bill is computed and prorated at closing; (b) describe why a property-tax lien generally outranks other liens; (c) name the transfer or recordation taxes that appear on a settlement statement and who customarily pays them; (d) state the three homeowner tax topics an agent is asked about and route every one of them to a CPA. Application: compute a proration and a transfer tax on a supplied settlement scenario.

6. VIDEO NOTE. All four are slide+narration straight from the lesson text, with one diagram each (the co-ownership death-transfer matrix, the three description systems, the lease-type cost matrix, the proration timeline). F1-U4’s government-survey diagram is the one animation worth building properly — it’s the classic thing a static page teaches badly. Avatar: none. These are reference units and a face adds nothing.

RE12 · F3-U3 and F3-U4 — the federal floor C5-U2 assumes

1. Where it sits. Track A, module F3, two new units.

2. What promises it. App B ¶394 <!--SCHOOL: C5-->: “Whether a given name may be called, texted, or mailed, and when, and how often, and what has to be captured before a machine-built list gets to work it, is its own set of rules — separate from fair housing, sitting right alongside it… Your state’s rule governs what that check has to confirm; the school’s compliance module carries it.” Addendum C5-U2’s DPOR note routes contact-and-consent to “B1 and B3” and the fair-housing framework to “F3.” C3-U2’s note routes a fair-housing question to “C5-U2, B1, and F3/B1’s fair-housing content.” References cite the 2024 settlement-driven practice changes.

3. What exists today. F3 has two units: mandatory disclosures, and fair housing. Nothing on the federal transaction statutes, nothing on the contact rules, nothing on antitrust.

4. The gap. Track C’s compliance units route students to F3 for law that F3 doesn’t teach. Contact-and-consent is the newest of Appendix B’s seven compliance terms and the one the book says “matters more now than it ever has, precisely because a well-built system makes it trivially easy to reach hundreds of people on a schedule” — and it is taught in no unit in any track. Antitrust is a standard tested topic, is the live legal issue of this decade in this profession, and is absent from all three tracks; a school that teaches a machine to apply one pricing or compensation rule uniformly across a roster and never mentions antitrust has a real hole.

5. What “deeper” means. Two units.

• F3-U3 — The Federal Rules That Ride Along With Every Transaction (25 min). Objectives: (a) name what RESPA governs and, specifically, what a referral fee or a thing of value may and may not be paid for in a settlement service; (b) explain the disclosure timing the federal lending rules impose and why a change to loan terms restarts a clock; (c) state what ECOA prohibits and how it interacts with fair housing; (d) name the contact rules as their own category — the do-not-call framework, the consent rules for texts and automated calls, and the commercial-email requirements — and identify what has to be captured about a contact before a machine-built list is worked. Application: take a list of six outreach scenarios and mark which are permitted, which need consent, and which need a record you don’t currently keep. You’ll know it worked when you can name the three fields a contact record must carry before an automated touch may go to it — which is exactly the three fields A-9’s check reads, and which C5-U2 assumes you already know.
• F3-U4 — Antitrust and the Compensation Conversation (22 min). Objectives: (a) name the four conduct categories antitrust law reaches in this profession — price fixing, market allocation, group boycotts, tying — and give a plain example of each as it actually shows up in an office; (b) explain why commission is negotiated between a licensee and a client, never set among competitors, and why casual conversation among agents is where the risk lives; (c) describe how buyer-side compensation is agreed today, and route the current mechanics to the living map because they have moved recently and continue to; (d) state the one habit that protects you: never discuss your fee with a competitor, in any room, ever. Application: read six overheard sentences and mark which ones you should walk away from.

6. VIDEO NOTE. F3-U3 is slide+narration with one strong diagram: the three fields a record needs before a machine touches it, drawn once and reused in C5-U5. F3-U4’s “six overheard sentences” is the best short-form video in Track A — six five-second scenes, one voice, plain text on screen. Avatar: one clip in F3-U4, ~20 seconds, on never discussing your fee with a competitor. It is a short, absolute rule and absolute rules land better from a person.

RE13 · B1-U7 — referral fees, RESPA §8, and what an unlicensed assistant may do

1. Where it sits. Track B, module B1, new seventh unit.

2. What promises it. App B ¶392 <!--SCHOOL: B1-->: “What counts as a permissible referral fee, and what an unlicensed assistant on a team may and may not do, is exactly the kind of detail your broker should walk you through on day one… Your state’s rule governs both — what a referral fee may be paid for and to whom, and what an unlicensed team member is permitted to do; the school’s licensing module carries it, current.”

3. What exists today. B1’s six units: VA licensing law, specific VA acts, escrow, VA disclosures, VA agency definitions and broker supervision, VA fair housing. None covers referral fees or unlicensed-assistant scope.

4. The gap. A named book sentence points at a named module for two specific answers the module does not contain. It is also the topic with the widest gap between what teams actually do and what the rules permit — unlicensed assistants performing licensed activity is one of the most common enforcement matters in the profession, and Track C’s C4 units are about to teach students to hand more of the file to more people.

5. What “deeper” means. One unit.

• B1-U7 — Referral Fees, Compensation, and Who May Do What (25 min). Objectives: (a) state who a referral fee may be paid to and for what, and why paying an unlicensed person for a referral of real estate business is a different act from thanking a friend; (b) explain what RESPA §8 prohibits in a settlement-service context and what an affiliated-business arrangement requires; (c) draw the line between licensed activity and administrative support, with a concrete list of tasks on each side — and name the three that surprise people; (d) describe a broker’s supervisory exposure when an unlicensed assistant crosses it. Build: take your own (or a supplied) team’s task list and sort every task into licensed, administrative, or “check the map” — then look up the three you weren’t sure about. You’ll know it worked when your “check the map” pile is empty and you can cite the rule for the two closest calls.

6. VIDEO NOTE. Slide+narration. The task-sorting exercise is a strong on-screen sequence: forty tasks flying into three columns, the ambiguous ones landing last. No screen recording. Avatar: none — this is a rule unit and the rules belong on-page with dates (RE01’s rule holds).

RE14 · B4 — Running the Office (new Track B module, 4 units)

1. Where it sits. Track B, new module B4, explicitly labeled post-exam and non-tested.

2. What promises it. CH19: “the Broker track is the upgrade path. It carries the licensing side of that step… What it adds beyond the exam is the same argument this book’s back half already made in scene: the difference between being the most experienced agent in the building and being the person responsible for the roster is a real difference, best walked into on purpose, with the infrastructure already understood, rather than discovered the week the first agent asks you why the office doesn’t have any of what her friend’s brokerage has.” App B ¶544 repeats it: “the part nobody warns you about, which is that being responsible for other people’s files is a different job than being excellent at your own.” Chapters 16–18 are written from the broker’s seat.

3. What exists today. B3-U3, “Practice of Real Estate (Brokerage Operations & Supervision)” — one unit inside a law module, with objectives limited to “a broker’s operational and supervisory responsibilities for licensees under them.”

4. The gap. The book promises the Broker track adds a whole second half beyond the exam, twice, in two different places. The design gives that second half one unit. Everything the book actually shows a broker doing — recruiting on infrastructure rather than split, running an office P&L, setting a policy manual, onboarding and offboarding agents, holding a roster to a standard — is either in one unit or in Track C, which a Broker-track student may never take.

5. What “deeper” means. Four units, clearly marked as post-licensing content carrying no exam weight (which also keeps them outside the ~945-item bank).

• B4-U1 — The Policy Manual and What It’s Actually For (22 min). Objectives: (a) name what a brokerage policy manual must settle before it settles anything else — supervision, advertising review, escrow handling, file retention, complaint handling; (b) distinguish policy that repeats law from policy that is stricter than law, and explain why the stricter one still governs your agents; (c) write a policy provision that an agent can actually follow on a Tuesday. Build: draft three provisions of your own manual and have someone who doesn’t work for you tell you what each one means. You’ll know it worked when their reading matches yours on all three.
• B4-U2 — Onboarding, Offboarding, and the Independent-Contractor Line (22 min). Objectives: (a) build an onboarding sequence that puts a new agent under supervision, insurance, and the office standard on day one; (b) name what an independent-contractor relationship permits you to require and what it does not, and why the distinction is both a tax question and a supervision question; (c) run an offboarding that protects the client, the file, and the record — and state plainly what travels with a departing agent (her contacts and client history) and what does not (the office’s machine, per A-15). Build: write your own day-one checklist and your own departure checklist, side by side. You’ll know it worked when every item on the departure list has an owner and a deadline.
• B4-U3 — The Office’s Numbers (25 min). Objectives: (a) read an office P&L and name the four costs that scale with headcount and the two that don’t; (b) model a commission plan — split, cap, fees — and compute the office’s revenue per agent at three production levels; (c) identify the production level below which an agent costs the office money, and decide what you do about it; (d) explain why recruiting on infrastructure outlasts recruiting on split. Build: model your own (or a supplied) roster of fifteen agents at real production distribution and find the crossover. You’ll know it worked when you can name the number of agents at which one more hire stops helping.
• B4-U4 — Supervision That Reads Off a Record (22 min). Objectives: (a) describe the difference between supervising by meeting and supervising by record; (b) name what a broker is personally accountable for regardless of how the office is run; (c) design a review rhythm that catches a file drifting before a client complains. Build: define the five things you would check on every file every week and where each one comes from. You’ll know it worked when all five can be answered without asking the agent.

6. VIDEO NOTE. B4-U3 is the module’s anchor and is slide+narration with live-building arithmetic — a commission plan computing across three production levels is the most watchable thing in Track B, same asset class as F5-U2. B4-U4 shares one screen recording with C4-U2 (the intake floor’s exception queue) — build it once, cut two versions. Avatar: one clip in B4-U1, ~30 seconds, on the difference between being the best agent in the building and being responsible for the roster. That sentence is the reason this module exists, and it is CH19’s own line.

RE15 · Depth passes on B1-U3, B2, and B3

1. Where it sits. Track B: B1-U3 (escrow), module B2 (three units), module B3 (four units).

2. What promises it. Design §3 sources each to a full 45-hour Moseley course, and flags Course 2 as “Best source for a finance/appraisal-basics unit.” B2’s objectives promise “broker-level calculation categories (yield, net operating income basics, amortization) at a depth beyond Track A’s math unit.” B3’s promise “draft-review-level contract issues a broker must catch that a salesperson course does not cover.”

3. What exists today. B1-U3 is one unit of objectives ending at “handling, recordkeeping, and reconciliation requirements.” B2 is three units for a 45-hour course. B3 is four units, and its objectives are analytic (“analyze a fact pattern”) with no stated exercise.

4. The gap. Three 45-hour courses compressed to thirteen units is defensible as a topic map and thin as teaching, and none of the thirteen names a single build, worked example or fact pattern. Escrow is the one topic where a broker’s personal liability is most direct and most arithmetic, and “describe the reconciliation requirements” is not the same skill as performing one. B2’s yield and NOI objectives have no worked scenario. B3’s contract unit has no document to mark up.

5. What “deeper” means. Not new modules — a build attached to five existing units, plus two additions.

• B1-U3 gains a build: reconcile a supplied three-month escrow ledger against bank statements with two planted errors — one timing, one misallocation. You’ll know it worked when you find both and can state which one would have been a violation even if nobody lost a dollar. (+10 min)
• B2-U1 gains a build: compare three financing structures on the same property and state which client circumstance selects each. B2-U2 gains a build: apply all three valuation approaches to one property and defend which governs. B2-U3 gains a build: a full settlement-statement lab — prorations, payoff, credits and debits both sides — which is the single most-tested arithmetic in the profession and currently has no home in either track.
• B2-U4 (new) — Income Property and Investment Analysis at broker level (NOI, cap rate, cash flow, debt service coverage), because the book’s own Chapter 12 puts a licensee in front of investment math and B2’s objectives name NOI without a unit that carries it.
• B3-U2 gains a fact-pattern lab: mark up a supplied (redrafted, anonymized) purchase contract and identify the five issues a salesperson course doesn’t cover. B3-U1 gains a fact-pattern set: four breach scenarios, one remedy each.

6. VIDEO NOTE. The escrow reconciliation and the settlement-statement lab are both screen-recorded spreadsheet walkthroughs — dry, and the two most-watched assets in any broker course, because students rewatch arithmetic. Everything else is slide+narration. Avatar: none in Track B except B4-U1’s single clip — Track B’s student is there for the exam, and a face between them and the material is friction, not warmth.

RE16 · C5-U3 — Reading the Ladder

1. Where it sits. Track C, module C5, new unit — and, despite its code, the first unit a Track C student takes. (See the note on R42 at the end of this entry.)

2. What promises it. Every one of Appendix A’s seventeen builds ends with a paragraph headed “Where it sits on the ladder.” App A’s preamble: “Each one closes by naming exactly where it stands on the authority ladder this book already taught you to read — prepare, propose, propose-with-track-record, authorized, and the one landing above all four that never goes away, human-on-exception.” CH10 coins the licensed ceiling. CH18 is an entire chapter about graduating a class of decisions. App B ¶97 applies it to a first-day database: “the queue begins at prepare, the same rung every new mechanism in this book begins at.”

3. What exists today. Nothing. The words prepare, propose, propose-with-track-record, authorized and human-on-exception appear in no unit’s objectives in either the design or the addendum. C5-U1’s objectives are about what never automates; C5-U2’s are about the relationship layer’s three edges.

4. The gap. The school’s most-used vocabulary is untaught. Every Track C unit’s build ends somewhere on this ladder, every addendum unit refers to approvals “graduating under a bound,” and no unit ever defines the rungs, the evidence that moves something between them, or the difference between a rung set by a record and a ceiling set by a rule. A student can complete the entire designed track without ever being taught the frame the whole book is written in.

5. What “deeper” means. One unit — short, early, and load-bearing.

• C5-U3 — Reading the Ladder (20 min). Objectives: (a) name the four rungs and the landing above them, and say what changes at each — what the machine does, what you do, and what evidence is being collected; (b) distinguish a rung that a record can move from a ceiling that only a rule can move, and give one example of each from your own work; (c) state what “graduation by class” means and why a birthday note and a listing description graduate differently even when the same routine drafts both; (d) name the two failure modes at the extremes — a queue nobody reads, and a class promoted on a feeling rather than a record. Build: take three things you already do by hand and place each on the ladder honestly, then write the specific evidence that would justify moving one of them up a rung. You’ll know it worked when at least one of your three is a ceiling rather than a rung, and you can say who sets it.

Note on R42 and sequence. R42 forbids a sixth module, and this unit is orientation, so it has no natural home in a five-module structure organized by subject. Placing it at C5 is correct by subject (authority and boundaries) and wrong by sequence (it must come first). The honest resolution inside R42 is to keep the code and state the sequence explicitly in the track’s curriculum map: Track C’s recommended order begins C5-U3 → C1-U1 → C1-U2 → …. This is worth one owner sentence rather than an invented module, and the alternative — burying the ladder inside C1-U1’s already-overloaded extension — would hide the school’s most important vocabulary inside a lead-generation lesson.

6. VIDEO NOTE. This is the single best avatar unit in the school and the only one where a face should carry most of the runtime. The ladder is a trust argument, not a procedure; it is the thing a skeptical licensee needs a person to say. Structure: ~90 seconds avatar on why authority is earned rather than configured, then slide+narration for the four rungs with one diagram (the ladder, with the ceiling drawn as a line that doesn’t move), then ~30 seconds avatar on the ceiling. One screen recording: an approval record showing proposals next to decisions, which is the artifact C5-U4 then reviews.

RE17 · C5-U4 — The Authority Review

1. Where it sits. Track C, module C5, new unit — the track’s capstone.

2. What promises it. CH18 <!--SCHOOL: C5-->, the whole chapter. App A-17: “an actual record to decide from about whether a class of decisions has earned its next rung — instead of a feeling about whether the machine seems to be doing fine lately… the one build in this appendix that isn’t really about a machine at all. It’s about you, reading your own record, and deciding what you’ve actually earned the right to stop personally checking.” App A’s closing: “A-17 is where all of it either turns into something real or doesn’t.”

3. What exists today. Nothing. C5-U1 enumerates what never graduates; nothing teaches the review that decides what does.

4. The gap. The book’s final mechanism, the one that makes every other build’s ladder position meaningful, has no unit. Without it, a Track C graduate has a stack of configured routines and no method for ever moving any of them — which means the school teaches the setup and skips the entire point.

5. What “deeper” means. One unit, and it is the natural last unit of the track.

• C5-U4 — The Authority Review (35 min). Objectives: (a) pull the record for one class of decisions and read the two columns — what was proposed, what you actually decided — including the reason logged at each divergence; (b) judge convergence off the log rather than off a feeling, and name the sample size below which you refuse to decide; (c) write a bound in words that would still be clear to someone else in six months: the ceiling, the qualifying file types, the conditions; (d) set the governor — what stops the line regardless of record, including everything sitting under the licensed ceiling — and state how you would narrow or revoke a bound as readily as you granted it. Build: run a real review on one class from your own configuration (or the supplied practice record), write the bound or the refusal in writing, and diarize the second look for six months out. You’ll know it worked when your written bound survives being read aloud to someone who wasn’t in the room, and your calendar has the re-review on it.

6. VIDEO NOTE. One long screen recording — the review record open, two columns, divergences read one at a time — because this is the only unit whose subject is literally reading a screen. Slide+narration for the bound-writing structure. Avatar: one clip, ~40 seconds, on revoking: “a bound worth having is one you’re as willing to pull in as push out.” That is App A’s own sentence and it is the note the track should end on.

RE18 · C5-U5 — Building the Check Before Send

1. Where it sits. Track C, module C5, new unit.

2. What promises it. CH10 ¶64 <!--SCHOOL: C5-->: “Your state’s rule on what that check has to confirm before a name gets touched lives in that same compliance module, C5.” App B ¶386: “This is where the compliance check before send earns its name — the device this book builds around exactly this problem.” App B ¶394 on contact and consent: “the same compliance check before send that stands in front of your advertising, pointed at a different question.” A-9 in full. CH17 ¶71 and CH12 ¶45 both invoke it.

3. What exists today. C5-U1 (what never automates, extended to add contact permission) and C5-U2 (the relationship layer’s three edges, whose build is a one-page boundary map — a document, not a device). Neither stands the check up.

4. The gap. The book’s most-named device is described everywhere and built nowhere. A-9 gives five concrete steps — three buckets, disclosure routing by stage, advertising language screening before a draft reaches you, three fields read on every contact before a message leaves the queue, and a clear rule about what clears automatically and what never does — and no unit walks any of them. C5-U2’s boundary map is a good deliverable and it is an audit of other units’ work, not a build.

5. What “deeper” means. One unit, sequenced after C1-U3 and C2-U2 so there are real outbound routines to put behind it.

• C5-U5 — Building the Check Before Send (35 min). Objectives: (a) name the three buckets the check holds separately — disclosure, agency, advertising (with fair-housing language and contact-and-consent riding inside the third) — and say why folding them into one “compliance” step loses the ability to say what failed; (b) route a stage-attached disclosure through the check rather than leaving it as an ordinary to-do, so it drafts, cites, dates, and waits on a name; (c) configure the three fields the check reads on every contact before an outbound message leaves a queue — where the record came from, what that person was told at the time, and any channel they’ve since asked to be left off — and demonstrate a hold; (d) state precisely what clears automatically when a rule stops applying to a file, and what never clears on any record. Build: put every outbound routine you have built so far behind the check, then deliberately break it — send one message from a record missing a consent field and one draft containing a phrase describing a buyer rather than a property — and confirm both are held before they reach a person. You’ll know it worked when you have two holds you caused on purpose, and you can say in one sentence which bucket caught each.

Relationship to C5-U2. C5-U2 stays as written and becomes the audit that follows this build: U5 stands up the check, U2 maps the boundary across the student’s whole configuration. Cross-reference both ways.

6. VIDEO NOTE. Screen recording is mandatory and is the unit — a message being held, the reason shown, the fix applied, the message clearing. That thirty-second loop is the most persuasive artifact the entire school can produce, because it turns a compliance argument into something visible. Slide+narration for the three buckets. Avatar: one clip, ~30 seconds, on the affirmative case — a rule applied by a machine is applied identically to the fortieth message and the first — which is owner rule R-F’s whole posture and should be said by a person, not a slide.

RE19 · C1-U6 — The Watch Layer

1. Where it sits. Track C, module C1, new unit, between C1-U2 and C1-U3.

2. What promises it. CH06 <!--SCHOOL: C1-->, the book’s strongest scene. A-4 in full: the nurture rhythm, the specific signal set (“a second visit to the same listing in one evening, a reply that asks a real question instead of returning silence or an emoji, a click on a link the name had ignored twice before, a saved search that suddenly widens”), the same-evening package with the whole thread attached, and the three-stage graduation of the flag itself. CH19 names “the hand-off that arrives with context attached” as part of module one.

3. What exists today. C1-U1 says “nurture” in one objective. C1-U2 re-scores a list nightly and surfaces stale records. C1-U3 builds a call list and drafted touches. None of the three configures a signal set, and none produces a same-evening flag with a thread attached.

4. The gap. The re-scoring queue (A-3) and the watch layer (A-4) are two different devices doing two different jobs — one ranks a list on a schedule, the other reacts to a single event within hours — and the addendum’s C1-U2 covers only the first. The book’s central scene, the one it says earns the whole argument, has no build. There is also an unbuilt half of A-4 nothing covers: the graceful path where a name that never moves produces no flag at all, and the A-8 route that puts a fallen-through file’s contact back into this layer.

5. What “deeper” means. One unit.

• C1-U6 — The Watch Layer: When a Name Actually Moves (30 min). Objectives: (a) distinguish a scheduled re-score from an event watch, and say which failure each one prevents; (b) define your own signal set — what, in your market and on your list, actually counts as “something changed” — and defend each signal against the false positive it invites; (c) configure the same-evening flag so it arrives with the whole thread attached rather than a score, and explain why a score is a guess dressed as math while a fact is actionable in one sentence; (d) state what the layer does when nothing moves, and why silence is the correct output rather than manufactured urgency. Build: turn the nurture rhythm on for one segment, define four signals, and run it for a week. Record every flag, mark each as real or noise, and adjust one signal. You’ll know it worked when you have at least one week where the layer said nothing at all, and you are comfortable with that.

6. VIDEO NOTE. Screen recording: a flag arriving with a thread attached, shown once. Slide+narration for the signal-set design, using the book’s own four examples as the worked case. Avatar: none — CH06’s emotional weight belongs to the story, and the story is better told in narration over the timeline of that evening than by a person recounting it.

RE20 · C1-U7 — One Listing Marketing Set

1. Where it sits. Track C, module C1, new unit.

2. What promises it. CH11 <!--SCHOOL: C1-->, an entire chapter. A-10 in full: one intake, five outputs (MLS description, single-property page, platform-sized social posts, a flyer, mailer copy), the one-wrong-number-fixed-everywhere property, and the same production line pointed at a just-sold mailer or a past-client packet. CH19 lists “Chapters Four through Seven and Eleven” as module C1’s content — Chapter Eleven is in the promise and absent from the module.

3. What exists today. Nothing. Track C has no unit touching listing marketing, and C2-U2’s client packet build is a different artifact for a different audience.

4. The gap. A whole chapter and a whole build with no unit — and it is the build with the most obvious daily payoff for the school’s core student, which makes it the most commercially costly omission in Track C. It is also the unit where the compliance argument lands hardest: every word in a marketing set is advertising, inherits fair-housing rules, and never graduates past a licensed look.

5. What “deeper” means. One unit.

• C1-U7 — One Intake, One Marketing Set (35 min). Objectives: (a) build the single intake — photos in one batch, facts confirmed once, comps already attached from the discovery sweep — and name what happens to the whole set when one fact is wrong; (b) read a drafted set against the actual house rather than against the draft, and say what only standing in the house tells you; (c) identify which parts of the set may graduate (formatting, resizing, scheduling an already-approved slot) and which never do (the words), and tie the second half to the check in C5-U5; (d) point the same production line at a different audience — a just-sold mailer, a past-client packet — and say what changes and what doesn’t; (e) name when to stop producing, because five easy variations don’t all deserve to exist. Build: run one real or supplied listing through a single intake, review the full set against the property, and deliberately plant one wrong number to confirm it corrects everywhere at once. You’ll know it worked when correcting the square footage once changes it in all five outputs and you can name the two lines in the set you would still write by hand.

6. VIDEO NOTE. This is the school’s best screen recording and its best marketing asset: one intake, then five finished outputs appearing. Ninety seconds, no narration needed over the reveal. Slide+narration for the graduation rule and the fair-housing boundary. Avatar: one clip, ~25 seconds, on knowing when to stop producing — the honest counterweight that keeps this unit from reading as a product demo.

RE21 · C1-U8 — The Dual Pipeline

1. Where it sits. Track C, module C1, new unit.

2. What promises it. CH12 <!--SCHOOL: C5--> (the compliance half) and A-11 in full: one overnight sweep, one comps pass, the routing that splits a candidate into a deal queue with a defensible offer number or a lead queue with a listing-pitch price, the both-queues case shown side by side, and a drafted opening message in each queue’s own voice. CH12’s own line: “every one of those messages, deal-queue or lead-queue, routes through the same compliance boundary before it ever reaches your outbox.”

3. What exists today. C5 carries the compliance half — disclosure, agency, advertising for a dual-hat licensee — and nothing builds the pipeline.

4. The gap. The routing itself is untaught: what criteria split a candidate, what a licensed person does differently from an unlicensed investor working the same feed, and what the both-queues case actually asks a person to decide. Note the boundary App A draws and the school must respect: the underwriting that produces the cash number “is the investor-side machine, built and documented in full in this series’ companion volume” — this unit teaches the licensed side (routing, the listing-conversation price, disclosure), not the offer math.

5. What “deeper” means. One unit.

• C1-U8 — Two Queues, One Sweep (30 min). Objectives: (a) configure one overnight scan that feeds both queues from the same raw feed and the same comps pass, and say why checking twice is worse than checking once; (b) write the routing criteria that split a candidate, and state what happens to the ones that qualify for both; (c) name what changes when the person working the feed holds a license — every message goes out under a name and a license both — and route the disclosure question to C5 and the current rule to B1/B3; (d) draft the two opening messages in two honestly different voices, and say what each one may not claim. Build: run a week of your own market’s FSBO and expired feed, route every candidate, and hand-check ten routings against your own judgment. You’ll know it worked when you can name the three routings you’d have made differently and say what rule would have caught them.

6. VIDEO NOTE. Screen recording of the split: one feed in, two queues out, one property landing in both with the two computations side by side. Slide+narration for the routing criteria. Avatar: one clip, ~30 seconds, on the dual-hat disclosure — a licensee buying from a seller they could instead have listed is the sharpest ethical moment in the book and deserves a face saying it plainly.

RE22 · C2-U5 — The Self-Building File

1. Where it sits. Track C, module C2, new unit.

2. What promises it. CH08, an entire chapter, and CH03’s one-shape claim underneath it. A-7 in full: deadlines computed off the contract’s own language and cited to the clause that created them, sorted into five families, routed to the party who has to act, counted backward from the day the deadline is created, with anything unfamiliar kicked back rather than guessed. CH13 and A-12 depend on it entirely. App B ¶316 promises the same five families to a Track A reader.

3. What exists today. C2-U4 builds the stage tracker — half of A-7 step 5 — and nothing else. Deadline computation, clause citation, the five families and the routing are taught in no unit in any track.

4. The gap. This is the most-used device in the book and it has no unit. Chapter Nine’s digest reads off it, Chapter Thirteen’s TC argument is entirely about it, Chapter Ten’s disclosure timing rides on it, and Appendix B’s transaction anatomy is its Track A shadow. Without this unit, C2-U4’s stage tracker is a checklist with nothing computing underneath it.

5. What “deeper” means. One unit.

• C2-U5 — The File That Builds Itself (35 min). Objectives: (a) explain what it means for a deadline to be computed from a contract’s own language and cited back to the clause that created it, and why a citation is what makes a computed date checkable rather than trusted; (b) sort every date into its family — money, inspection, financing, title, closing — and route each family’s reminders to the party who actually has to act; (c) count backward from a deadline on the day it is created, so a lender hears from the file with lead time rather than a day late; (d) distinguish what drafts and holds for a look from what carries a signature or legal weight and never graduates on any record; (e) name what the file does with what it cannot parse — a margin note, an unfamiliar clause, an unresponsive party — and why kicking back beats guessing. Build: run one executed contract (real or supplied) through the build, then compare every computed date against your own hand-count from F8-U4’s method, and find the one the file cited that you missed. You’ll know it worked when every date on your timeline names the clause that created it, and you have found at least one disagreement between the file and your own count and resolved which was right.

6. VIDEO NOTE. Screen recording: a contract going in, a cited timeline coming out, one date traced back to its clause. Reuse the five-families diagram built for F8-U4 — same asset, second use. Avatar: one clip, ~25 seconds, on the honest limit: a wrong date computed with total confidence is worse than no date at all. That is A-7’s own line and it is the sentence that earns the reader’s trust in the whole device.

RE23 · C2-U6 — The Four-Line Morning Digest

1. Where it sits. Track C, module C2, new unit.

2. What promises it. CH09 <!--SCHOOL: C2-->, an entire chapter, and the book’s title moment for a solo agent. A-8 in full: one line per file showing only what’s due, a calm line for a file with nothing due, a separate quiet-file signal with the last specific thing the client actually said attached, a quiet window set once and differing by file type (“a buyer between showings can reasonably go two weeks… a listing three weeks into a marketing period with no showing feedback logged cannot… a file inside a contingency window is loud by definition”), and the route that sends a fallen-through file’s contact back into the watch layer.

3. What exists today. C2-U2 sets a quiet-file window as one objective among five, aimed at post-close relationship risk. Nothing builds the digest, and nothing teaches the by-file-type window differentiation, which is the whole craft of it.

4. The gap. The chapter that most readers will remember has no build. And the design’s version of the quiet-file flag is materially weaker than the book’s: C2-U2 asks the student to “set the quiet-file window and defend the number,” singular, where A-8 is explicit that one window is a design error — “A flag that fires on the wrong window teaches you to ignore it, which is worse than not having one.”

5. What “deeper” means. One unit.

• C2-U6 — Four Files, One Calm Morning (30 min). Objectives: (a) build one daily view from the stage trackers and deadline flags already running, showing only what is actually due today; (b) explain why a file with nothing due must read as a calm line rather than a reassurance, and why removing the confirming-nothing’s-wrong tax is the actual product; (c) set quiet windows by file type and defend each number in both directions — what a too-short window costs, and what a too-long one costs; (d) distinguish a deadline risk from a relationship risk on the same view and act differently on each; (e) route a fallen-through file’s contact back into the watch layer so the relationship survives the transaction. Build: stand up the digest across at least three files, set three different quiet windows, and read one real week. Mark every line as needed-me, didn’t-need-me, or should-have-fired-and-didn’t. You’ll know it worked when at least one morning’s digest requires nothing from you at all and you trust it enough not to go look.

6. VIDEO NOTE. Screen recording, and it should be the school’s shortest and most-shared clip: four lines, three calm, one needing thirty seconds. Slide+narration for the window-setting logic with a small diagram (three file types, three windows, what each protects against). Avatar: none — the digest sells itself on screen and a person talking over it slows the one asset whose whole point is speed.

RE24 · C2-U7 — The KPI Queue at Roster Scale

1. Where it sits. Track C, module C2, new unit.

2. What promises it. CH15 <!--SCHOOL: C2-->, an entire chapter. A-14: source performance watched continuously “across which agents, over enough weeks to mean something,” conversion followed by stage rather than totaled monthly, files at risk across a whole roster, every line arriving with its reasoning and a small-sample line saying so plainly. CH15’s own closing: “What never graduates is the read on what a files-at-risk flag or a conversion dip actually means for the person behind it.”

3. What exists today. C2-U1, written for one licensee: “Identify the handful of numbers that actually predict production; build a dashboard that calculates conversion funnels automatically; use a dashboard’s proposal queue to catch at-risk files/leads.” Every objective is singular and self-directed.

4. The gap. The team-scale version is a different unit with different hazards: comparing agents rather than weeks, sample sizes small enough to lie, a proposal that reallocates another person’s leads, and the boundary CH15 spends its last page on — a number is a fact and its meaning is a conversation. C2-U3 (time-value) already exists for the process-health half; nothing carries the roster half, and C4’s units will hand students a roster before C2 teaches them how to read one.

5. What “deeper” means. One unit.

• C2-U7 — The Numbers a Team Can’t Argue With (32 min). Objectives: (a) follow conversion by stage rather than by month, so a dip surfaces near where it happened rather than six weeks later as an unexplained bad month; (b) compare source performance across agents with the sample size and territory shown, and refuse to act on a line that says small-sample; (c) run files-at-risk across a roster, and set the flags that matter at team scale — a financing deadline closing with no lender confirmation, a listing with no showing feedback logged in weeks; (d) approve what matches your read and ask a question rather than render a verdict on what doesn’t; (e) state plainly what the queue’s job ends at, and why what a number means for the person behind it was never on the ladder. Build: run one month of real or supplied roster data, work the queue, and for every line you didn’t approve, log the reason — then check next month whether the queue’s next proposal accounted for it. You’ll know it worked when you have at least one line you overrode with a reason, and the following month’s proposal reflects it.

6. VIDEO NOTE. Screen recording of a proposal with its reasoning and sample size attached, including one deliberately small-sample line so students see the honest hedge. Slide+narration for conversion-by-stage. Avatar: one clip, ~35 seconds, on the last objective — a bad month that has nothing to do with skill. This is the one place in Track C where the material is about a person’s dignity and it should not be narrated by a slide.

RE25 · C3-U3 — The Rented-Edge Audit

1. Where it sits. Track C, module C3, new unit — and the natural first action in the track, taken right after C5-U3.

2. What promises it. CH02 carries two markers, C1 and C3. A-1 in full: list every recurring charge, write what each was sold as in the pitch’s own sentence, write what you still do by hand around it, ask what your own machine would do with the same underlying data, and total the number before acting on it. The preflight names it credibility build #2.

3. What exists today. Nothing. C3-U1 is revenue-gated budgeting — the rule that gates future spend — and C3-U2 applies the same gating to hours. Neither looks at the spend already running.

4. The gap. The book’s Exhibit A has no unit, and it is the cheapest, highest-conviction exercise in the entire school: it requires no platform account, produces a number in one sitting, and is the exact thing that makes a skeptical licensee want the rest of the track. It is also the honest prerequisite for C3-U1 — you cannot gate spend against revenue until you know what the spend is.

5. What “deeper” means. One unit, explicitly a worksheet rather than a build (the design’s own precedent: A-1 “isn’t a machine at all — it’s a worksheet”).

• C3-U3 — The Rented-Edge Audit (25 min). Objectives: (a) itemize every recurring charge in your business and write, next to each, the single sentence it was sold to you as — the pitch, not the feature list; (b) write what you still do by hand around each tool to make it actually work, which is the column nobody keeps; (c) ask of each line what your own process would do with the same underlying data if you owned it rather than rented access to it; (d) total the monthly and annual number and sit with it before deciding anything — and name the lines that pass this test and stay, because the audit is not a mandate to cancel. Build: complete the audit on your own real statement, then rank the lines by what you’d keep if you had to cut a third. You’ll know it worked when the annual total surprises you, and at least one line on your list is one you’re keeping on purpose rather than out of habit.

6. VIDEO NOTE. Slide+narration, with the audit table filling in row by row on screen from a de-identified example — the owner’s own itemized stack is [OWNER-STORY-1] in the preflight and is Lane 1 material for the book, not automatically for the school; use an illustrative stack unless the owner clears his. Avatar: one clip, ~30 seconds, on the sentence that carries the chapter: you weren’t buying the edge, you were renting it. Highest-conviction line in Book 3’s front half, and it is the school’s best acquisition asset.

RE26 · C4-U2 — The Team Intake Standard

1. Where it sits. Track C, module C4, new unit.

2. What promises it. CH14 ¶79, the most explicit hand-off in the entire book: “The screens behind all of this — the actual routing rules, the floor’s exact fields, the exception paths and how you configure who they come back to — belong to the school’s team module, walked step by step; this book stops at the argument.” A-13 in full: five floor items (a recorded source in plain words, a verified contact method, one true note from the first conversation, a first touch logged inside the window the team agreed to, and a next step with a date on it), continuous checking against the same floor for everyone, automatic flag-back to a shared queue, and a narrow exception class routing to the team lead until an agent’s own record earns it.

3. What exists today. C4-U1, whose objectives are candidate sourcing, onboarding, and an accountability workflow that “reads off the record instead of off a meeting.” No intake floor, no fields, no exception path, no routing.

4. The gap. The book names a step-by-step and assigns it to this module by name, and the module does not contain it. This is the single clearest unmet pointer in Book 3 — clearer even than the missing law map, because the sentence says “walked step by step” and there are no steps.

5. What “deeper” means. One unit.

• C4-U2 — The Floor Under Every File (32 min). Objectives: (a) set a floor short enough that everyone on the roster can recite it, and defend each of your five items against the failure it prevents; (b) apply it identically to the newest hire and the fifteen-year veteran, and explain why it was never reading tenure; (c) configure the flag-back: a file that misses the floor returns to a shared queue visible on the exception path, not to one agent’s private list; (d) define a narrow, named exception class that routes to the team lead, and the evidence that would move it under an agent’s own name; (e) name what the floor is not — a judgment on anyone’s discipline — and say why it stays in place forever. Build: write your own five-item floor, configure the check across at least three files, and force two failures: a first touch never logged, and a scheduled follow-up that silently didn’t happen. You’ll know it worked when both failures reach the shared queue without anyone reporting them, and you can state which of your five items is the one people will argue about.

6. VIDEO NOTE. Screen recording of a file failing the floor and appearing on the exception queue — shared with B4-U4, cut twice. Slide+narration for the five-item design. Avatar: one clip, ~30 seconds, on the last objective: the floor exists so a client’s experience doesn’t depend on which agent answered the phone. That is A-13’s own framing and it is what keeps this unit from reading as surveillance.

RE27 · C4-U3 — The TC Handoff

1. Where it sits. Track C, module C4, new unit.

2. What promises it. CH13 <!--SCHOOL: C4--> (plus a second C3 marker for the budget half). A-12: look at what a coordinator’s day is actually made of, watch A-7 and A-8 already doing nearly all of it, read your own files as A-8 presents them, and find the genuinely human residue — “reading whether a lender’s silence means nothing or means a file about to fall apart, and the relationship call that only a person makes.” CH13’s open questions about whether you’re contractually required to route files through a broker’s in-house coordination service.

3. What exists today. Nothing. C4-U1 covers hiring generally; C3-U1 covers budget generally. Neither addresses the argument the chapter is entirely about.

4. The gap. “Automate before you hire” is one of the book’s two commercial arguments (the other is the broker rollout) and it has no unit. It is also the unit that makes the file devices pay off: a student who has built C2-U5 and C2-U6 and then does this exercise sees their own return, in a job description they no longer need to write.

5. What “deeper” means. One unit — deliberately an analysis rather than a new configuration, exactly as A-12 is.

• C4-U3 — The Hire You Don’t Have to Make Yet (28 min). Objectives: (a) write out what a coordinator’s day is actually made of, task by task, before looking at any of your own automation; (b) mark each task against what your self-building file and digest already do, honestly, including the ones they do badly; (c) name the residue — the judgment calls and the relationship calls — and say what a hire would actually be for; (d) state what does not move for a hire any more than it moves for a machine: the licensed ceiling, drafted and waiting on whoever’s name belongs on it; (e) answer the two questions the book poses rather than answers — whether you are contractually required to route files through an in-house service, and what that costs you — as a question for your own agreement and your own broker. Build: write the job description you would have written six months ago, then strike every line your build already covers, and cost what’s left. You’ll know it worked when the struck lines outnumber the survivors and you can say what the survivors are actually worth per file.

6. VIDEO NOTE. Slide+narration, with the job description striking itself line by line on screen — a genuinely satisfying automated animation and one of the few places where the slideshow pipeline beats a screen recording. One short screen recording reusing C2-U6’s digest. Avatar: none. The chapter’s own honesty (this doesn’t tell you whether to hire) is better carried in narration; save the module’s avatar for C4-U2.

RE28 · C4-U4 — The Broker Rollout

1. Where it sits. Track C, module C4, new unit.

2. What promises it. CH16 <!--SCHOOL: C4-->, an entire chapter. A-15: enroll a roster under one membership rather than individual seats, every agent waking up already inside the sweep and the watch discipline and the file discipline, an admin view showing adoption across the roster, and nothing an agent already likes being taken away. CH19’s group-membership pitch and its warning about “thirty separate agents each individually deciding whether to bother enrolling, each individually falling off the track the first busy month.” A-15’s ownership line: an agent’s contacts are hers and exportable; the machine on top is the office’s.

3. What exists today. Nothing. C4-U1 is hiring workflow at team scale.

4. The gap. The broker-scale offer — the school’s largest commercial unit of sale, per CH19’s two-tier close — has no unit. It is also where the school’s own delivery problem shows up: a broker enrolling a roster is precisely the case §6’s localStorage-only, no-account model cannot serve, which is flagged in RE29.

5. What “deeper” means. One unit.

• C4-U4 — One Flip, Not Thirty Logins (30 min). Objectives: (a) describe what changes when the machine is enrolled at the office rather than adopted agent by agent, and why individual adoption reliably fails on the first busy month; (b) set the office-wide defaults every new agent inherits on day one — the sweep for the office’s territory, the nurture discipline on her own book, the file discipline on every executed contract; (c) read an adoption view across a roster and name the three things it shows you that thirty separate logins never could; (d) state the ownership line precisely — what is the agent’s and travels with her, and what is the office’s and doesn’t — and say it out loud to a recruit without hedging; (e) route the volume of exceptions this size creates, including the licensed-ceiling items that stop for a person regardless of whose file they’re in. Build: write your own office rollout plan: the defaults, the exception routing, the ownership statement in the words you would actually use with an agent, and the first thing you would check thirty days in. You’ll know it worked when your ownership statement survives being read by an agent who is considering leaving.

6. VIDEO NOTE. Screen recording of the adoption view across a roster — the single most useful thing a prospective broker customer can be shown, and the closest thing this school has to a sales asset. Slide+narration for the defaults and the exception routing. Avatar: one clip, ~35 seconds, on the ownership line — a broker saying plainly what travels with a departing agent is a trust moment, and it is exactly the sentence a recruit will repeat back.

RE29 · Design §6 — what delivery must add before lesson one

1. Where it sits. Design §6 (delivery), plus the addendum’s §4 delivery note.

2. What promises it. CH19: “What they do is what a book never can: stay current, and let you actually build it, in your own account, against your own real database, with somebody checking your work instead of you guessing whether you configured it the way the chapter meant.” Addendum §4: “the two craft units (C1-U4, C1-U5) are the first units in the school whose paid tier involves reviewing something a student submits… which the current static-site model does not support.” Design §6’s own flagged list (cross-device progress, DPOR hour logging, proctoring).

3. What exists today. Static chaptered HTML, one page per unit; practice exams as inline JSON scored client-side; progress in localStorage.

4. The gap. Three of §6’s prerequisites are treated as future enhancements and are actually blockers. (a) A state selector. RE01’s rule cards are worthless without a way for a reader to say which state they’re in and have every rule card on every page resolve to it — and that is a page-architecture decision, not a content one, so it must land before pages are generated rather than after. (b) A submitted-work path. Two units’ paid tier is defined as reviewing something a student submits, and CH19 sells “somebody checking your work” as the school’s differentiator against the book itself. Static pages cannot accept a submission. (c) A per-page version stamp. Every rule card carries a verified_on, so every page needs a visible last-reviewed date and a change history, or the currency claim is unverifiable by the reader it was made to. Additionally, C4-U4 teaches a broker to enroll a roster, which implies accounts the current model does not have — and that collides with §6’s own “cross-device progress needs an account layer” flag.

5. What “deeper” means. Not lessons — four delivery decisions, in this order.

• The state selector and the rule-card component. One choice, remembered per browser, resolving every rule card on every page. Ships with the federal floor plus Virginia verified and every other state honestly labeled.
• A submission path for the two craft units, which can be as small as a form that sends a written call record or stage list to a reviewer and returns comments. It does not need an LMS; it needs an inbox with a defined turnaround and a stated limit.
• A last-reviewed date on every page, plus a visible change note when a rule card underneath it moves. This is what makes the school’s central claim checkable rather than asserted.
• An explicit v1 boundary, published: what the static model does and doesn’t do, so no marketing sentence outruns it — no cross-device progress, no seat-time record, no proctoring, and therefore no DPOR-hour claim (which §7 already forbids anyway). Then a decision on whether the broker group tier CH19 sells requires an account layer in v1 or ships as a manual enrollment.

6. VIDEO NOTE. None — this is infrastructure. It does constrain video in one way worth stating: because rules never go in the audio (standing rule 1), the video for a law unit is deliberately shorter than its page, and the page must be built so a viewer who watched the video still has a reason to read it. Design the page that way rather than treating video as the primary and text as the transcript.

RE30 · The Track C assessment basis and the free/paid split

1. Where it sits. Design §5’s Track C row; addendum §4 and §5.1–§5.2.

2. What promises it. Design §5: Track C is “Scenario assessments only, for course-completion, not licensure… 10/module × 5 = 50.” Addendum §4: “Proposed: move Track C’s completion assessments to 10 per unit × 14 units = 140 items… This is a proposal, not a decision.” Addendum §5.1: “The free/paid split… Inferred, not read off a decision. If one exists elsewhere, it wins.”

3. What exists today. Two open questions, both flagged honestly, neither answered.

4. The gap. With the units proposed above, Track C is 24 units, so the per-unit basis is 240 items rather than 140 — and the gap between the two proposals widens from 90 items to 190, which is a resourcing decision, not a formatting one. The free/paid line matters more than the addendum treats it: C5-U2’s lesson is free in full “deliberately — a boundary a student can’t see because it’s behind a paywall is a boundary that doesn’t do its job.” That reasoning applies with equal force to C5-U3 (the ladder, which every other unit’s build depends on), C5-U5 (the check itself), F5-U4 (safety), and F8-U5 (wire fraud). A school that paywalls the wire-fraud lesson and gives away the marketing-set lesson has its ethics backwards.

5. What “deeper” means. Two decisions, written down.

• Assessment basis: per-unit at 10 items × 24 Track C units = 240, in a separate pool, still entirely outside the licensing bank. The alternative (module-level at 50) leaves nineteen of twenty-four units with no quiz, which contradicts the design’s own §5 tagging model where every item belongs to a unit.
• A free-tier rule rather than a free-tier list: every unit whose subject is a boundary, a safety habit, or a fraud is free in full, permanently, regardless of track — C5-U2, C5-U3, C5-U5, F3-U3, F5-U4, F8-U5 — and the paid tier is the guided build, the review of submitted work, and the completion record. That is a rule a drafter can apply without asking, and it is defensible in public.

6. VIDEO NOTE. The free-tier rule doubles as the video-publishing rule: the free units’ videos are the school’s public acquisition reel, so they should be produced first and to the highest standard. The wire-fraud clip and the ladder clip are the two most shareable things this school will ever make.

RE31 · C6 — the cost-defense unit, and where it actually belongs

1. Where it sits. Currently proposed as C6 in COST-DEFENSE-LESSON-2026-09-02.md Part 4. Recommended home: the REI school, with an optional licensed-side variant as C1-U9.

2. What promises it. The cost-defense document itself: “Fits the Track C module table in SCHOOL-RE-CURRICULUM-DESIGN-2026-09-01.md §4 as a sixth module, following the same columns.” Against it: addendum §1, “No sixth module (integration doc R42). Track C stays five modules.”

3. What exists today. A complete, well-made unit spec — title, four objectives, a five-part lesson outline, a two-frame numeric exercise with a “you’ll know it worked when,” and five drafted assessment items with rationales. It is more finished than anything else in this corpus, and it is homeless.

4. The gap. Two problems, only one of which is the module number. First, R42 is binding and C6 violates it. Second, and more substantive: the content is investor-side. Its subject is presenting a cash-offer stack — after-repair value, repair bid, project-management cushion, holding costs, contingency, margin — to a seller you intend to buy from. That is Book 2’s audience. Book 3’s Track C student is a licensed agent, and the one place Book 3 touches this material (CH12/A-11) explicitly routes the underwriting to Automating REI: “that underwriting is the investor-side machine, built and documented in full in this series’ companion volume.” Dropping the investor’s margin stack into the licensed-agent capstone track imports the exact content Book 3 drew a line around.

5. What “deeper” means. A ruling, not a build.

• Primary recommendation: the unit ships in the REI school, where its audience and its source material both live, with its exercise and five items intact. Nothing is wasted.
• If it must appear in this school, it becomes C1-U9 — Explaining a Number to a Seller inside the dual-pipeline sequence, rewritten for the licensed side: the same discipline (agree on the value first, walk the costs in the same order every time, disclose rather than pad, put the number last, say what a link can and cannot protect) applied to the two conversations an agent actually has — a listing price a seller doesn’t want to hear, and a dual-hat disclosure when the licensee is the buyer. The investor’s margin line becomes the licensee’s disclosed-interest line, which is Chapter Twelve’s actual subject. Does not become a sixth module under any framing.
• Either way, the drafted assessment items need one pass against the addendum’s added originality rule — checking titles, objectives and topics against the IP review’s REFRAME list, not only against exam-source PDFs.

6. VIDEO NOTE. If built here: slide+narration with the stack computing line by line, the number arriving last — the same asset class as F5-U2 and B4-U3. Avatar: one clip on the honesty argument (a disclosed number is stronger than a hidden one), which is the unit’s whole thesis. If it ships to the REI school instead, that school’s audit owns the note.

RE32 · Two open judgments the book can help settle

1. Where it sits. C1-U4 / C1-U5 (merge or keep separate), and the naming of the four-topic rapport structure.

2. What promises it. Addendum §5.3: “Whether C1-U4 and C1-U5 should be one unit or two… this is a judgment, not a finding.” §5.5: “Whether the four-topic rapport structure should be named even in the school… the naming decision on top of it is still the owner’s.”

3. What exists today. Two craft units kept separate provisionally, and a naming decision deferred.

4. The gap. Both are cheap to settle and both block page generation. The book actually supplies evidence for the first one, which the addendum didn’t use.

5. What “deeper” means. Two recommendations, offered rather than assumed.

• Keep them separate. Book 3 names the two moments in two different places, for two different readers. CH19 names the first: “the instant a machine hands you someone and the conversation becomes entirely yours to run” — a stranger, a first contact, nothing signed. App B ¶153 names the second: “the actual craft of a listing consultation, practiced and refined” — a client, in a file, with documents and disclosures attached to specific stages. Merging them would put a first-contact rapport lesson and a disclosure-timing lesson in one 1,500-word page, and the second half would lose its DPOR routing. Keep two, and cross-reference.
• Name the four-topic structure, in the school only. The IP review found it pre-existing and non-proprietary, R40 already proposes naming it, and a listening structure that cannot be named is a listening structure students cannot look up, practice against, or discuss with each other. The constraint that matters is the one already in force: it is named as a listening structure, never as a script, and never with any source’s framing attached.

6. VIDEO NOTE. Both craft units are the school’s recorded role-play assets — two-voice conversation audio over minimal slides, which the cloned voice produces natively and which no other school in this space will have. If they merge, that asset halves; another reason to keep two.

RE33 · Book↔︎school pointer reconciliation

1. Where it sits. Appendix B (eight pointers), CH08, CH19, and the design’s track eligibility rules.

2. What promises it. App B ¶384: “Your state’s rule governs; the school’s licensing module carries it, current” — the licensing module is B1. ¶382: “the school’s agency-and-contracts module carries the current answer” — B3. ¶386, ¶388, ¶394: “the school’s compliance module” — C5. ¶390 and ¶392: B1 again. CH08 ¶67: “that’s the school’s agency-and-contracts module, B3.” CH10 ¶35/¶45/¶51/¶70: B1, B3, C5, B1.

3. What exists today. Design §1: Track B is “Broker-license upgrade pre-licensing”, recommended sequence A → B → C, “available as a standalone track for students who already hold a Salesperson license.” Track C is “open to any student who has finished A or B, and to already-licensed agents.”

4. The gap. Appendix B’s reader is defined by the book as not yet licensed — HOW-TO-USE ¶11 introduces them as “the not-yet-licensed reader” and sends them to Appendix B and the Foundations track. Appendix B then points that reader, eight times, at modules in the Broker track and the capstone track. As designed, the sentence “the school’s licensing module carries it, current” is false for the person it was written for: the licensing module is a broker-upgrade course they cannot take, containing another state’s law. CH08 does the same thing to a working salesperson — sends them to B3, a broker module. Three smaller items belong in the same pass: CH08 ¶63 promises the conversation craft (“the companion school teaches it as its own subject”) with no SCHOOL marker at all; App B ¶153/¶167 promise the listing-consultation craft under a Track A marker while the addendum designs it as C1-U5 in Track C; and CH19’s “It runs five modules” is accurate today and becomes a count the school must not quietly break — five modules is a printed number and R42 is what keeps it true.

5. What “deeper” means. Not a lesson — one reconciliation, and it determines where several units above actually live.

• Make the living map cross-track (RE01), and make the school’s public answer to “where does the current rule live” the map, not a module. The book’s sentences then become true for every reader: the map is reachable from Track A’s F5-U1 and from Track B’s B1-U0, and B1/B3 remain where a Broker student studies that law for an exam. This is the smallest change that fixes eight pointers at once.
• Open Track C’s compliance units to any reader. C5-U2 is already free in full by the addendum’s own reasoning; C5-U3 and C5-U5 should join it, and the eligibility rule should say plainly that the compliance units carry no prerequisite. Appendix B points an unlicensed reader at C5 three times; that must be reachable.
• Fix the two craft routings: add a C1-U5 marker where App B ¶153 currently reads Track A, or — better, since the book is final — build both the Track A version (F7-U3, RE09) and the Track C version (C1-U5), and have the school’s Foundations map say so. A final book cannot be re-marked; the school can be built to make its markers true.
• Publish the eligibility grid — which track a person may take at which licensure status — because three book sentences currently depend on an answer no document states.

6. VIDEO NOTE. None directly. One consequence: because the living map becomes cross-track, its explainer video (RE01’s avatar clip) is watched by every student in every track and should be produced first, at the highest quality, and kept to under a minute.

COUNTS

Shape of the work. Five new modules (Track A: F5, F6, F7, F8 — seventeen units; Track B: B4 — four units), one new cross-track asset (the living map), thirteen new Track C units (C1-U6, C1-U7, C1-U8, C2-U5, C2-U6, C2-U7, C3-U3, C4-U2, C4-U3, C4-U4, C5-U3, C5-U4, C5-U5), nine new units inside existing modules (F1-U3, F1-U4, F2-U4, F3-U3, F3-U4, F4-U5, B1-U0, B1-U7, B2-U4), three unit rewrites for roadmap honesty, six depth-pass builds attached to existing Track B units, one registry reconciliation, one unit-format spec, four delivery decisions, and three open rulings. Track A goes from 4 modules / 11 units to 8 modules / 34 units. Track B goes from 3 / 13 to 4 / 20. Track C stays at 5 modules — R42 holds — and goes from 14 units to 27. The program goes from 12 modules / 38 units to 17 modules / 81 units. The licensing-prep item bank goes from ≈945 to roughly 1,200 at the design’s own 10-per-unit rate (340 Track A quiz items, 160 for Track B’s sixteen tested units — B4 is post-licensing and carries none — plus the 705 full-length-exam items unchanged), and Track C’s separate completion pool from 140 proposed to 270.

Of the thirty-three, exactly four must happen before a single lesson page is written, and none of them is a lesson: RE01 (the living map’s schema and state scope), RE04 (the registry), RE05 (the unit format), and RE29’s state selector (page architecture). Everything else is cheaper after those four and more expensive before them.

IF WE COULD ONLY BUILD THREE

Build RE01 (the living state-law map), RE16 (C5-U3, Reading the Ladder), and RE07 (F5, Entering the Profession) — in that order.

RE01 first because it is the only need on this list that Book 3 asserts as a fact on its copyright page and repeats twelve times, and because it is not a lesson at all: it is a schema, a verification cadence, a date on a card, and an honest sentence for every state that hasn’t been checked yet. It can ship with the federal floor and Virginia verified and still make every one of those twelve book sentences true, because the promise the book made was never “we have all fifty states” — it was “we keep it current and we tell you where it came from.” Nothing else in this school can be marketed honestly until this exists, and everything else in it gets cheaper once it does, because eight Appendix B pointers, four Chapter Ten pointers, Track A’s missing state section, and the practice bank’s staleness problem are all the same problem wearing four hats.

RE16 second because it is one twenty-minute unit that makes twenty-six other units teachable. Every build in this school ends with a claim about where a thing sits on the authority ladder, and no student has ever been taught the ladder. It is short, it is free-tier by the rule in RE30, it is the best avatar asset in the corpus, and it is the difference between a track that teaches configuration and a track that teaches judgment — which is the only thing the book actually promised that a vendor’s feature list doesn’t.

RE07 third because it is where the school stops being a book with homework and starts being worth money to the person Appendix B was written for. Five units — the rules and where they live, the brokerage agreement, the money, safety, the ninety-day map — answer five of the twenty-four Track A markers directly, require no platform account, and are drafted from Appendix B’s own material at a depth the appendix explicitly said it wasn’t going to reach. They are also the cheapest video in the school: arithmetic building on screen in a cloned voice, one avatar clip about walking out of a showing that feels wrong, and not one sentence that goes stale when a legislature moves.

The fourth, if there is a fourth, is RE26 — the team intake standard — because Chapter Fourteen printed the words “belong to the school’s team module, walked step by step” and there are, today, no steps.

REI school — 28 needs

2026-09-02 · lens: the investor school Book 2 points its reader at · per audit/AUDIT-BRIEF.md

State of this school

The REI school is a complete architecture with no content in it. Two charter documents name twenty modules — RU1–RU5 (concepts), RP1–RP10 (one working automation each), RA1–RA5 (business-model electives) — and every one of those twenty exists as a single table row averaging about fifteen words. Zero module files exist. The general ProcessAutomater school has twenty-four written lesson files in the nine-section SCHOOL-MODULE-SPEC format (U1–U5, P01–P10, A1–A3, D1–D6); the REI school has none, and nothing in /school/ is written to a real-estate learner. What the REI school does have is 15,042 words of genuinely excellent source material that has never been converted into a lesson: a cell-by-cell decode of the competitor’s deal analyzer (2,184 w), a clean-sheet offer-engine spec with a verification ledger (1,323 w), a build list of eighteen automations, a lesson ledger with eleven cards assigned to SCHOOL-Investor, and one drafted unit (the Cost Defense, C6-U1) that was filed into the wrong school — it sits in the Book 3 agent school’s Track C table, where it does not belong, because everything it teaches is investor-side.

Meanwhile Book 2 has been finished around this school. Twenty chapters, an appendix of seven builds, and a first-rung appendix all assume the school is there. The book points at it explicitly five times — the state-by-state wholesaling map, private lending depth, the four alternative asset paths, credit strategy, and the whole-school promise in Chapter 20 (“full builds, module by module, for every structure and every stage this book has walked you through”) — and it makes three membership promises in present tense that are stronger than pointers: the term-sheet automation is “there for any reader of this book, free, inside the membership section” (Ch10, Ch12, echoed Ch17), and the Property Management Machine “is available to any reader of this book through membership” (Ch14). The sibling charter says RP9, the funding feeder, “ships only when the funding funnel is live + compliance/disclosure cleared.” Those two sentences cannot both be true at print. That is the single most urgent finding in this audit and it is not a curriculum gap — it is a truth-doctrine conflict with rule R-H, and it is flagged as R-20 below.

Ranking rule used here, since almost nothing exists and a flat wall of P0 would be useless: P0 = the book sends the reader here by explicit pointer or membership promise, or a chapter’s central machine has no module at all. P1 = a teaching-grade source artifact exists but has never been made into a lesson, or the lesson exists and is misfiled/too thin for the promised outcome. P2 = a named elective the book deliberately deferred. P3 = worth building, nothing promised it.

Ranked needs

R-01 · RP1 — The Nightly Sweep — P0

Where it sits. RE-PLATFORM, RP1 (“Lists: Where Deals Hide”), expanded from one module to four units: RP1-U1 the nightly pull, RP1-U2 the watch, RP1-U3 dedupe and the stack count, RP1-U4 the Monday queue and correcting it.

What promises it. Ch4: “None of that requires the machine to decide anything. It requires the machine to pull every one of those lists on its own clock… and land every record in one place instead of a browser full of tabs.” And the marker left in the manuscript itself: <!--BUILD-APPENDIX: the nightly-pull configuration — which lists, which refresh cadence, where records land-->. App A-1 promises “a ranked queue waiting for you on Monday morning.” Build list #2 (wholesaler-list intake) and #1 (expired-listing machine).

What exists today. school/REI-SCHOOL-SIBLING-CHARTER.md §2, one row: “RP1 | Lists: Where Deals Hide | Wholesaler-list intake: get on lists, parse inbound, staleness timers | 2”. Fifteen words. No file.

The gap. The single most-referenced machine in Book 2 — every chapter from 5 onward assumes a queue already exists — has no lesson, no build steps, no failure list, and no place for the learner’s own market thesis. App A-1’s seven steps are the book’s depth (“what you do, what comes back”); the school owes the configuration a learner actually performs, plus the one skill Ch4 says nobody teaches: stacking. A module that names “stack count” and never has the learner build one is exactly the thin module this audit is looking for.

What “deeper” means. - RP1-U1 — The nightly pull. Objectives: name the five list categories your own county actually publishes and which office publishes each; set a refresh cadence per list and defend it against how fast that list rots; land every record in one place instead of four tabs. Build: connect two real lists for your own county and run one overnight pull. You’ll know it worked when tomorrow morning’s records are in one place and you didn’t open a single county website to get them. 15 min. - RP1-U2 — The watch. Objectives: distinguish a scheduled pull from a continuous watch and say which of the four events belongs to which; set the four watches (new listing, price cut, probate filing, code violation); explain why a fresh probate filing is a slower clock than a price cut. Build: set all four, then wait for the first one to fire and write down the hours between the event and the notice. You’ll know it worked when you can name the exact hour a price cut happened, not the week. 15 min. - RP1-U3 — The stack count. Objectives: normalize an address so two counties’ spellings stop being two houses; merge duplicates without losing which list each copy came from; choose which lists to stack and defend that choice as your own market thesis. Build: run a dedupe pass on last week’s raw records, then rank by stack count and compare the top six against the raw eighty. You’ll know it worked when your six-row queue contains at least one address you’d have missed in the eighty. 20 min. - RP1-U4 — Correcting the queue. Objectives: identify a signal that means something different in your market than in the data; teach the queue that exception once so it holds every week; read a merge you disagree with and say why. Build: find one neighborhood where a pressure signal is structurally meaningless (Ch4’s student-rental absentee example), correct the weighting, and re-run. You’ll know it worked when next Monday’s queue is shorter and the reason you shortened it is written down. 15 min.

Video note. Slide+narration: the five list categories and where each comes from; the four watch events; the address-normalization example; the stacked-response-rate figure with its one-operator caveat spoken as written. Screen recording (required): the overnight pull landing, the dedupe merge view, and the Monday queue with stack counts visible — this is the module’s whole payoff and cannot be a slide. Avatar clip (one, ~40 s): the correction moment — “the judgment that never moves off you is which lists to stack in the first place.” That sentence is the school staking a position against every list vendor in the business and it should come from a face.

R-02 · RP3 — Feeding the Hunter — P0

Where it sits. RE-PLATFORM, RP3 (“Comps Without the Ritual”), expanded to four units: RP3-U1 comps before the call ends, RP3-U2 the county record, RP3-U3 the portal with no door (and its terms of use), RP3-U4 skip tracing and the one-good-source discipline.

What promises it. Ch5: “comps get pulled and adjusted before the call ends, not after. Not ‘we’ll run the numbers and follow up.’” And the hard line the school must carry: “Read the terms of use on the specific site before you automate anything against it. Where a site says no, that’s the whole answer — don’t do it. Where a site says nothing, treat the silence as a question worth a phone call, not as an invitation.” Build list #6 (comps collector), #3 (jurisdiction-letters pipeline).

What exists today. One charter row: “RP3 | Comps Without the Ritual | Address in → comps/records collected → results drawer | 6”. Fourteen words. The component map in REI-SCHOOL-CHARTER-AND-SOURCE-INVENTORY.md §4 has one line about the export/import ritual. No file.

The gap. This is the school’s largest liability surface and it has no lesson. Ch5 spends four sections on the wall — 484 MLSs each with its own contract, 2,690 of 3,100+ counties covered by the best-funded aggregator, Yardi’s two-years-and-three-clients gate — and then tells the reader to read a county’s terms of use before pointing anything at it. A learner who builds RP1’s sweep and RP3’s feeds without that lesson will automate against a county that forbids it, and the school will have taught them to. Separately, the “one good source beats five stale ones” discipline and the subscription-tax audit are the only defense the reader has against the exact cost pattern the narrator confesses to in Ch5 — and neither is anywhere in the school.

What “deeper” means. - RP3-U1 — Comps before the call ends. Objectives: state what a comp set must be adjusted for before it’s usable on a live call; attach a comp set to an address automatically at the moment it clears the RP1 queue; read a thin comp set and say so out loud instead of using it. Build: attach comps to five queue addresses, then time yourself answering “what’s it worth” without opening anything. You’ll know it worked when you can say a range out loud with the seller still on the line. 15 min. - RP3-U2 — The county record. Objectives: name the record facts that change the conversation, not just the number (years behind, other liens, lis pendens, heirs named, hold length); attach record history to the same lead; distinguish “tax-delinquent” from “three years behind, no other liens, sole heir named eight weeks ago.” Build: pull full record history on three queue addresses and write one sentence per address predicting the kind of call it will be. 15 min. - RP3-U3 — When the county has no door. Objectives: describe what to do when a record lives only behind a search form; read a specific county’s terms of use and classify it as forbids / silent / permits before automating anything against it; recognize what a scanned 1987 deed image does to accuracy and check the thin ones by hand. Build: pick one county you actually work, find and read its terms-of-use page, write down the classification and the date you read it, and only then set the collection. You’ll know it worked when you have a dated, written classification for every source you automate against — and at least one source you decided not to. 20 min. This unit’s lesson content is free in full, deliberately; a boundary behind a paywall is a boundary that doesn’t work. - RP3-U4 — One good source. Objectives: name the honest hit-rate range on skip-traced records and budget for the bad share instead of being surprised by it; run the subscription audit (list every feed, its monthly cost, and what it uniquely gives you); prune rather than stack. Build: put every data invoice you pay on one page in one sitting, mark the overlaps, and cancel one. You’ll know it worked when you can name which one source you’d keep if you could keep only one, and why. 15 min.

Video note. Slide+narration: the API-wall numbers (484 MLSs, 2,690/3,100 counties, 957 apps at 27% connected, the Yardi gate) — these are pure data-slide material and every one has a citation in the backmatter to display. Screen recording (required): a collection running against a real search form and coming back into the same intake as a connectored source — that side-by-side is the lesson in U3. Avatar clip (one, ~45 s, mandatory): the terms-of-use rule, spoken. “Where a site says no, that’s the whole answer.” A compliance boundary the school will be judged on should not be a slide with a robot voice.

R-03 · Every Way to Buy — the six structures, computed — P0

Where it sits. RU2 (“The Offer Family”) becomes a concept unit, and a new applied family carries the depth: RA-S1 cash & wholesale (with novation and wholetail), RA-S2 subject-to, RA-S3 some-now-some-later and hybrid, RA-S4 the wrap and its three gates.

What promises it. Ch9: “Six real structures need six real sets of fields, six sets of floors, six receipts — not one box with everybody’s numbers crammed into it.” And the wrap’s gates: “The down payment your buyer puts up has to leave you real profit on day one… the equity built into the note you’re holding has to clear a number you set in advance… and what’s left over every month… has to clear a floor of its own.” The offer-engine spec’s Design Law 5 says the same thing from the tool side: “Every deal structure is a first-class citizen.”

What exists today. RU2’s row: “The Offer Family | Cash / seller-finance / hybrid structures in plain words; when each fits; their own worksheet of a sample deal three ways.” Twenty-eight words, and it names three structures where the book teaches six plus two cousins. OFFER-ENGINE-SPEC §3.2–3.3 has the math for all of them.

The gap. The charter is a version behind the book. RU2 promises a worksheet of “a sample deal three ways”; Ch9 runs five structures side by side on one lead and then a full wrap end to end. Novation and wholetailing don’t appear in the charter at all, and both now carry a live legal caution (the 2025–26 equitable-interest statutes). Subject-to’s due-on-sale risk — the one structure in the book with a named, unhedged downside — has no lesson anywhere. And the wrap’s three gates, which the decode proved lived in a PDF and were enforced by nobody, are the school’s single best worked example of policy-as-code and it is unwritten.

What “deeper” means. - RA-S1 — Cash, wholesale, and the two cousins. Objectives: say when the edge on a house was finding it rather than owning it; distinguish assignment, novation, and wholetail by who is on the hook to the seller at closing; state the caution that rides with novation without stating any state’s law. Build: run one real queue address as a cash offer and as a wholesale assignment and compare the two receipts. 15 min. - RA-S2 — Subject-to. Objectives: explain what you are actually buying when you buy a debt’s position; state the due-on-sale clause plainly, including that “usually fine” is not “never happens”; identify the two seller situations where sub-to beats cash. Build: run a sub-to structure on a lead with a real existing loan and read the rolled-forward payoff at exit. 20 min. Compliance line at close, plus attorney pointer. - RA-S3 — Some-now-some-later and hybrid. Objectives: separate walk-away cash, kept mortgage, and seller note into three named inputs and never overload one for another; find the seller problem that is actually two problems; solve for the maximum walk-away the deal supports at your floors. Build: take one seller scenario and produce both an evaluate and a solve run, then say what the difference told you about your negotiation ceiling. You’ll know it worked when you have a walk-away ceiling you’d defend out loud, computed rather than guessed. 20 min. - RA-S4 — The wrap and its three gates. Objectives: describe the wrap as a buy side plus a sell side and name which structures its buy side borrows; set the three gates as your own named policy (day-one profit, note equity floor, monthly spread floor); read a note-equity break-even month. Build: configure the three gates with your own numbers, then run a wrap that fails exactly one of them and read which. You’ll know it worked when the engine tells you which gate broke before you’d have noticed it yourself. 20 min. Includes the “verify with a licensed professional” flag on seller-finance pricing compliance.

Video note. Slide+narration: the five-structure comparison on one lead (Ch9’s $240k/$48k example) as a build-up animation, one column at a time — this is the best pure slide sequence in the whole curriculum. The wrap’s parallel-amortization spread is a two-line chart. Screen recording: the structures computing side by side with each receipt expanding. Avatar clip (one, ~35 s, in RA-S2 only): the due-on-sale paragraph. A named risk in a structure the industry sells as a loophole is exactly where a face buys credibility, and it is the one place in this family where a slide would read as burying it.

R-04 · Stress, Floors, and the Walk-Away Ceiling — P0

Where it sits. New RP4-U4, sitting under the offer-engine module (R-21) rather than free-standing, because Ch10 is the discipline that makes Ch8’s number safe to send.

What promises it. Ch10: “when the cushion is thin, the stressed number isn’t a caution you glance at and set aside — it’s the number that goes to the seller, full stop.” And: “A floor you can override in the moment you most want to override it was never a floor.” Manuscript marker: <!--BUILD-APPENDIX: the stress toggle, step by step-->.

What exists today. Nothing in the school. OFFER-ENGINE-SPEC Design Law 8 has one line (“Stress is one switch: ARV −6%, rehab +15%, longest-DOM hold”) and §2 lists the policy gates as editable inputs. App A-4 is the book’s four-hundred-word walkthrough.

The gap. The floors are the only place in the entire system where trust runs from the machine to the investor, and that inversion has never been taught. There is also a live doctrine question the school has to answer and hasn’t: App A-4 says the stress toggle “doesn’t climb the ladder… it’s the one part of this whole system built to hold its ground,” which reads adjacent to a permanent gate. Ch10 resolves it correctly — “you can change every one of them, any time you decide they need changing” — and the module must carry that resolution explicitly, or the school ships a permanent gate in violation of the Graduation Law register.

What “deeper” means. One unit, four objectives: set three floors (profit, return, profit-as-percent-of-value) on a calm afternoon and write down why each number is that number; flip one switch and read which floor breaks first under a softer sale, an overrun rehab, and the slowest comp’s hold, all at once; say the difference between a stress case and a forecast; explain why the floors are policy you own and can revise, and why revising them mid-negotiation is the one thing the system won’t help you do. Build: run two leads through the toggle — one with real daylight, one thin — and write the sentence you’d say to each seller. You’ll know it worked when one of your two leads comes back with the stressed number as the actual offer and you can say, in one line, which floor made it so. 20 min.

Video note. Slide+narration: the two worked deals ($146,263 baseline versus ~$131,000 stressed; the thin deal’s $29,000 collapse), plus the Dodd-Frank stress-test parallel as a single credibility slide. Screen recording (required): one flip of one switch, every number in the deal recomputing — the whole point is that it is one switch and not four columns, which only a recording can show. Avatar clip (one, ~40 s): “the machine tells me the most I can pay before I ever fall in love with the house.” This is the founder’s own confession about kitchens and it is the emotional load-bearing beat of the entire offer track.

R-05 · Speed to Lead — P0

Where it sits. RP6 (“Follow-Up That Never Forgets”) re-scoped and split: RP6-U1 the nineteen minutes, RP6-U2 the sequence that stops on reply, RP6-U3 channels and consent.

What promises it. Ch11: “Nineteen minutes. Lead to a real, considered, stress-tested, receipt-backed offer sitting in a seller’s hands.” And the honesty half: “the rules on how and how often you can contact someone vary by state and by the consent you actually have — check your own before you turn a cadence like this loose on a real list.” Two manuscript markers: the timestamped offer run, and the no-reply follow-up routine.

What exists today. RP6’s row: “Follow-Up That Never Forgets | Compliant outreach sequences; reply-halt; review queue | 7 (dispo variant).” Eleven words of scope. The general school’s A2 (A2-sales-and-leads.md, 1,431 w) teaches the review-queue and auto-stop pattern well and is the closest thing to a template — but it is a B2B sales module with no seller, no offer package, and no distressed-channel consent law.

The gap. The chapter’s whole claim is a clock, and a clock cannot be taught by description — the learner has to run one and see their own timestamps. Nothing in the school does that. Worse, Ch11 authorizes a narrow class of unattended send (“you authorized this one narrow class of send along with it — check-ins scheduled against offers you’d personally signed off on”) and that is the first place in the reader’s journey where something goes out without a per-message look. Teaching that authorization badly is how a learner turns a cadence loose on a purchased list. The channel-consent unit is not optional and it is not written.

What “deeper” means. - RP6-U1 — The nineteen minutes. Objectives: chain RP1→RP3→RP4 into one run and read the timestamps; identify which lead types should be slower and why a flagged borderline number is the system working; state what the speed edge actually is (“first with the real number, never first with a guess”). Build: run one real lead end to end with the clock on and record every timestamp. You’ll know it worked when you have your own timestamps from watch-list match to sent package, and you can point at the minute that took the longest. 20 min. - RP6-U2 — The sequence that stops. Objectives: set a spaced cadence whose messages differ from each other rather than repeat; verify the auto-stop on any reply before the first real run; read a cadence that got complaints instead of silence and shorten it rather than defend it. Build: run the cadence against two dead leads and one live one and confirm the live one’s sequence ended the instant they answered. 15 min. - RP6-U3 — Channels and consent. Objectives: distinguish “I found the number” from “I’m cleared to use it”; name which channel your machine defaults to when consent is unknown, and why letters survive that test; state that your outbound templates are a compliance surface and read yours once, carefully, before it runs. Build: read your own three outbound templates aloud and mark every sentence that makes a claim you couldn’t defend. You’ll know it worked when at least one sentence in your own templates changed. 15 min. Free in full.

Video note. Slide+narration: the timestamp ladder (6:52 → 7:11) as a build-up, and the MIT lead-response numbers with their honest caveat spoken as the book states it. Screen recording (required): the run itself, on a real clock, in one unbroken take — a cut in this recording destroys the claim it exists to prove. Avatar clip (one, ~30 s): “no web page stops a phone pointed at a screen.” The one place a tool this good tempts an owner into overselling, said by a person.

R-06 · Magic Math — the BRRRR engine — P0

Where it sits. RU3 (“Magic Math”) promoted from one concept row to a four-unit suite inside RE-UNIVERSAL, with the calculator itself in RE-PLATFORM.

What promises it. L-004 (ASSIGNED: TOOL priority build + B2-Ch12 + SCHOOL-Investor Magic Math module), verbatim: “School version MUST be more sophisticated than the book example: include private-money / fix-flip loan cost (~11% ± + closing costs), refi closing costs + points BOTH ways.” Plus L-014’s requirement that cash-left-in is a first-class output with a rate-sensitivity view, and L-011’s CKC feed keeping the lender terms current. Ch12: “the tool doesn’t hand you one answer — it runs the same deal across a range of rate scenarios, so you can see exactly where your market sits on the easy-to-hard curve.”

What exists today. RU3’s row (24 words, and it currently describes something narrower than the ledger card: “BRRRR full-capital-return explained honestly: appreciation + rent + depreciation”). L-004 and L-014 are detailed ledger cards, not lessons. No file.

The gap. The ledger already ruled that the school version must exceed the book’s example, and named exactly what it must add — money cost both ways, both refinance bases, cash-left-in as a dollar figure, rate sensitivity. None of it exists. This is also the module most exposed to going stale: every number in it is a lender term with an expiration date, which is precisely why L-013’s Depth Test sent it here. Building it without the live-terms dependency (R-20) resolved produces a module that teaches with frozen rates — the exact failure the book accuses guru books of.

What “deeper” means. - RU3-U1 — What the math actually solves. Objectives: work backward from a lender’s coverage floor to the minimum rent a specific house has to command; price both refinance bases (appraised value versus purchase-plus-documented-improvements) and say which your lender actually uses; name every input that has an expiration date. Build: solve required rent for one real house at your own lender’s current terms. 20 min. - RU3-U2 — Cash left in. Objectives: read cash-left-in as a dollar figure rather than a pass/fail; run one deal across a rate range and locate your market on the easy-to-hard curve; state what a deal that traps cash is and isn’t. Build: run the same house at three rate scenarios and write the sentence that changes between them. You’ll know it worked when you can say, in dollars, what a two-point rate move costs this exact house. 20 min. - RU3-U3 — The cost of the money, both ways. Objectives: include fix-and-flip loan cost, points, and closing costs on the way in and refinance points and closing on the way out; show what the book’s example leaves out and what it does to the answer. Build: re-run the book’s own worked shape with full money cost and report the difference. 20 min. (This unit is the ledger’s explicit “more sophisticated than the book” requirement, discharged.) - RU3-U4 — The conditions, and when they don’t cooperate. Objectives: name the two conditions a free-house outcome depends on; describe what happens when a lender pulls approval mid-pipeline or an appraisal comes in soft; size reserves against that case. Build: stress one BRRRR against a soft appraisal and a rate move and record what it would have cost you to close the gap. 15 min. Hard disclaimer, CPA/lender pointer.

Video note. Slide+narration: the required-rent solve, the two refinance bases, the rate-sensitivity curve (one chart, three lines) — all clean, all automatable from lesson text. Screen recording: the calculator run, with cash-left-in visible as a number. Avatar clip (one, ~35 s): “In a 5% world the house buys itself back. In an 8% world you tip it.” The ledger flagged that as a teaching line; it is the one sentence in this module a learner will repeat, and it should carry a face.

R-07 · Entity & the Ecosystem — P0

Where it sits. New RU6, RE-UNIVERSAL, two units.

What promises it. L-002 (ASSIGNED: B2-Ch16 + SCHOOL-Investor Structure module) and L-003 (ASSIGNED: SCHOOL-Investor Ecosystem module), whose card carries the load-bearing line: “a construction company you own — legitimate accounting, no shell, actually hires and manages the buildout.” Ch12: “That line only holds because nothing about it is fake. A shell that invoices itself for work nobody did isn’t a lever, it’s fraud.” Ch16 adds the mortgage-count ceiling and the personal-guarantee caveat.

What exists today. Two ledger cards, both marked DISCLAIMER REQUIRED. Nothing in either charter. App F’s “First Rung: Entities” section (about 700 words) is the reader-side floor and explicitly stops at “ask a licensed professional.”

The gap. This is the highest-risk topic assigned to the investor school and it has no module, no disclaimer footer, and no attorney/CPA pointer structure. The ecosystem play (own the construction company, own the PM company, bill at real margins) is genuinely valuable and one sentence away from teaching people to build a shell. The legitimacy line has to be the module’s spine, not a caveat at the end — and there is currently no module for it to be the spine of.

What “deeper” means. - RU6-U1 — Whose name is on the house. Objectives: compare personal-name and entity ownership on three axes (liability, financing shape, mortgage-count ceiling); explain why a DSCR loan underwrites the property rather than the person; state plainly that most lenders still want a personal guarantee, so the entity separates you from a tenant’s lawsuit better than from the loan. Build: map your own next three purchases to a structure and write the two questions you’d take to an attorney. 15 min. - RU6-U2 — The businesses inside the business. Objectives: describe the customer-and-supplier loop (construction, management, ownership) and what each does for the others; state the legitimacy test — real payroll, real permits, real invoices, priced as it would be with a stranger; identify when a related-party transaction stops being a lever and becomes a problem. Build: price one rehab twice — as a stranger’s bid and as your own company’s bid — and justify the margin in one sentence a CPA would sign. You’ll know it worked when the two numbers differ and you can defend the difference as real work at a real price. 20 min.

Video note. Slide+narration: the three-entity loop as a simple diagram; the liability/financing comparison table. Screen recording: none needed — this module has no platform surface, which is fine and worth noticing. Avatar clip (one, ~40 s, mandatory): the legitimacy line, verbatim from Ch12. This is the single sentence in the REI curriculum most likely to be quoted back at the school by someone who did it wrong, and it must be on record in the founder’s own voice, followed by the not-legal-advice line in the same clip.

R-08 · Tax Awareness — P0

Where it sits. New RU7, RE-UNIVERSAL, two units.

What promises it. L-006 (Flip/wholesale vs hold — the tax fork; ASSIGNED SCHOOL-Investor Tax Awareness module, with an OPEN item: “owner to enumerate — capture the four” other reasons) and L-007 (REPS, VERIFY: IRS code cites, HARD DISCLAIMER). Ch12: “Cross a federal threshold… and the tax code stops treating you as a passive landlord and starts treating you as a real estate professional, licensed or not.” Ch16 repeats it at portfolio scale.

What exists today. Two ledger cards. App F’s “Taxes” and “1031 Exchanges” sections (~1,100 words combined) are reader-floor material that ends at “ask a licensed CPA.” Nothing in the school.

The gap. Two chapters name REPS with a specific federal threshold and both hand off to the school implicitly by treating it as “the reason to have that conversation.” L-006 has an unresolved OPEN item — the four other reasons to hold — which means the module can’t be drafted to completion without an owner pass. L-007 carries a VERIFY flag on the code cites that must clear before anything is published. Neither has moved since 2026-08-24.

What “deeper” means. - RU7-U1 — The fork. Objectives: state what changes at tax time between a flip/wholesale and a hold, in plain words; explain depreciation as a paper loss against a real asset without overstating it; name what a 1031 defers and what it does not. Build: take one deal you actually ran and write both versions of its tax treatment, then list the three questions you’d bring a CPA. 15 min. - RU7-U2 — Hours that change your status. Objectives: describe the two-part federal test (an hours threshold and a majority-of-working-hours test, with material participation) as a reason to track, never as advice; name which activities the code counts; explain why serious hold investors log hours like payroll. Build: start an hours log for one month and categorize a real week. You’ll know it worked when you have a month of real hours logged and a written question for your CPA that isn’t “am I a real estate professional?” 15 min.

Constraint flags. (a) L-006’s four reasons are OPEN — draft blocks until captured. (b) L-007’s IRC cites must be verified at draft. (c) HARD DISCLAIMER in the module footer per L-012, plus a CPA pointer in-body, not just at the end. (d) The living-map rule applies: thresholds change, so the specific numbers live in a dated, revisable panel, not in narration.

Video note. Slide+narration for everything, with one exception — and put every threshold on a slide with a visible “current as of” date so a re-render updates one asset rather than a whole video. Screen recording: none. Avatar clip (one, ~25 s): the disclaimer itself. A tax module is the one place where a person saying “this is education, not advice, and here is why you’re calling a CPA” is worth more than the same words in text.

R-09 · Close & Multiply — P0

Where it sits. New RP-TC (“The Transaction Chain”), RE-PLATFORM, four units. Absorbs build-list #7 (dispo blaster).

What promises it. Ch13’s entire chapter, which contains the school’s cleanest reusable insight: “a real estate transaction can have a hundred variables or a thousand, but they’re the same hundred or thousand variables, on every file… which is precisely what makes the work priceable and, eventually, automatable.” And the gate: “A change to wiring instructions is one of them… Real estate closings lost more than $275 million to wire fraud in a single year.”

What exists today. Nothing. Not even a charter row — RP1–RP10 has no transaction-coordination module; the closest is RP8 (PM turnover), which is a different chapter.

The gap. An entire chapter of Book 2, and one of the most obviously automatable stretches of the business, has no home in the charter at all. The buyer-list-from-deed-records technique (a recorded deed with no mortgage alongside it) is a genuinely teachable, genuinely differentiated skill that appears nowhere in the school. The draw-approval discipline and the wire-fraud gate are both hard boundaries the school must teach and currently doesn’t.

What “deeper” means. - RP-TC-U1 — Who’s buying this. Objectives: read a deed-without-mortgage pairing as a cash-purchase signal; build a buy-box profile from what an entity has actually bought; prune a buyer list the same way you prune a seller list. Build: build a live cash-buyer list for one zip code from public records and age-rank it. You’ll know it worked when your list contains a buyer you’d never heard of and excludes one you’d have called out of habit. 20 min. - RP-TC-U2 — The same hundred variables. Objectives: write the closing checklist that recurs on every file; set per-file clocks that know the difference between two days late on a five-day window and on a thirty-day one; route only the exceptions to yourself. Build: run three files through one checklist for one week and count how many times you were interrupted. 20 min. - RP-TC-U3 — Draws and verified-is-payable. Objectives: match a photo set against a specific scope line before a dollar moves; release a routine match without eyeballing every photo; kick back anything that doesn’t obviously match, with no relationship-based benefit of the doubt. Build: run two draws, one clean and one deliberately ambiguous, and read what each did. 15 min. (Shares its spine with R-25.) - RP-TC-U4 — The gates with no path up, and the ones with one. Objectives: distinguish an owner-set policy gate from the permanent licensed/consent ceiling; state the wire-instruction rule (confirm by phone, to a number you already had, before a dollar moves) and why it stays; explain why a missed contingency deadline is handed to you rather than guessed at. Build: write your own gate list — which tasks may graduate, which you’re holding, and the date you’ll revisit each. You’ll know it worked when every gate on your list has a review date, including the ones you never intend to open. 15 min.

Video note. Slide+narration: the checklist sequence; the JV split conventions; the $275M wire-fraud figure. Screen recording (required): the three-files-one-morning view — the flag that fires because a bid is late against a five-day window is the whole argument and it has to be seen firing. Avatar clip (one, ~30 s): the wiring-instructions rule. Nothing else in this module earns one.

R-10 · The Property Management Machine — P0

Where it sits. RA3 (“The Management Company”), RE-APPLIED, three units. Build-list #10 and #11 feed it.

What promises it. Ch14, in present tense, as a member benefit: “It’s available to any reader of this book through membership. Your data is always yours; you can see it, export it, and take it with you if you ever leave.” Manuscript marker: <!--BUILD-APPENDIX: the PM morning digest — setup, categories, notice-schedule configuration-->. App A-6 walks the digest.

What exists today. RA3’s row: “The Management Company | Tenant self-showings, prequalification-as-automation, the escalation ladders, newsletter engine.” Thirteen words — and note it describes a different product than the book does. The book’s PM machine is triage, rent tracking, maintenance routing, and a morning digest; the charter row is showings and prequalification. Neither list contains the other. The general school’s D1 demo (demos/D1-morning-rent-check.md) is a 3-portal rent check for a property manager persona and is the nearest existing artifact.

The gap. A named product is promised to every reader of the book, by name, as available through membership — and the school’s own charter describes a different feature set. Before a module can be drafted someone has to reconcile which one is real. Beyond that: the module has to teach the escalation-with-action pattern (bucket under the leak, water off at the main, walk the breaker box) that Ch14 spends a full paragraph on, because that pattern is the difference between a triage bot and a machine that reduces damage, and it appears in no spec anywhere.

What “deeper” means. - RA3-U1 — Triage and the drafted reply. Objectives: sort an inbound tenant message into question / maintenance / payment; distinguish the zero-weight answers that may go out on their own (trash day, gate code) from everything that drafts and waits; explain why a tenant reads every message as a signal about their landlord. Build: run one week of real messages with everything drafted, and count how many you sent unchanged. 15 min. - RA3-U2 — Documentation and the first action. Objectives: ask for the proof that actually documents this issue (a video of a disposal humming, not a photo of a sink); pair every urgent category with the tenant’s first action, walked step by step; route to the right vendor category with the documentation attached. Build: write the first-action script for your own three worst emergency categories. You’ll know it worked when a tenant could shut off the water from your message alone, without calling you. 20 min. - RA3-U3 — The morning digest. Objectives: read one list in the order that matters instead of forty notifications at random hours; set the rent-reminder schedule that firms up on a written cadence rather than your mood; name what stays human (leases, evictions, walking a property) and why an absent landlord isn’t an automated one. Build: stand up the digest and read one week of it. 15 min. Landlord-tenant disclaimer with a local-law pointer.

Video note. Slide+narration: the four things happening every day; the message-category taxonomy. Screen recording (required): one real digest, read top to bottom in five minutes — Ch14’s proof is “done before the coffee finished brewing” and that is a recording, not a claim. Avatar clip (one, ~35 s): “a landlord who never talks to a tenant, never walks a property, never makes a judgment call isn’t automated, he’s absent.” The honest ceiling of the whole PM track.

R-11 · The Portfolio Machine — P0

Where it sits. RP8 (“The Turnover Chain”) expanded from one module to five units, absorbing build-list #10 and reaching into #12 (books).

What promises it. Ch15’s whole chapter. The arithmetic the school must teach: “a unit rents for $2,500 a month. That’s roughly $82 a day in rent you don’t get back for every day the unit sits unrecovered.” And the gate: “This is one of the few places where I’ve written the gate myself and left no path up: a filing that puts a tenant out of their home stops at me, every time.”

What exists today. RP8’s row: “The Turnover Chain | PM turnover: dispatch → verification follow-up → lockbox/photos → listing | 10.” Twelve words for a chapter that covers turns, the vendor bench, renewals, the delinquency-to-filing chain, short-term calendar pricing, and bookkeeping.

The gap. One charter row is carrying six distinct businesses. Short-term/mid-term operation — a growing slice of the audience by the book’s own account, with a completely different failure mode (a weekend priced wrong is gone permanently) — has no charter presence at all. Bookkeeping/extraction is on the build list as a main-school item (#12) even though Ch15 makes it a portfolio requirement. And there is a doctrine problem to resolve: Ch15 says the eviction gate has “no path up,” while Ch17 says “I expect that to change” and R-F says the ceiling is a current position, not a permanent principle. The module must teach it as an owner-set policy gate with a review date — the licensed/consent ceiling is the only permanent one — or the school ships a contradiction.

What “deeper” means. - RP8-U1 — The turn that dispatches itself. Objectives: build a turn checklist off a property’s own history rather than a generic template; rank a vendor bench by availability and price for a job that size; state what an emergency-call price costs you when the bench has a hole. Build: build your bench and run one turn checklist. 20 min. - RP8-U2 — Verified is done. Objectives: apply “scheduled is not done, verified is done, and verified is what gets paid” to a photo against a line item; clear a routine match and kick back an ambiguous one; explain why the standard is a better deal for the good contractors. Build: qualify three real payouts. 15 min. - RP8-U3 — Renewals ninety days out. Objectives: pull a comparable-rent read against a policy band you set; let the ordinary case go out on its own and stop the flagged ones; name what a lease sitting under market for a year actually costs. Build: run one renewal cycle. 15 min. - RP8-U4 — The chain you hope you never need. Objectives: compute your own per-day cost of an unrecovered unit; verify each step in the notice chain actually happened on the day it was supposed to; set the filing gate as your own policy with a review date, and say why that is different from the permanent licensed ceiling. Build: write your own chain with your own numbers and your own gate list. You’ll know it worked when every step has a checked date and the one step that stops at you has a date you’ll revisit it. 20 min. Eviction law varies; attorney pointer, no law stated. - RP8-U5 — Pricing a calendar, and the books. Objectives: set a floor rate and blackout rules before any pricing proposal is allowed to move; catch a comparable-read skew before approving on autopilot; land documents in the books already categorized and spot-check until the pattern earns the check off. Build: run one week of queued rate proposals and reject one with a written reason. You’ll know it worked when you caught a bad proposal and the same skew never fooled the read twice. 20 min.

Video note. Slide+narration: the $82-a-day arithmetic; the notice chain as a sequence diagram; the bench economics. Screen recording: the turn checklist building itself off a property record; the rate queue with a rejection. Avatar clip (one, ~30 s, in U4 only): the filing gate — the founder saying which line he drew himself and that he’ll revisit it, which resolves the doctrine conflict on camera instead of in a footnote.

R-12 · The Freedom Number & the Long Game — P0

Where it sits. New RU8, RE-UNIVERSAL, two units; the calculator ships inside the Magic Math suite per L-015.

What promises it. L-015 (ASSIGNED: B2-Ch16 beat + SCHOOL-Investor Freedom Number calculator inside Magic Math suite), permission AUTHORIZED 2026-08-24 with attribution required. Ch16: “Feed it the income you’re trying to replace and your market’s real cashflow-per-door, and it hands back the door count; feed it a rate change or a rent shift next year and it re-runs the same number instead of leaving you holding a napkin calculation from two years ago.” Manuscript markers for both the Freedom Number run and the hold/sell/refinance run.

What exists today. L-015 (a detailed card with the full twenty-house model and the authorization evidence). No module, no calculator, no charter row.

The gap. The chapter names a tool, describes its inputs and outputs, and promises it re-runs when rates move. It doesn’t exist. The hold/sell/refinance three-way comparison — the thing that makes a portfolio a decision rather than a pile — has the same problem. And the attribution obligation travels with the material: the school owes Connor Steinbrook in-text credit anywhere the model or the term appears, plus the passage-preview courtesy the ledger records as still owed.

What “deeper” means. - RU8-U1 — Your number. Objectives: compute doors-needed from target income and real per-door cashflow; say why per-door cashflow must be measured after reserves and vacancy, not before; explain what changes when the number is a count instead of “more.” Build: compute your own number with your own market’s real per-door figure, then re-run it at a rate two points higher. You’ll know it worked when you have a door count you’d put on a calendar and a second one you’d put on a worse calendar. 15 min. - RU8-U2 — Hold, sell, or refinance. Objectives: price a sale off real net operating income against a real market cap rate; project the hold against where the amortization curve actually sits now; price a refinance at current loan-to-value against a real coverage floor. Build: run all three on one house you actually hold and log which you chose and why. 20 min. Model-not-guarantee disclaimer, verbatim in the footer.

Video note. Slide+narration: the whole twenty-house model, step by step — it is the single best slide sequence available anywhere in the curriculum, because the arithmetic is the argument and every figure is cited. Screen recording: the calculator re-running at a second rate. Avatar clip (one, ~20 s): the attribution — naming Connor Steinbrook and the Investor Army channel out loud. Attribution spoken by a person is both the courtesy the ledger promised and the school’s own truth-doctrine posture, visible.

R-13 · The Five Dials — P0

Where it sits. New RP-KPI, RE-PLATFORM, two units. Absorbs build-list #16 (digest) on the ownership side.

What promises it. Ch17: “It’s five dials — one per task class, exactly as many as the tasks actually need — each one earning its own room to move at the speed its own stakes allow.” And R-F’s affirmative case, which Ch17 states directly: “a reminder a machine sends goes out in the same words, on the same schedule, to every tenant… which is exactly the posture fair-housing compliance asks for, and it’s a posture that’s easier to prove than a person applying the same rule from memory on a bad day.” Manuscript marker: the maintenance-spend qualification gate.

What exists today. Nothing — no charter row, no card. The general school’s A3 (A3-ops-and-finance.md, 1,486 w) is the closest pattern and is about invoicing and expenses, not portfolio KPIs.

The gap. Ch17 is where the authority ladder finally becomes concrete — five task classes climbing at five different speeds — and it is the clearest teachable statement of the Graduation Law anywhere in either book. It has no module. The affirmative compliance case (R-F) is stated once in the book and nowhere in the school, which means the school currently has no lesson making the argument that automation increases fair-housing defensibility.

What “deeper” means. - RP-KPI-U1 — Five numbers that watch themselves. Objectives: name the five (occupancy, delinquency, maintenance spend against a property’s own history, rent-to-market, refinance opportunity) and what each catches early; replace a budgeting rule of thumb with a running total checked against that property’s own history; read a renewal proposal with its comps attached. Build: stand up the panel and read one week. 20 min. - RP-KPI-U2 — Five dials, five speeds. Objectives: place each of the five on the ladder and defend the placement by what being wrong costs; make the affirmative compliance argument — one rule applied identically to every tenant is provable in a way memory isn’t; identify the one class you’re holding and write its review date. Build: write your own five-dial placement with a cost-of-being-wrong sentence for each. You’ll know it worked when two of your five sit on different rungs and you can say why in one sentence each. 20 min.

Video note. Slide+narration: the five metrics with their industry figures (34.4 vacant days, $3,872 per turnover, the 91%/89% late-fee finding, the landlord rent-increase gap) — a strong data sequence, every figure cited in the backmatter. Screen recording: the panel; a payout clearing under a cap and one kicking back. Avatar clip (one, ~40 s): the fair-housing argument. R-F says this case is affirmative, not apologetic — an apologetic delivery is exactly what a synthesized read of hedged text produces, so this one needs a person making the case.

R-14 · The Double Edge, investor side — P0

Where it sits. New RA6, RE-APPLIED, two units. Build-list #1 (expired-listing machine, RP5) is its engine.

What promises it. Ch18, and — critically — App A’s exclusion: “One machine this book names doesn’t get a walkthrough here: the double edge Chapter Eighteen touches on is a licensee’s build, and its full step-by-step lives in Automating Real Estate Agency.” Ch18 also draws the line: “Read your own state’s current statute, or have a real estate attorney licensed where you buy read it for you, before you scale any outreach past the size where a mistake is cheap.”

What exists today. Nothing. RP5 exists as a row (“Offers at Scale | Expired-listing machine: scrape → identify → offer → walkthrough scheduling | 1”) and is one of the three modules the sibling charter names for the beta bundle — but the agent-respecting fork, which is the whole point of Ch18 and Ch20’s Marcus build, is not in that row.

The gap. The book explicitly routes this build out of Book 2’s appendix and into Automating Real Estate Agency — a different book for a different school. But everything in Ch18’s first half is available to an unlicensed investor today and requires no license. So the investor school owes the build, and neither school currently has it. The forking logic (agent attached → note goes to her; genuine FSBO → note goes to the owner) is the specific thing the book teaches investors not to skip, and the outbound templates it produces are, in the book’s own words, “a compliance surface.”

What “deeper” means. - RA6-U1 — The buyer agents call first. Objectives: answer an address with a number and its basis on the same call; give a fast honest no and understand why it’s worth nearly as much as a yes; keep an agent relationship alive with a current buy-box description and a quiet flag when someone’s gone silent for six months. Build: write your one-paragraph buy box and set the relationship-gone-quiet flag. 15 min. - RA6-U2 — Expired listings, the right way round. Objectives: fork every expired address on whether an agent was attached before anything drafts; write the note that goes to her — buyer not competitor, number and basis, contract and window, and the offer to let her represent someone; verify current listing status immediately before sending so a note about an expired listing never lands on a relisted one. Build: run one expired sweep, produce both branches, and read your own outbound language once against your state’s rule before anything sends. You’ll know it worked when your sweep produced two different drafts for two different addresses and you changed a sentence in one of them. 20 min. Licensed-activity line at close; routes to R-19.

Video note. Slide+narration: the fork as a decision diagram; the “what goes wrong” list (relisted status, protection periods). Screen recording (required): the sweep producing both branches from the same night’s data. Avatar clip (one, ~45 s): “the agent is not the competition” and the cost of routing around her. This is the module’s entire thesis and it argues against something the learner has heard from a stage; a stage argument needs a face to answer it.

R-15 · Earned Authority: the Review — P0

Where it sits. New RP-AUTH, RE-PLATFORM, one unit, positioned immediately before the capstone.

What promises it. Ch19: “I do this myself every quarter, and it’s less ceremony than it sounds. I sit down with the log — every offer the machine proposed in the period, sitting next to what I actually did with it.” Two manuscript markers: reading a track-record log by decision class, and pulling an authorization back when a market shifts. App A-7 is the book’s walkthrough.

What exists today. Nothing. The general school’s U4 (U4-trust-receipts-quiet-contract.md, 1,529 w) teaches the trust ladder generically and is the right template — but it has no decision classes, no dollar caps, no structure bounds, and nothing about real estate.

The gap. The book’s final teaching chapter — the graduation the whole arc walks toward — has no lesson. And it is the one module that discharges the Graduation Law directly: the ladder is prepare → propose → propose-with-track-record → authorized, with human-on-exception as the only permanent landing. Without this module the school has no place to teach that a gate has a review date, which is exactly the failure that would let R-11’s eviction gate and R-09’s wire gate read as permanent architecture.

What “deeper” means. One unit, four objectives: read a track-record log by decision class and tell convergence from noise; name a class narrowly enough to authorize it (dollar cap, structure set, market conditions) and write the bounds down; state what still stops the line no matter how good the record gets, and why that landing never graduates; narrow or revoke an authorization the same way you granted it, in writing, the moment the record stops backing it. Build: pull one class’s log — the follow-up cadence from R-05 is the fastest, because a cadence returns a verdict in weeks — read it, and either sign a bounded authorization or write down what the record still owes you. You’ll know it worked when you have either a signed, bounded, dated authorization or a written statement of exactly what evidence would earn one. 20 min.

Video note. Slide+narration: the four rungs plus the permanent landing; the worked quarter (140 proposals: unchanged / adjusted / rejected) as three stacked bars. Screen recording (required): a real log with proposals next to decisions, and a bound being set — the whole module is “read your own record,” and reading is a screen act. Avatar clip (one, ~50 s): “you didn’t hand over judgment… it’s the highest form of owning it.” The book’s thesis, and the last thing a graduating member should hear from a person rather than a slide.

R-16 · The Community of Owners — capstone and gallery — P0

Where it sits. RP10 (CAPSTONE) plus the community-rules unit the sibling charter §4 already drafts in policy form.

What promises it. Ch20, the whole-school promise: “The school exists past that door, and it’s real, and it goes as deep as anyone wants to take it — full builds, module by module, for every structure and every stage this book has walked you through, taught by people who’d rather show you the whole thing working than gate it behind a homework assignment.” And: “what you build there is yours, and it doesn’t have to stay locked inside your own account.”

What exists today. RP10’s row: “CAPSTONE — Your Deal Pipeline | Chain RP1→RP8 into one Routine; publish the build; install someone else’s.” Sixteen words. The sibling charter’s §4 has eight community rules — the AI host disclosed as AI, the always-on moderator, the no-queue consultant, real-results-only testimony, affiliation disclosure, escalation to a person — which is real policy and not yet a lesson.

The gap. Two things. First, the capstone cannot be written until RP1–RP8 exist, so it is the module most dependent on everything else in this list. Second, Ch20 is careful about what it claims — the shelf of installable builds is described “as the direction, not as something you can browse this afternoon” — and the school’s capstone must inherit exactly that hedge under R-H. A capstone that says “install someone else’s build” before the gallery ships would break the book’s own honesty.

What “deeper” means. - RP10-U1 — Chain it. Objectives: chain RP1→RP3→RP4→RP6 into one pipeline with one review gate and one schedule; identify the two points in your own chain that genuinely need you; name what your chain does badly and what rule would fix it. Build: run the whole chain on one real lead, end to end. You’ll know it worked when one lead moved from a county filing to a drafted, receipted offer without you opening four tools. 25 min. - RP10-U2 — Publish it. Objectives: publish a build with its receipt and its before/after number; state what a testimony may claim (numbers you got, not numbers you hope for) and why the community’s credibility is the product’s; identify what tuning any installed build still costs the next owner. Build: publish one build and write its honest limits alongside it. 15 min. Carries the affiliation-disclosure rule and the AI-host disclosure verbatim from the charter.

Video note. Slide+narration: the chain diagram; the community rules read as a covenant. Screen recording (required): the chain running once, unbroken, and a build being published. Avatar clip (one, ~60 s — the longest in the curriculum, and the only one that should be): the standing invitation from the end of Ch20, and the honest hedge about the shelf. The founder’s own closing argument is the one place a cloned avatar is doing exactly what an avatar should do.

R-17 · Funding & the First Rung — P0

Where it sits. New RA7, RE-APPLIED, two units; L-001 calls it the “Funding module.”

What promises it. App F, explicitly handing off: “Credit strategy — which cards, which lenders, how to sequence financing across a growing portfolio — is exactly the kind of fast-moving, situation-specific territory this book treats as the school’s job to keep current, not the book’s job to freeze in print.” L-001 (ASSIGNED: B2-Ch10/12 + SCHOOL-Investor Funding module + F pointer), including the naming rule: “the third-party services stay UNNAMED in all reader-facing text; the route is presented through Century Key Capital’s financing partnerships, NAMED with URL.”

What exists today. L-001, a dense ledger card with the owner’s own numbers and a hard disclaimer flag. App F’s Credit section (~450 words) ending in the pointer above. Nothing in the school.

The gap. The book deliberately removed content and named the school as its destination. That destination does not exist, which converts a graceful hand-off into a dead link. There is also an unresolved naming conflict the drafter cannot decide alone: L-001 orders CKC named with a URL in reader-facing text, while VOICE-AND-RULES §5 bans internal project names and Book 2 itself never names it (it says “the automation in this book’s membership section”). The school’s public framing per L-011 is “our live lender data.” Owner decision required before drafting.

What “deeper” means. - RA7-U1 — Credit is a track record. Objectives: describe personal and business credit as two separate track records and keep the finances genuinely separate from day one; name what stronger credit actually buys (options, rate, down payment) and what property-income lending offers when credit is thin; tend credit before the week you need it. Build: separate your accounts and pull both files. 15 min. - RA7-U2 — The capital stack, current. Objectives: describe the two-part purchase structure (down payment plus a rehab loan replaced at refinance); name the one skill that actually matters in card-funded capital — which terms of service prohibit real-estate use — without naming any service; compute the effective cost of borrowed capital across a refinance window, both ways. Build: price your own next down payment three ways and write the effective cost of each. You’ll know it worked when your cheapest-looking source isn’t the one you picked, and you can say why in one line. 20 min. HARD DISCLAIMER; affiliation disclosure per charter §4 rule 7; every rate on a dated panel per R-G.

Video note. Slide+narration: the capital stack diagram; the effective-cost arithmetic, on dated slides so a re-render updates one asset. Screen recording: the term-sheet automation, if and only if R-20 resolves — otherwise none, and the module says so plainly. Avatar clip (one, ~40 s, mandatory): the disclaimer plus the affiliation disclosure in one breath. Charter rule 7 requires the relationship be disclosed plainly; plainly means a person saying it.

R-18 · Private Lending & Raising Capital — P0

Where it sits. New RA8, RE-APPLIED, two units.

What promises it. Ch9, twice in one paragraph: “it’s worth a paragraph here even though the real depth of it belongs to the school” and “the school carries the rest of this at the depth it actually deserves.” The paragraph also draws the line the school inherits: “raising money from people you already have a relationship with, quietly, one deal at a time, is a very different thing under securities law than advertising publicly for lenders or pooling several people’s money into one fund.”

What exists today. Nothing — no module, no row, no ledger card. The book’s paragraph is currently the deepest treatment anywhere in the corpus.

The gap. A named, doubled pointer with nothing behind it, on the one topic in Book 2 that touches securities law. The channels the book names (investor meetups, self-directed retirement accounts, county recorder records, title and closing referral networks) are all findable and teachable; the line between a quiet single note and a pooled fund is the part that must be taught carefully and currently isn’t taught at all.

What “deeper” means. - RA8-U1 — Where the money actually is. Objectives: name four findable channels and what each one is good for; read a county recorder’s records to identify who financed a deal like yours last year; approach a lender with a specific deal rather than a general ask. Build: produce a list of ten recorded private lenders in your own county from public records. You’ll know it worked when at least three names on your list financed a deal shaped like the one you’re working. 20 min. - RA8-U2 — The line you don’t cross without a lawyer. Objectives: distinguish a single private note secured by a recorded deed of trust from a pooled offering; explain why advertising publicly changes the question; identify the exact moment in your own plan where a securities attorney gets called. Build: write your own one-page “before I do this, I call” list with the trigger for each item. 15 min. Free in full; hard disclaimer; attorney pointer in-body.

Video note. Slide+narration throughout — this is a concept module with no platform surface except the recorder search. Screen recording: the recorder search producing lender names. Avatar clip (one, ~35 s): the securities line, said plainly, ending on “before you advertise for lenders or pool anyone’s money.”

R-19 · The Living State-Law Map — P0

Where it sits. New RU9, RE-UNIVERSAL, but structurally unlike any other module: a dated, revisable reference with a lesson wrapped around it. Shared with the agent school per L-008 (“SCHOOL-Investor/Agent”).

What promises it. Three separate chapters. Ch9: “The school keeps the state-by-state version of that current; what belongs here is the shape and the caution, not the map.” Ch13: “Know your state’s current line before you scale a dispo pipeline past the size where a mistake there is cheap.” Ch18: “‘your state’s rule governs’ is not me dodging. It is the actual answer.” L-008, RULED 2026-08-24: “School treatment: the living state-by-state module, kept current, with citations, revised as legislatures move.”

What exists today. L-008’s card. One backmatter citation. No module, no map, no revision process.

The gap. The most-promised single item in the entire book and it does not exist in any form. It is also the school’s clearest membership driver under L-013 — the book teaches the permanent principle and ages gracefully; the map has a pulse and only the school can hold it. And it is by far the cheapest P0 on this list to stand up in a first version. The design constraint from the RE design’s §0 is absolute: state-specific law is a living map, never printed as fact — so the module teaches how to read your own state’s current rule and when to call an attorney, and the map itself is a dated, cited, separately-revised artifact, not narration.

What “deeper” means. - RU9-U1 — Reading your own state. Objectives: name the three activities whose licensing treatment varies most (wholesale marketing, dispo of a contract you hold, paying an unlicensed party for an introduction); find your own state’s current statute or regulator page and record what you found with the date you found it; identify the trigger that means “stop and call an attorney licensed where you buy.” Build: produce your own one-page, dated state sheet with citations. You’ll know it worked when your sheet has a date on it and a calendar reminder to re-check it. 20 min. - RU9-U2 — The stable principles underneath. Objectives: explain double closing and novation as structures, independent of any state’s treatment of them; describe the disclosure duties and cancellation windows the 2025–26 wave of statutes tends to attach; state why a principle survives a legislature and a number doesn’t. Build: write the two-sentence version of each principle in your own words. 15 min. - The map artifact. A dated, cited, per-state panel maintained on its own revision cadence, with an explicit “current as of” stamp and a standing instruction that it is education, not legal advice. Owner decision needed on revision cadence and who signs off.

Video note. Slide+narration for the principles — and the map itself must never be narrated into a video, because a rendered video is exactly the frozen artifact this module exists to avoid. Build the map as a live page the lesson links to; the video says “here is how to read it,” never “here is what your state says.” Screen recording: finding a real statute page and dating the entry. Avatar clip (one, ~30 s): “your state’s rule governs is not me dodging — it is the actual answer.” Said once, by a person, so it doesn’t read as a legal shrug.

R-20 · RP9 — Funding on Autopilot — P0 · BLOCKING FLAG

Where it sits. RP9 (“Funding on Autopilot”), marked (GATED) in the sibling charter.

What promises it. Ch10, present tense, unhedged: “Where I’ve routed that piece is through the automation in this book’s membership section: point it at the financing you’re shopping and it goes and gets current, real term-sheet data across multiple lenders at once… It’s there for any reader of this book, free, inside the membership section.” Ch12 repeats it at length. Ch17 leans on it again: “against current terms pulled from that same membership-section term-sheet automation.”

What exists today. RP9’s charter row, with its own condition attached: “Funding on Autopilot (GATED) | The CKC feeder… Ships only when the funding funnel is live + compliance/disclosure cleared.” Plus REI-EXPANSION-PRICING-AND-BUILD-LIST.md §3, which flags lending as licensed activity state-by-state, requires affiliation disclosure to members, raises RESPA-adjacent referral rules, and notes plainly: “the CKC funnel isn’t live yet (V2 engine still being polished) — the feeder ships when the funnel does.”

The gap. This is not a curriculum gap. Book 2 tells every reader, three times, that a working term-sheet automation is available to them free through membership; the school’s own charter says it cannot ship until a funnel is live and compliance is cleared. Under rule R-H — “No present-tense claim that an automation is running… or that the platform already has a feature, unless confirmed” — one of these two documents is currently wrong, and the book is the one that goes to print. Two further chapters (R-06 Magic Math, R-13 the five dials) have their currency dependency wired to this feature, so a slip here degrades two other modules from “live terms” to “frozen numbers.”

What “deeper” means. Before any lesson: an owner decision, in writing, on which of three paths. (a) The automation is live for members at print — then RP9 gets a module and the compliance workstream must close first. (b) It is not live — then Ch10, Ch12 and Ch17 need their three sentences re-cast as direction rather than availability, exactly the way Ch20 hedges the gallery shelf. (c) A narrower thing is live (rates and terms displayed, no application submitted) — then the book’s sentences shrink to match it, and RP9’s module teaches only what actually runs. Whichever path: the module, when it exists, carries affiliation disclosure per charter §4 rule 7 in-body, not in a footer.

Video note. No video until the decision lands. Recording a module for a feature that isn’t live is the exact failure Chapter Two is about.

R-21 · RP4 — The Offer Engine — P1

Where it sits. RP4 (“The Offer Engine”), the school’s named differentiator, expanded to four units (U4 is R-04’s stress unit).

What promises it. Ch8, the book’s central inversion: “you enter what’s true, and the maximum defensible offer comes out. Not a suggestion buried among three others. Not a rubric that grades a number you already typed.” And the receipt: “Every offer the system hands back tells you which rule bound it.” The sibling charter names RP4 as one of three beta-bundle modules and calls it “the differentiator no coaching program has.”

What exists today. school/offer-engine/OFFER-ENGINE-SPEC-2026-08-23.md, 1,323 words: ten design laws each traceable to a decoded defect, a complete input surface, three engines described precisely, a museum case for the four legacy heuristics, and a verification ledger with real numbers (the closed-form second lien matching the template’s 18-term series to the penny at 9,965.99; the MAO inversion reference deal at 146,263 with the return floor binding at exactly 0.150). It is excellent. It is also a design spec written for a builder, not a lesson written for a learner: no objectives, no build steps, no failure list, no glossary, no check-yourself.

The gap. The best source artifact in the corpus has never been converted. Everything a module needs is in the spec except the pedagogy — and the pedagogy is the entire job, because Ch8’s claim is not “the math is right,” it’s “you can see why every number is the way it is.” A learner who is handed the tool and not the receipt-reading skill is back in Chapter Seven, trusting a green light.

What “deeper” means. - RP4-U1 — Offers are outputs. Objectives: name every input the engine needs and where each comes from (comps from RP3, repairs from the estimator, timeline from your own market’s closings); set your floors in your own numbers, written down away from a closing table; read a computed offer and name the binding constraint out loud. Build: run one real lead and say which floor bound it. You’ll know it worked when you can point at the rule that capped the number, not just the number. 20 min. - RP4-U2 — Reading the receipt. Objectives: expand any headline figure to its source in one step; override a number, type the one-line reason, and understand what the log now holds; explain why the four legacy heuristics sit in a display case rather than a vote. Build: override one number with a real reason, then find that entry in the log a week later. 15 min. - RP4-U3 — Every constant has a why. Objectives: identify a magic constant in any tool (as-is = value minus 1.5× repairs; $1,000 of rehab equals a day) and label it a heuristic; set your own policy values with a defended reason for each; state what changes when a constant becomes an editable, named policy. Build: write your own policy sheet — every threshold, its value, and one sentence of why. You’ll know it worked when every number in your policy sheet has a reason you’d defend to a partner. 20 min. - RP4-U4 — the stress unit from R-04.

Video note. Slide+narration: the inversion in one sentence; the four legacy heuristics landing $39k apart on one house; the ten design laws as ten short cards. Screen recording (required, and this is the school’s single most important recording): facts in, offer out, receipt expanding to a comp — the inversion has to be seen once, or it stays a claim. Avatar clip (one, ~40 s): “the crime was never the arithmetic — the crime was hiding sound math behind a locked door.” The founder answering the villain, on camera, is the moment the school earns the differentiator claim.

R-22 · The Blank Cell — auditing a tool that hands you a verdict — P1

Where it sits. New RU10, RE-UNIVERSAL, one unit, free tier — an acquisition lesson by design, because it teaches a skill a non-member can perform on the tool they already own.

What promises it. Ch7: “The test takes ten minutes on any analyzer you’re evaluating. Pick the field that decides your offer — the number you’d actually say to a seller — and ask: where does this come from?” And the four buckets the chapter names as a checkable procedure.

What exists today. school/DEAL-ANALYZER-DECODED-2026-08-23.md, 2,184 words — a cell-by-cell forensic decode: ten sheets, ~10,600 formulas, 20 of 22 named ranges resolving to #REF!, the eighteen-term hidden solver duplicated three times, the vendor’s own rules doc instructing students to overload the wrong input cells. It is the best villain evidence in the corpus. It is also a forensic report about one identified vendor file.

The gap. Two problems, both solvable. First, no lesson exists — the reader gets a ten-minute test in Ch7 and nowhere to practice it. Second, and binding: Book 2’s own backmatter deliberately generalizes this material (“reflects what I’ve seen, as a paying user, across more than one program in this category”; “The worked numbers in this chapter… are mine: a constructed example”). The school module must inherit that generalization exactly — no vendor file, no document IDs, no reproduced sheet, no quoted rules doc. The decode stays internal source; the lesson teaches the test.

What “deeper” means. One unit, four objectives: run the visibility check (how much of the math happens where you can see it, and does anything discourage you from looking); run the disagreement check (ask the tool the one question it exists to answer and count how many answers you get); run the workaround check (read the tool’s own written instructions for a sentence telling you to enter a number that isn’t what the field says it is); run the trace check (pick the field that decides your offer and walk it backward — computed, or typed?). Build: run all four against a tool you actually own, on a real house, and write a one-page finding. You’ll know it worked when you can name, for your own tool, whether the number you’d say to a seller is computed or typed. 20 min.

Video note. Slide+narration: the four buckets as four checks; the four disagreeing rules of thumb on the constructed $180k/$30k house, animated to land at a $64,000 spread — the strongest single slide in the curriculum. Screen recording: a generic constructed workbook built for the lesson, never the vendor’s file. Avatar clip (one, ~45 s): “I hadn’t bought a tool that decided the offer. I’d bought a very elaborate rubric that graded whatever offer I decided on my own.” The founder’s own confession is the reason anyone trusts the rest of the school; it cannot be narrated by a clone reading a slide.

R-23 · The Cost Defense — relocate and expand — P1

Where it sits. Rules on placement: it belongs in the REI investor school, not the Book 3 agent school. It should be RA-S5 in the REI structures family (or a unit pair under RP4/RP6), taking its unit codes from the REI scheme. It is currently written into the agent school’s Track C as “C6.”

What promises it. The book already carries the teaching — the Ch8 insertion and the Ch11 insertion are both merged into the current drafts (Ch8: “Show them the stack, in order, and you’re not arguing with their number”; Ch11: “Sent doesn’t mean the number went out. It means the link did”). The unit is what teaches a learner to build it.

What exists today. book/book2/COST-DEFENSE-LESSON-2026-09-02.md, 5,499 words: the two book insertions, two finished seller-facing scripts (flip frame and rental-buyer frame, fully placeholdered), a Track C unit spec (C6-U1) with four objectives, a five-part lesson outline, a two-frame exercise with a “you’ll know it worked when” line, and five assessment questions with rationales. It is the only lesson-shaped artifact in the entire REI corpus.

The gap. Three, in order of severity. (1) Misfiled. It sits in SCHOOL-RE-CURRICULUM-DESIGN-2026-09-01.md’s Track C table — the agent school, whose Track C is Book 3’s automated-agency capstone and which the 2026-09-02 addendum explicitly caps at five modules (“No sixth module (integration doc R42). Track C stays five modules”). The cost defense is investor-side: it presents a buyer’s own cost stack to a seller, including a disclosed investor margin, which is not a licensee’s act. Filed where it is, it both violates the addendum’s ruling and lands in front of the wrong learner. (2) One unit where three are needed. The cost stack, the two frames, and the hosted-link honesty each carry a full unit’s worth of objectives; one unit at the school’s 900–1,500-word ceiling cannot hold them, and “deeper” here means splitting, not lengthening. (3) Format mismatch. It is written to the agent school’s unit format (objectives / outline / exercise / assessment questions), not the SCHOOL-MODULE-SPEC nine-section format the REI school inherits. The assessment questions are usable as “Check yourself” seeds but need to be cut to three per module.

What “deeper” means. - U1 — The stack, in order. Objectives: build the stack in the seller’s own order, value first and number last; name the project-management cushion and the contingency as two different real costs; state why a stack that changes shape from seller to seller stops being honest. Build: the existing $210,000 exercise, flip frame. You’ll know it worked when your stack lands on a number you’d say out loud and every line above it has a two-sentence reason. 20 min. - U2 — Two frames, one engine. Objectives: read what a seller has actually said about their own house and pick the frame from their words, not your plan; say what changes between frames and why changing it is honest rather than convenient; default explicitly, and mark the default as a default, when the seller has said nothing. Build: the same house, rental-buyer frame, plus the two-sentence justification. 20 min. - U3 — The link, not the number. Objectives: explain why a link is a better carrier than a call or a raw number; set an expiry and say plainly what an expiring offer is telling the truth about; state exactly what a hosted page deters, what it traces, and what it cannot prevent — and never claim more. Build: send yourself one offer link and try to break your own claim about it. You’ll know it worked when you can state the page’s protection in one sentence you’d defend to a seller. 15 min. Free in full. - Also relocate: PropFlow doctrine proposal P9 in that same file is a platform artifact, not school content, and its own format note says the P1–P8 files were unreachable — it needs re-shaping against the real proposal set before filing, and it should not travel with the school module.

Video note. Slide+narration: the stack building line by line, with the number arriving last — the animation is the pedagogy here and it is perfectly automatable. Screen recording: a real hosted offer page opening with a name and address on it, and expiring. Avatar clip (one, ~45 s): the margin paragraph — “a seller who sees margin written down, named, with a real number next to it, has nothing left to discover.” It argues against the instinct of every new investor watching, and that argument needs a face.

R-24 · The First Walk-In — judge, don’t chase — P1

Where it sits. New RU0, RE-UNIVERSAL, the school’s orientation unit, placed before RU1.

What promises it. Ch6, which coins the posture: “judge, don’t chase. Not because chasing is beneath anybody… The term names something narrower than that. It names where the center of gravity of the work moves.” And the honest cost: “You won’t get there in a morning, and you shouldn’t expect to.”

What exists today. Nothing. The general school’s U1 (U1-welcome-second-shift.md, 1,143 w) is the structural template — orientation, the reader contract, “write your why in three sentences” — but it is written to a general business owner, not an investor.

The gap. The REI school’s recommended path (RU1→RU4 → RP1→RP5 → RU5 → RP6→RP8) starts at concepts and never establishes the posture the whole path depends on. Every downstream module assumes a learner who reads a queue instead of building one, and nothing teaches that switch or warns that it takes a year of small corrections. Without it, RP1’s queue is just another list, and the first noisy week talks the learner out of the school.

What “deeper” means. One unit, four objectives: describe the swap from building a list to reading one, and why reading is judgment and building is effort; write your own before-morning honestly (where your first hour actually goes today); name three corrections you already know you’d have to teach a queue about your own market; set the promise — the list gets opened before the phone. Build: pen and paper — log one real week’s first hour, then write the three corrections. You’ll know it worked when you can name, in writing, the exact habit you’ll break on the first Monday the queue arrives. 15 min.

Video note. Slide+narration: the two mornings side by side; the posture named. Screen recording: none — this is deliberately a pen-and-paper unit, as the general school’s UNIVERSAL track allows. Avatar clip (one, ~30 s): “you won’t get there in a morning, and you shouldn’t expect to.” An orientation module that over-promises poisons everything after it, and the founder saying the honest version first is worth more than any feature slide.

R-25 · Rehab at a Distance & the Vendor Bench — P1

Where it sits. RP7 (“Rehab at a Distance”) and RA5 (“Contractor & Rehab Ops”), two charter rows that should be one three-unit module. Build-list #8 (vendor-discovery monitor) and #9 (rehab payout qualifier).

What promises it. Ch15’s bench paragraph (“whoever answers first during an emergency call charges an emergency-call price”) and its payout standard (“scheduled is not done, verified is done, and verified is what gets paid”), Ch13’s draw logic, Ch17’s maintenance qualification gate and its manuscript marker.

What exists today. Two charter rows totalling 24 words. The payout standard is stated three times in Book 2 and specified nowhere in the school.

The gap. “Verified is payable” is the single most reusable operating rule in the whole REI corpus — it appears in draws (Ch13), turns (Ch15), and maintenance spend (Ch17) — and it has no home. Its logic is currently scattered as one paragraph inside three different proposed modules, which means three drafters will write it three ways. It needs one canonical unit that the others reference.

What “deeper” means. Three units: U1 the bench (find and rank vendors by trade and region before you need one; record who is good at what for what price, once; measure what a hole in the bench costs at emergency-call prices). U2 verified is payable (match a photo against a specific scope line; clear routine matches under a cap; kick back anything ambiguous with no relationship-based benefit of the doubt — You’ll know it worked when a payout kicked back and the contractor’s next submission was better documented). U3 vendor onboarding (self-serve signup, insurance-spec request, approve gate, doc forwarding). 15–20 min each.

Video note. Slide+narration: the bench economics; the qualification decision tree. Screen recording (required): a photo clearing a line item and a second one kicking back — that contrast is the unit. Avatar clip: none needed; the bathtub-faucet-installed-upside-down story is vivid enough as narration over a still.

R-26 · Every Other Way to Buy — the deferred elective — P2

Where it sits. New RA9, RE-APPLIED elective, four short units.

What promises it. Ch9, deferring explicitly: “A handful of other ways to buy deserve a name and a sentence each, even though none of them gets the full build here… Each of those is a real path with real rules, and each one gets the depth it needs in the school rather than a rushed paragraph pretending to cover it here.” Land, notes, mobile and manufactured homes, auctions and tax liens.

What exists today. Nothing. RA4 (Land Intelligence) touches one of the four from a different angle.

The gap. Four named paths, four pointers, nothing behind them. This is a real promise — but it is the book’s own judgment that these are electives rather than spine, so it ranks below the twenty things a reader hits on their first deal.

What “deeper” means. One short unit each (12–15 min), each with the same shape: what makes this asset trade differently, what financing looks like, the one rule that loses people money fastest, and a build that produces a real candidate list in the learner’s own market. Land: use and access instead of comps. Notes: stepping into a lender’s position. Mobile/manufactured: real property versus personal property and what titling changes. Auctions and tax liens: redemption periods and the compressed clock.

Video note. Slide+narration for all four — these are definitional units and fully automatable from lesson text. Screen recording: one per unit, showing the candidate list being built. Avatar clip: none. Reserve avatar time for spine modules; an elective that opens with a founder clip signals a priority the elective doesn’t have.

R-27 · Driving for Dollars, Automated — P2

Where it sits. RP1-U5, an elective unit under the sweep.

What promises it. Ch4’s own section, including its honest limit: “Street-level imagery is not live… That makes it a candidate generator, not a verification: it tells you where to look, not what’s true today.”

What exists today. Nothing.

The gap. A whole named technique with a genuinely interesting second-pass mechanic (a tarp blue in March and still blue in June is a stronger signal than either photo alone) and an honest ceiling. It is a delightful module and it is not on anyone’s first-deal path, which is exactly why it ranks P2.

What “deeper” means. One unit, three objectives: point an image walk at a street and check it against a named signal list (roofline, boarded windows, tarps, overgrowth, posted notices, empty driveways); read a timestamped second pass as a comparison rather than a second look; state plainly that this generates candidates and never verifies them, and keep the short drive that the machine earned you. Build: run one ZIP code, then drive the ten addresses it flagged and score its accuracy. You’ll know it worked when you can say what fraction of its flags survived your own windshield. 20 min.

Video note. Slide+narration: the signal list; the March/June tarp comparison as a two-image slide, which is the whole idea in one frame. Screen recording: the walk running across a street. Avatar clip: none.

R-28 · The Exit Menu Keeps Growing — P3

Where it sits. RA4 (“Land Intelligence”), extended with an accessory-dwelling unit.

What promises it. Ch12: “a house on a big enough lot can add a second legal unit… the after-repair value isn’t just a renovated house anymore, it’s a renovated house plus a second rentable address, and the refinance math above runs on a bigger number entirely.” Ch12 itself explains why it stays a line and not a chapter: “Rules like this differ block to block and change on a legislative calendar.”

What exists today. RA4’s row: “Land Intelligence | Zoning-fringe monitoring, storage-unit lots, franchise-pad corners — parameters, not promises.” Thirteen words.

The gap. Nothing is promised to the reader beyond a line on an exit menu, and the topic is jurisdictional enough that it belongs behind R-19’s living-map discipline rather than in front of it. Build it after the map exists, so it inherits the map’s revision cadence instead of inventing a second one.

What “deeper” means. One unit, three objectives: set zoning and lot parameters as a monitor rather than a promise; describe what a ministerial approval process changes about underwriting a second unit; run the value math on a house plus a second address and say which input you’d verify first. Build: set one parameter monitor for your own county. 15 min. Routes every legal question to R-19.

Video note. Slide+narration only; no avatar. Every specific in this module has an expiration date, so keep the video generic and put the jurisdictional specifics on the live map page.

Build list → lesson mapping (all 18)

REI-EXPANSION-PRICING-AND-BUILD-LIST.md §8 lists eighteen automations. Every one now has a named lesson need, or an explicit reason it doesn’t belong to this school.

Three of the eighteen (11, 14, 15) are main-school automations with no REI lesson obligation; #18 is a platform capability. The other fourteen all land on a need above.

Ledger cards → needs

L-005 is worth one extra line: it is the smallest card on the list and the fastest win — negotiating a wholesaler’s fee onto the back end is one paragraph in Ch9 and a fifteen-minute unit in RA-S1, and it teaches a learner something they can use on their next deal without building anything.

Counts

Proposed new units across all needs: 68 (spanning 20 charter modules that exist only as table rows, plus 12 modules that do not exist in either charter). Existing REI lesson files: 0. Existing REI source words that convert into lessons rather than needing fresh research: 9,006 (the decode, the offer-engine spec, the cost-defense lesson).

Modules earning an avatar clip: 19 of 28, one clip each, 20–60 seconds, always at the honesty beat or the compliance line — never at the feature demo. Modules requiring a screen recording of the platform: 21 of 28. Modules that are pure slide+narration with no recording at all: R-07, R-08, R-12 (partly), R-18 (mostly), R-24, R-26, R-28 — which is the useful finding for production sequencing: the concept track can be produced from lesson text alone, in parallel, before a single platform screen is ready to film.

Two items block drafting rather than merely waiting on it: R-20 (the book promises a live automation the charter gates) and R-08 (an OPEN ledger item plus an unverified code cite). R-17 needs one naming ruling. Everything else can start.

If we could only build three

Build R-21 (the Offer Engine), R-02 (Feeding the Hunter), and R-19 (the Living State-Law Map) — in that order, and for three different reasons. R-21 because the spec is already written to verification-ledger depth, the tool exists, and it is the only module that answers the book’s villain: a learner who watches facts go in and a receipted offer come out has seen the school’s entire differentiator in one sitting, and the sibling charter already ranks it in the beta bundle. R-02 rather than R-01 because a sweep feeding garbage is worse than no sweep — the book says so in Ch4’s closing line — and because RP3-U3 carries the terms-of-use rule, which is the school’s largest liability surface and the one lesson that must exist before a single member points a collection at a county website. R-19 because it is the cheapest P0 on this list and the most-promised item in the entire book: three chapters send the reader to a map that does not exist, and a first version is a dated page with citations and a review cadence, not a build. It is also the honest membership driver L-013 designed the whole book-to-school hand-off around — the book teaches the principle and ages gracefully, the school holds the thing with a pulse. Those three give a member a working offer engine, clean data under it, and a current answer to the one question that can cost them their business, which is a defensible beta on its own. What they do not give is a lead queue (R-01) or a closing chain (R-09), so the fourth build is R-01 and the honest framing at launch is “the judging half, first” — the same order the book itself builds in.

How a module becomes a lesson video (the pipeline, in plain words)

You write one module the way you already write them — the nine sections, the plain-words register, the build. This turns that one file into a finished lesson: a slide deck, a video with your narration timed to the second, and captions — without anyone touching a slide tool or a video editor.

What goes in

• The module markdown. One file, school/P03-whatever.md, in the same format every module already uses. Nothing else to prepare — the script reads the nine sections straight off the page.
• Your cloned voice. Not needed to try this today (see “the one command” below), but when you’re ready: an ELEVENLABS_API_KEY and a VOICE_ID in the environment, never written into any file this script touches.
• Avatar clips — optional, and only for a few spots. The script always knows where you’d naturally show up on camera: the open, the moment it turns into the actual build, the close. Everywhere else stays slides. Drop a clip in as avatar/cold-open.mp4 (or build-step-1.mp4, or outro.mp4) and the next run uses your real footage there instead of the placeholder card. Skip it, and the placeholder just says what’s supposed to go there and for how long.

The one command

python3 lesson_to_video.py school/P03-whole-list-dynamic-variables.md --tts none

That’s it. --tts none is the default — it builds the whole package with a silent track, timed the same way real narration would time it (word count, spoken pace, a little breathing room). Good for previewing the deck and the pacing before you spend anything on voice. Swap in --tts elevenlabs once your key and voice ID are set, and the same command renders it with your actual voice and re-times every slide to match exactly how long you take to say it.

What comes out

Everything lands in video/out/<module-name>/:

How a lesson re-renders when you change the text

Every package remembers a checksum of the module file it came from. Edit the markdown — fix a sentence, add a failure mode, reword a step — and run the same command again. Nothing to clean up, nothing to remember to delete: it rebuilds the deck, the frames, the timing, and the video from scratch, and the new checksum shows the lesson is current with what you last wrote. If you haven’t touched the module, running it again just produces the same package.

The three things only you supply

Everything else here is mechanical — the segmenting, the slide design, the timing, the render. Three things aren’t, and the script doesn’t pretend otherwise:

1. The voice clone itself. The script sends it text and gets back audio; it doesn’t make the voice.
2. The avatar clips. A handful of short moments where you’re actually on camera — the script leaves the exact spot and the exact length open and waits for the file.
3. A final watch. Before anything goes in front of a learner, watch the whole thing once. The script gets the words and the timing right; it doesn’t know what a slide feels like next to the one before it, and you’re the only one who can catch that.

The course view

Once you’ve got more than one lesson built, course_to_playlist.py reads school/CURRICULUM-MAP.md — the order the school is meant to be taken in — and writes course.json and course.html. The page lists every lesson in that order; anything already rendered expands into its segment list, and every segment is a link straight into that lesson’s deck at that slide. Lessons you haven’t rendered yet just show up as not-yet-produced, in their place in the order, instead of disappearing — so the page always shows the whole shape of the school, not just the part that’s finished.

python3 course_to_playlist.py

Re-run it any time you’ve rendered another lesson; it just rebuilds the two files from whatever’s in out/ right now.

What this run actually did

Both sample lessons (P03 — Do It for the Whole List and U2 — The Show-It-Once Mindset) were built end to end with --tts none and checked: the deck opens and navigates, the frame count matches the segment count, and ffprobe confirms lesson.mp4 is H.264/1080p/30fps/AAC at the timeline’s exact runtime minus the crossfade overlaps. Numbers are in the final report handed back alongside this file.

One honest gap: the deck is styled with the school’s real palette and the specified type (Fraunces for headlines, Source Serif 4 for body), but this machine has no network access to Google Fonts and neither face is installed locally, so Chromium substitutes its own serif when it renders frames. The CSS asks for the right fonts — on a machine that has them (or with the two font files installed here), the exact same script produces the exact intended look with no code changes.

The production brief the pipeline was built to

Owner: “I don’t have videos, but we can clone my voice and create slideshows that are timelined to topics in a course so it looks and feels designed intentionally, but is 100% automated in how they are designed and created, so I just provide the cloned voice and my cloned avatar for some of clips in between.”

What to build (prototype, runnable today, in /home/claude/work/school/video/)

lesson_to_video.py — a single script: one school module markdown → a finished lesson package: 1. Segmenter: parse the module’s 9 sections (SCHOOL-MODULE-SPEC order) into a SEGMENT LIST: {id, kind, title, narration, slide, duration_s, avatar}. Kinds: cold-open (avatar, ≤20 s), concept, why, prereq, build-step (one per numbered step), verify (“You’ll know it worked when…”), sideways (one per failure), words, show, check (one per question), outro (avatar). Narration is the module’s own prose lightly turned into spoken register (contractions, second person, no headings read aloud, no URLs; numbers spelled the way you’d say them). Default avatar moments: cold-open, the pivot into “The build”, the outro — nothing else (the owner supplies avatar clips only “for some of clips in between”). 2. Slide designer: for each segment, a slide spec {layout, headline, bullets|steps|callout|term-def|question, visual-hint} chosen by kind (title card / statement / 3-bullet / numbered step with “step N of M” / two-column term-definition / question card / receipt-style callout). Design tokens shared with the review site (fonts Fraunces/Source Serif 4, the school palette P_SC in review-site/_build/make.py — copy the hex values). Render slides as one self-contained HTML deck (deck.html, keyboard-navigable, one section per slide, data-seg ids, 16:9) AND as PNG frames via Playwright/Chromium (frames/NNN.png, 1920×1080) — Chromium is at /opt/pw-browsers, do not download browsers. 3. Timing: duration_s = narration word count ÷ 2.6 words/s + 0.8 s pad (min 4 s). Emit timeline.json (segments in order with start/end) and narration.json (one entry per segment, the exact text to send to the voice clone, plus voice: "owner-clone" and ssml_hint). 4. Voice: a tts step with a pluggable backend: --tts none (default: silent track, timing from the formula), --tts elevenlabs (reads ELEVENLABS_API_KEY + VOICE_ID from env; POST per segment; writes audio/NNN.mp3; re-times segments to the real audio length via ffprobe). Never embed a key. If a backend fails, fall back to silent and say so. 5. Assembler: ffmpeg concat — each frame held for its segment duration, with the segment’s audio when present, crossfade 0.3 s; avatar segments render a placeholder card (“[avatar clip: cold open — 14 s]”) sized to the narration so the owner’s real clip drops in later by filename (avatar/<seg-id>.mp4 overrides the placeholder if present). Output lesson.mp4 (H.264, 1080p, 30 fps, AAC) + lesson.vtt captions from the narration + package.json manifest (module id, title, segments, total runtime, checksum of the source markdown so a re-run knows when the lesson changed). 6. Course timeline: course_to_playlist.py — given CURRICULUM-MAP’s learning path, emit course.json (ordered lessons with runtimes) and a course.html chapter-navigation page (lesson list → segment list, each segment a deep link into the deck at that slide) so the whole track “feels designed intentionally”.

Run it on P03-whole-list-dynamic-variables.md and U2-show-it-once-mindset.md (both in /home/claude/work/school/school/) with --tts none and produce their packages under /home/claude/work/school/video/out/<module>/. Verify lesson.mp4 exists and plays (ffprobe duration ≈ timeline total), the deck opens, frames exist. Keep everything in plain Python + Playwright + ffmpeg; pip install what you need with –break-system-packages. Write a README.md in /home/claude/work/school/video/ that explains the pipeline in the owner’s plain-words register (no vendor pitch): inputs (module markdown, cloned voice id, optional avatar clips), one command, outputs, how a lesson re-renders when its text changes, and the three things only the owner supplies (voice clone, avatar clips, a final watch). Report: what runs, runtimes, frame counts, anything that failed.

Rules: no product/internal names in any narration or slide (the module text already obeys the secrecy line; keep it). The narration must not add facts the module doesn’t contain.