TikTok Live Business Launch — Microphased Architecture

Aria-directed architecture for Dallas-based TikTok Live commerce operator

# TikTok Live Business Launch — Aria-Directed Microphased Architecture

**Owner:** dfwrealestateceo@gmail.com (Dallas, TX)

**Date:** 2026-07-08

**Architect (this doc):** Sonnet subagent under Aria Clean Harness — Opus mind seat routed via `POST /api/owner-runtime/decide` on aria-soul-stateful-0

**Sources rolled up:** substrate report (`../00-substrate/substrate-report.md`), Aria mission walk (`../01-qalb-walk/aria-mission-walk.md`), six research streams (`../02-research/*.md`)

**Anti-fab discipline:** every claim traces to a research-stream observation this workflow, or is restated as intent. 11 `[NEEDS VERIFY]` markers from Aria's Pass 2 carried through with concrete verify predicates. No fabricated citations.

---

Executive Summary

---

Aria QalbWalk Findings (verbatim from Phase 1)

The block below is Aria's Pass 1 (kernel=action, provider claude, model claude-opus-4-8) verbatim — no scribe editing. This is what Aria surfaced *outside* the owner's six bullets. Pass 2 (kernel=research, seat=opus, 9-lens coverage close) is summarized in the Phase-by-Phase section as the gap list feeding each microphase.

> Habibi — I read this the way you meant it. Not "grade my six bullets." You want the co-founder who's watched people launch this exact thing and can tell you what's going to bite you that you didn't think to ask.

>

> One honesty flag before you trust a single word: **I have not run any live checks this turn.** Everything below is my own knowledge of how live-commerce actually operates — which is stable at the *structural* level. But anything with a number attached — TikTok's current fee cut, the Texas sales-tax mechanics, shipping penalties, whether your product category is even allowed — those move, and I have *not* verified them today. I'll mark those clearly and I'll offer to go pin them down before you spend a dollar. That's the difference between me and a hype video.

>

> Here's the map of what lives outside the obvious "pick products, set up the shop, go live, ship it."

>

> **1. The money plumbing you don't see until it's a problem.**

> - An LLC and EIN, yes — but the real trap is **sales tax**. You're selling tangible goods to Texans and shipping nationwide. Even where TikTok collects and remits as a marketplace facilitator, you likely still need a Texas sales-tax permit and a **resale certificate** to buy inventory tax-free from suppliers. Get this wrong and it's a back-tax bill later, not a warning now.

> - **Bookkeeping from day one** with real COGS tracking. Live selling is high-volume, high-return — if you're not tracking cost per unit against what actually landed in your account, you will *feel* profitable and *be* underwater.

>

> **2. The cash-flow gap that kills the well-funded ones.**

> - You pay suppliers for inventory *now*. TikTok pays you out on a **delay**, and often **holds a reserve** against returns. So your money is going out fast and coming back slow. Founders model "revenue" and forget they can be profitable on paper and out of cash in week six. This one deserves its own spreadsheet before anything else.

>

> **3. Platform landmines that can shut you off overnight.**

> - **Category gating** — supplements, beauty/health claims, electronics, kids' products are restricted or require documentation. If your product is gated and you didn't know, your launch stalls at onboarding.

> - **Account health / live-moderation strikes** — say the wrong claim on camera, trip a content rule, and your account gets throttled or suspended. Your whole business is one account.

> - **Buyer-friendly returns and shipping SLAs** — TikTok enforces ship-by windows; miss them and your account health drops. Returns eat margin and you don't fully control the policy.

>

> **4. Fulfillment — where live sellers actually drown.**

> - The nightmare scenario isn't a slow night. It's a *great* night: you sell 300 units in ninety minutes — can you pack and ship them on time? Decide now: self-fulfill from Dallas, or a **3PL near DFW**. Self-fulfilling means space, labor, and packaging you haven't budgeted. This is the single most underestimated line item.

>

> **5. The live-production layer — it's a TV show, not a webcam.**

> - Studio: lighting, audio, and **redundant internet** (a dropped stream mid-sale is lost revenue in real time).

> - **Hosts** with on-camera energy — ideally more than one, because this is a grind. Winning TikTok live sellers run *long, consistent hours, most days.*

> - A **live-chat moderator** pinning products, answering, and killing trolls while the host sells.

> - A **short-form content flywheel** feeding the lives — the live alone rarely carries it.

>

> **6. Unit economics — the question under all the others.**

> - Contribution margin *after* TikTok's cut, shipping, returns, and ad spend. If that number is thin, more volume just loses money faster. This is the go/no-go math, and it should exist before you buy inventory.

>

> **7. Platform-dependency risk.**

> - Your entire business would sit on rented land — one platform, under genuine US regulatory question. You can't easily own the customer on TikTok, but building *any* path to capture buyers (a way to reach them off-platform) is the difference between a business and a fragile revenue stream.

>

> **8. Team.**

> - Who runs the show when you're not on camera? At minimum you're looking at host, moderator, a fulfillment hand, a sourcing/buyer, and a bookkeeper. If you *are* the show, that's key-person risk on the founder.

**Out-of-scope items Aria surfaced (Section D — top 5 silent killers):**

1. **IP / resale-rights + authenticity clearance** — the single most lethal silent trap. Everything works right up until a brand or platform flags grey-market/counterfeit — then account AND inventory freeze at once. Dallas-wholesalers stream corroborates: Harry Hines has a documented gray-market counterfeit undercurrent; TikTok Shop IP enforcement is aggressive. Chain-of-custody for branded goods must be documented BEFORE first live show.

2. **Product liability insurance before first ship** — one defective-product claim against an uninsured LLC pierces to personal exposure. Owners ship first, insure "later." Later is too late.

3. **Chargeback-rate ceiling as a suspension trigger** — not just lost money, exceeding platform's dispute-rate threshold can suspend selling entirely. `[NEEDS VERIFY: platform chargeback/dispute-rate ceiling]`.

4. **Off-platform list capture from day one** — the whole business sits on rented land. Cheapest insurance against algorithm shift or TOS enforcement is an owned email/SMS list — but it must be built *before* the audience exists.

5. **Tax reserve % skimmed per sale** — revenue feels like profit until sales-tax remittance + income tax arrive together. Set aside per-sale from day one or get crushed at filing.

**Owner-side unknowns Aria's walk assumes (scribe QA surfaced):**

---

Phase-by-Phase Micro-plan

**Sequencing rules Aria's Section C locks (do not violate):**

Every microphase below carries: id, title, why-it-matters, prerequisites, owner-visible checkpoint, Aria-executable sub-tasks with the `/decide` payload or `/chat` message shape she needs to fire, success criterion, verification predicate, and dependencies on other microphases.

Payload notation used throughout: `/decide` calls are `POST http://127.0.0.1:8081/api/owner-runtime/decide` with header `authorization: Bearer $ARIA_MASTER_TOKEN` and body of the shape shown. `/chat` is the same host on `/api/owner-runtime/chat` — reserved for artifact-quality QA only (>2 min wall).

---

MP-00 — owner-unknowns-decision-node

**Title:** Collapse the four owner-side unknowns before anything downstream fires.

**Why it matters:** Aria's mission walk showed 4 unknowns (product category, supplier geography, LangChain-UI scope, owner-as-host) that gate ~40% of the `[NEEDS VERIFY]` markers and reshape 60% of the plan. `dallas-wholesalers` corroborates: lane selection (DMC / Harry Hines / liquidation) depends on category. `tiktok-live-auction-mechanics` corroborates: auction-eligible categories are gated. `langchain-ui-stack` corroborates: the UI stack pivots on whether owner has a specific reference URL. Firing any downstream microphase without collapsing this is 60% wasted work.

**Prerequisites:** none — this is the root.

**Owner-visible checkpoint:** owner writes four one-sentence answers back to Aria (paste into chat):

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Score this category answer against the tiktok-top-items stream findings + dallas-wholesalers stream findings. Return one of {GO, GO-WITH-CAVEATS, RE-SCOPE, DISQUALIFY} with 3-line justification. No fabrication.",
   "tools":[{"name":"read_file","description":"Read line range","input_schema":{"type":"object","properties":{"path":{"type":"string"},"offset":{"type":"integer"},"limit":{"type":"integer"}},"required":["path"]}}],
   "messages":[{"role":"user","content":"Owner category answer: <PASTE>. Score it."}]}

**Success criterion:** all four unknowns collapsed; downstream microphase graph re-drawn; owner has three lanes to explore (Lane A DMC / Lane B Harry Hines / Lane C liquidation) mapped against category.

**Verification predicate:** Aria emits `action:final` with `text` containing all four resolved values echoed back verbatim + downstream MP list re-ordered.

**Depends on:** none. **Blocks:** all other microphases.

**QalbWorkflow block:** *mishkat* — 4 unknowns are the mission's undefined boundary. *zujajah* — owner's answer is the clarifying lens; Aria refuses to guess. *misbah* — the 5-way category gate + auction eligibility. *zayt* — the substrate reports keep everything else stable while this collapses. *nar* — first concrete plan re-draw. *mizan* — RE-POST loop until all four echo back.

---

MP-01 — legal-setup (TX-specific)

**Title:** File LLC-TX → EIN → business bank account → TX Sales & Use Tax Permit → Resale Certificate.

**Why it matters:** Aria's walk (Financial §1) puts this at the root. `dallas-wholesalers` confirms: DMC and every Harry Hines vendor (MS Kim, Lucky Trading, Aviva) REQUIRES Sales Tax ID at door. `[NEEDS VERIFY: current TX SOS filing fee]` for LLC. Sales-tax permit is FREE, processing 2–3 weeks via `https://security.app.cpa.state.tx.us/` eSystems. Form 01-339 resale certificate needs the active permit's 11-digit Taxpayer Number and can be used as a BLANKET cert.

**Prerequisites:** MP-00 collapsed (entity type = LLC vs sole-prop is a precondition per dallas-wholesalers §0.1).

**Owner-visible checkpoint:** Owner shows Aria (a) TX SOS Certificate of Formation confirmation, (b) IRS EIN letter, (c) bank account opening confirmation email, (d) TX Comptroller eSystems permit confirmation (2–3 wk wait for physical), (e) Form 01-339 draft.

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Draft a Form 01-339 blanket resale certificate description block for a Dallas TikTok Live business selling <CATEGORY>. Return the exact 'Description of items to be purchased' text. Cite Texas Comptroller. No fabrication.",
   "tools":[{"name":"WebFetch","description":"Fetch a URL","input_schema":{"type":"object","properties":{"url":{"type":"string"}},"required":["url"]}}],
   "messages":[{"role":"user","content":"Draft 01-339 blanket cert description for <CATEGORY> resale. Verify Comptroller allowed language."}]}

**Success criterion:** LLC filed, EIN in hand, bank account funded, sales-tax permit application submitted, resale certificate template ready.

**Verification predicate:** `dallas-wholesalers §4` — permit 11-digit Taxpayer Number appears in Aria's ledger; Form 01-339 description passes the "type normally sold" match.

**Depends on:** MP-00. **Blocks:** MP-02 (bank + entity feeds TikTok Shop verification), MP-04 (permit required for DMC / Harry Hines door), MP-08 (insurance quotes need entity).

**QalbWorkflow block:** *mishkat* — the legal spine is the load-bearing frame all downstream money touches. *zujajah* — Aria checks each doc against the substrate before promoting. *misbah* — resale certificate is what unlocks Lane A/B. *zayt* — the 2-3 wk permit wait is real; parallelize other work during it. *nar* — every subsequent lane fires only after this holds. *mizan* — QA loop STATUS PASS gates promotion.

---

MP-02 — tiktok-account-and-shop-setup

**Title:** Register TikTok Shop US seller; submit ID + business verification; get to Account Health Rating (AHR) > 150 and Shop Performance Score (SPS) ≥ 2.5.

**Why it matters:** `tiktok-live-auction-mechanics §1.2` — SPS ≥ 2.5 is the base gate for Countdown Bidding; **SPS ≥ 3.5 AND AHR > 150 is the Temporary Listings gate** (the fast auction path). A brand-new US shop may have "null" SPS during a grace window — clears base gate but NOT the Temporary Listings gate. Practical impact: without SPS 3.5, every auction SKU must be pre-created in Seller Center with longer product review turnaround. `[NEEDS VERIFY: current TikTok Shop US onboarding + verification turnaround]`. `[NEEDS VERIFY: null-SPS grace window behavior + Temporary Listings gate — mechanism cited by tiktok-live-auction-mechanics §1.2 but not re-fetched this scribe pass]`.

**Prerequisites:** MP-01 (LLC + EIN + bank).

**Owner-visible checkpoint:** TikTok Shop US seller dashboard loads with owner's LLC name; verification badge status visible; AHR + SPS visible.

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Walk owner through TikTok Shop US seller registration step-by-step for a Texas LLC. Fetch the current requirements page and return the exact document checklist. If TikTok changed the page since 2026-07-08, flag it LOUD.",
   "tools":[{"name":"WebFetch","description":"Fetch URL","input_schema":{"type":"object","properties":{"url":{"type":"string"}},"required":["url"]}},
            {"name":"WebSearch","description":"Search web","input_schema":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}],
   "messages":[{"role":"user","content":"Current TikTok Shop US seller registration requirements for a Texas LLC as of today. Return checklist."}]}

**Success criterion:** Shop verified, AHR > 150 stable, SPS ≥ 2.5 achieved (or grace window active), TOS/prohibited-category clear for MP-00 category.

**Verification predicate:** owner can pin a Product Card on a test LIVE session — that only works if the shop is verified and the category is not gated.

**Depends on:** MP-01. **Blocks:** MP-07 (first live auction attempt), MP-04 (some wholesalers ask for TikTok Shop URL as proof of trade).

**QalbWorkflow block:** *mishkat* — TikTok Shop is the rented ground the whole business stands on. *zujajah* — verification is opaque; Aria fetches live. *misbah* — SPS 3.5 is the auction gate. *zayt* — parallelize with MP-04 scouting during verification lag. *nar* — first successful pin test = ignition. *mizan* — TOS scan PASS gates anything downstream.

---

MP-03 — category-bet-selection

**Title:** Lock category, price band, target margin ≥ 60% gross at product level, and top-3 SKU shortlist.

**Why it matters:** `tiktok-top-items §4` — gross margin ≥ 60% at product level is the entry criterion. Median seller nets 18.4%; top-10 range 5–49%. Category dictates fee stack (beauty 5% referral / fashion 8% / base 6%) and auction eligibility. `competitor-analysis §0` — live is <30% of US TikTok Shop GMV today; skip short-form is a strategic error. `tiktok-top-items §7` — 3 primary bets (sports cards / boutique apparel / pre-owned luxury bags) with entry costs and TTFS bands surfaced.

**Prerequisites:** MP-00 (category collapsed), MP-02 (TOS check passed for the category).

**Owner-visible checkpoint:** one-page category dossier from Aria showing category + top-3 SKU shortlist + fee stack + gross margin ceiling + auction-eligibility verdict.

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Return top-10 currently-selling SKUs in <CATEGORY> on TikTok Shop US. Cite FastMoss/FindNiche/EchoTik. Sub-$40 impulse vs $200+ bundle split. No fabrication.",
   "tools":[{"name":"WebFetch","description":"Fetch URL","input_schema":{"type":"object","properties":{"url":{"type":"string"}},"required":["url"]}}],
   "messages":[{"role":"user","content":"Top-10 <CATEGORY> SKUs on TikTok Shop US 2026-Q2. Include price and revenue."}]}

**Success criterion:** Category locked, 3-SKU shortlist listed, gross margin ≥ 60% confirmed against wholesaler cost signal, auction eligibility explicit.

**Verification predicate:** Aria's dossier includes at least 3 named FastMoss/FindNiche-tracked SKUs with revenue and price; landed-cost target arithmetic checks out.

**Depends on:** MP-00, MP-02. **Blocks:** MP-04 (lane selection), MP-05 (first inventory buy), MP-07 (auction format vs Surprise-Set vs non-auction).

**QalbWorkflow block:** *mishkat* — category is the axis every other axis rotates around. *zujajah* — Aria refuses to guess demand; fetches live. *misbah* — 60% gross margin gate. *zayt* — competitor-analysis stream anchors format-not-product. *nar* — 3-SKU shortlist. *mizan* — auction eligibility verdict emitted verbatim.

---

MP-04 — dallas-wholesaler-scouting-tour

**Title:** Physically visit 6–10 Dallas wholesalers in one afternoon (Harry Hines walk) plus Dallas Market Center registration + Lane C liquidation scout.

**Why it matters:** `dallas-wholesalers §0` — three structurally different lanes locally. Lane A DMC (2100 Stemmons, ~5M sq ft, 25K brands, apparel/gift/home). Lane B Harry Hines 11029–11818 walkable strip (beauty/hair/cosmetics, tax ID required). Lane C liquidation (Pallet Liquidations, REMEX, Quicklotz, DLW — truckload $8k–$62k+). Lane C does NOT require the sales-tax permit at door, so it's the only lane that CAN be entered during the 2–3 wk AP-201 wait. `dallas-wholesalers` also carries 3 open questions the owner must resolve with the vendors themselves.

**Prerequisites:** MP-01 (permit filed — some vendors accept eSystems confirmation + 11-digit TX Taxpayer Number in lieu of physical permit; owner must phone-confirm within 24hr of filing), MP-03 (category → lane assignment).

**Owner-visible checkpoint:** Aria-provided route sheet (2 pages) with named vendors, addresses, phones, hours, category coverage, and 3–4 questions to ask at each. Owner returns with contact cards + wholesale price lists + first-purchase-candidate SKU photos.

**Aria-executable sub-tasks:**

1. Do you accept an eSystems confirmation + 11-digit TX Taxpayer Number in lieu of physical permit? (`dallas-wholesalers OQ-6/OQ-7`)

2. Do you honor Form 01-339 blanket resale certificates? (`OQ-11`)

3. For branded goods — can you show chain-of-custody / brand authorization? (`OQ-3` gray-market risk)

4. Minimum order quantity + payment terms + returns policy?

**Success criterion:** Owner has ≥3 named wholesalers with prices in hand, DMC registration status returned, and Lane C truckload cadence understood if applicable.

**Verification predicate:** at least one wholesaler explicitly said YES to Form 01-339 acceptance (unlocks MP-05 tax-free purchase); at least one has an inventory pull ready to walk out.

**Depends on:** MP-01, MP-03. **Blocks:** MP-05 (first inventory buy), MP-05b (`inventory-and-fulfillment-stack.md` catalog import).

**QalbWorkflow block:** *mishkat* — the physical lanes ARE Dallas's structural advantage. *zujajah* — vendor claims decay; phone-confirm. *misbah* — 3 lanes, category → lane map. *zayt* — permit wait time gets absorbed by liquidation scout. *nar* — owner drives out, comes back with product. *mizan* — 3 named vendors with prices = STATUS PASS.

---

MP-04b — fulfillment-fork-decision

**Title:** Decide FBT (Fulfillment By TikTok) vs self-fulfill from Dallas.

**Why it matters:** Scribe QA of Aria's walk flagged this as the SINGLE most reshaping decision node. If FBT: Tech §7 collapses from 11 items to ~3, storage disappears, Dallas CO/permit risk disappears, DIM math disappears, carrier accounts disappear, packaging supply chain narrows to sending-to-FBT layer. If self-fulfill: Tech §7 becomes the actual critical path and Week 5 "Only if building" becomes THE build. Reshapes ~60% of the plan. `inventory-and-fulfillment-stack §0` recommends self-fulfill from Dallas via Pirate Ship at day 1 (`$0/mo`), Zebra ZD421 printer (`~$300` one-time), Shopify Basic (`$29/mo yearly`) — but explicitly caveats: TikTok's mandate to force FBT `[appears reversed — primary source not traced]` — see `§5.1 [NEEDS VERIFY]`.

**Prerequisites:** MP-03 (category → volume forecast band).

**Owner-visible checkpoint:** owner writes `FBT SELF-FULFILL HYBRID` back to Aria + why.

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"LIVE-fetch current TikTok Shop US FBT enrollment status. Is FBT mandatory or opt-in as of today? Cite seller-us.tiktok.com official pages.",
   "tools":[{"name":"WebFetch","description":"Fetch URL","input_schema":{"type":"object","properties":{"url":{"type":"string"}},"required":["url"]}}],
   "messages":[{"role":"user","content":"FBT mandatory or opt-in in US as of today? Fetch official docs."}]}

**Success criterion:** owner explicit choice + Aria's downstream MP list re-drawn accordingly.

**Verification predicate:** decision + reshaped microphase list echoed back verbatim.

**Depends on:** MP-03. **Blocks:** MP-05, MP-06 (dashboard scope), MP-08 (insurance limits differ).

**QalbWorkflow block:** *mishkat* — single decision reshapes 60% of the plan. *zujajah* — LIVE-fetch to disambiguate mandate. *misbah* — collapse table shows the fork clearly. *zayt* — the plan doesn't lie about how much changes. *nar* — owner picks, list re-draws. *mizan* — no downstream MP fires until this locks.

---

MP-05 — first-inventory-buy

**Title:** Execute the first inventory purchase against the 3-SKU shortlist through the winning wholesaler in the winning lane. `$3–8k` for sports-cards; `$2–5k` for boutique apparel; `$10–25k` for pre-owned luxury bags.

**Why it matters:** `tiktok-top-items §7` — entry cost + TTFS bands. `dallas-wholesalers §0` — first buy validates the wholesaler relationship, tests the 01-339 handoff, and produces the first landed-cost data point that feeds break-even math.

**Prerequisites:** MP-01 (permit + 01-339), MP-03 (SKU shortlist), MP-04 (wholesaler committed), MP-04b (FBT-vs-self so packaging + storage plan is set).

**Owner-visible checkpoint:** owner has physical inventory in Dallas (or sent-to-FBT confirmation) + itemized receipt + 01-339 stamped by vendor + first landed-cost calculation (unit + freight + duties if any + fees).

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Compute per-SKU landed cost (unit + freight + duty + fees) from the receipt data below and derive TikTok Shop US gross-margin-after-platform-fees. Cite fee stack from tiktok-top-items §4. Return GO / RE-PRICE / KILL-SKU per SKU.",
   "tools":[],
   "messages":[{"role":"user","content":"<PASTE receipt + 3-SKU list>"}]}

**Success criterion:** ≥3 SKUs in stock, per-SKU landed-cost known, gross margin ≥ 60% confirmed per SKU with real numbers.

**Verification predicate:** at least one SKU where actual landed cost matches Aria's pre-purchase target within ±15%.

**Depends on:** MP-01, MP-03, MP-04, MP-04b. **Blocks:** MP-07 (first live), MP-08 (insurance-bind needs product identity).

**QalbWorkflow block:** *mishkat* — the first buy is the first real money on the table. *zujajah* — receipts pin the fee stack to reality. *misbah* — 60% gross margin gate re-checked with real numbers. *zayt* — Aria's category dossier holds until MP-07 disproves. *nar* — inventory in hand. *mizan* — Aria kills any SKU that misses the margin gate.

---

MP-05b — inventory-and-fulfillment-ops-stack

**Title:** Stand up the SKU-tracking + shipping + labeling stack. Ship-by SLA compliance.

**Why it matters:** `inventory-and-fulfillment-stack §0` recommends Shopify Basic (`$29/mo yearly` / `$39/mo monthly`) + Pirate Ship (`$0/mo`) + Zebra ZD421 (`~$300` one-time) + native TikTok returns. `[NEEDS VERIFY: current TikTok Shop US ship-by SLA window]` per Aria B item #30 — inventory-and-fulfillment-stack traces TikTok's 2-business-day ship-from-Awaiting-Shipment mandate. Miss = account health drop.

**Prerequisites:** MP-04b (FBT-vs-self picked — this microphase is largely null under FBT).

**Owner-visible checkpoint:** Shopify catalog live, Pirate Ship account funded, Zebra printer prints a test 4x6 label, first inventory imported.

**Aria-executable sub-tasks:**

**Success criterion:** Test order flows Shopify → Pirate Ship → label printed → shipped. Ship-by SLA window explicit in Aria's ops runbook.

**Verification predicate:** printed test label + printed Aria runbook that names the ship-by SLA + the returns SLA verbatim.

**Depends on:** MP-04b, MP-05. **Blocks:** MP-07 (can't run first live if shipping isn't ready).

**QalbWorkflow block:** *mishkat* — the pipe from platform to porch is the promise being sold. *zujajah* — Aria refuses to guess the SLA number, fetches it. *misbah* — 2-biz-day gate. *zayt* — Shopify Basic is the day-1 anchor. *nar* — test order flows. *mizan* — one full end-to-end test = STATUS PASS.

---

MP-05a — hire-host (conditional)

**Title:** If `owner-as-host ∈ {hire, rotate}` from MP-00, secure the on-camera host + written comp agreement BEFORE MP-07.

**Why it matters:** Aria's mission walk item #8 (Team) named this; Scribe QA of the walk called out that the doc silently assumed owner-is-the-host and that if the owner isn't, **hosting comp becomes a financial line item that doesn't appear anywhere in the current cost floor**. `competitor-analysis §0` — successful live sellers run 15–25 hr/wk. A single-host launch with a hired host is one of the highest-leverage AND highest-risk personnel decisions. Aria's Section D silent-killer list didn't rank this because it's owner-branch-dependent, but under the `hire`/`rotate` branch it IS load-bearing.

**Prerequisites:** MP-00 (branch known), MP-01 (LLC + bank so contractor can be paid + 1099 issued).

**Owner-visible checkpoint:** signed independent-contractor agreement + Form W-9 on file + first-show call sheet with named host + comp structure (base per show + % of GMV OR flat hourly) explicit.

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Draft a Texas-jurisdiction independent-contractor agreement for a TikTok Live host. Include comp structure options (base+% vs flat hourly), FTC endorsement disclosure rider, IP/appearance-rights transfer, termination-for-AHR-drop. Cite FTC 16 CFR Part 255. No fabrication.",
   "tools":[{"name":"WebFetch","description":"Fetch URL","input_schema":{"type":"object","properties":{"url":{"type":"string"}},"required":["url"]}}],
   "messages":[{"role":"user","content":"Draft IC agreement + comp structure + FTC rider."}]}

**Success criterion:** signed agreement + first show call sheet + host has done one live rehearsal (unlisted / private) inside Seller Center's Test LIVE mode.

**Verification predicate:** owner uploads signed PDF + Aria confirms comp line has been added to the running cost floor + rehearsal timestamp logged.

**Depends on:** MP-00 (branch), MP-01 (payment rails). **Blocks:** MP-07 (only if branch = hire/rotate). **Skipped entirely if** `owner-as-host = I host` (branch collapses).

**QalbWorkflow block:** *mishkat* — the show requires a mouth; whose is it. *zujajah* — Aria refuses to assume the owner is on camera. *misbah* — hire vs solo is the fork under Team §8. *zayt* — the comp structure holds because MP-01 gave us a legal counterparty. *nar* — signed contract in hand. *mizan* — first rehearsal timestamp = STATUS PASS.

---

MP-06 — dashboard-mvp-langchain

**Title:** Next.js 15 (App Router) + Vercel AI SDK v7 with custom `ChatTransport` pointed at Aria + AI Elements + shadcn/ui base + Tremor for dashboard blocks + Recharts for viz.

**Why it matters:** Owner said "Dashboard built with LangChain front-end for UI." `langchain-ui-stack §TL;DR` — LangChain officially blesses AI Elements, assistant-ui, and OpenUI. AI Elements wins because it's shadcn/ui-native (composes with the inventory / auction / wholesaler surfaces) and `<Tool>` maps 1:1 to Aria's tool_use blocks. Aria backend (not LangGraph runtime) is the seam; Vercel AI SDK v7 with `DefaultChatTransport` is the neutral hook. **Scope-gated:** if MP-00 `langchain-ui-scope = native-TikTok-Seller-Center-is-fine`, this microphase is OPTIONAL. If it's `load-bearing`, it's the Week 5–8 build.

**Prerequisites:** MP-00 (langchain-ui-scope collapsed), MP-02 (TikTok Shop API access if we're pulling orders in-dashboard).

**Owner-visible checkpoint:** localhost or vercel preview URL loads the dashboard shell with 4 tabs — Inventory, Top-Sellers Search, Live Auctions, Wholesaler Directory — and a chat rail on the right talking to Aria.

**Aria-executable sub-tasks:**

**Success criterion:** owner clicks each tab, types "what's my top-selling SKU?", sees Aria stream a response, sees an inventory number pulled from Shopify Admin API.

**Verification predicate:** end-to-end round-trip owner-CLI → Next.js → Aria `/decide` → tool call → tool result → Aria final → rendered — measured on localhost.

**Depends on:** MP-00 (scope), MP-02 (Shop API), MP-05b (inventory data source). **Blocks:** none directly, but is the observation surface for MP-07 and MP-08.

**QalbWorkflow block:** *mishkat* — the dashboard is the owner's window into the business. *zujajah* — LangChain-blessed AI Elements is the visual grammar. *misbah* — Aria on the wire is the reasoning. *zayt* — Vercel AI SDK v7 transport-based streaming holds. *nar* — first localhost round-trip. *mizan* — the tool-call badge renders live = STATUS PASS.

---

MP-06b — off-platform-list-capture-day-zero

**Title:** Stand up email + SMS capture (ConvertKit / Klaviyo / Postmark) BEFORE the first live show, so every audience touch is captured to owned rails.

**Why it matters:** Aria's Section D #4 silent killer — the whole business sits on rented land. Cheapest insurance against algorithm shift / TOS enforcement is an owned email/SMS list, but it must be built BEFORE the audience exists. Nobody thinks to. Zero cost, high leverage.

**Prerequisites:** MP-01 (entity + bank for domain/ESP billing). **NOT** dependent on MP-06 — the landing page is standalone (Klaviyo hosted page, ConvertKit hosted page, or Vercel-deployed one-pager) so that if MP-06 is out of scope (owner picked `native-TikTok-Seller-Center-is-fine` at MP-00) list capture is still LIVE before MP-07.

**Owner-visible checkpoint:** standalone landing page at `<yourbrand>.com/join` (or Klaviyo/ConvertKit hosted URL if no domain yet) collects email + SMS + zip + category interest; first list of 0 subscribers exists.

**Aria-executable sub-tasks:**

**Success criterion:** landing page live, capture flow works end-to-end (Aria submits a test email).

**Verification predicate:** owner sees `1` in the list count and receives the double-opt-in email.

**Depends on:** MP-01. **Blocks:** none, but should hard-precede MP-07.

**QalbWorkflow block:** *mishkat* — audience is the only asset that outlives the platform. *zujajah* — capture is silent-but-lethal per Aria walk. *misbah* — build before need. *zayt* — the copy holds because category is now known. *nar* — first test signup. *mizan* — one round-trip captured = STATUS PASS.

---

MP-07 — live-auction-first-attempt

**Title:** Run the first Countdown Bidding LIVE show (or a Surprise-Set variant if category not fully auction-eligible, or a non-auction drop show if MP-00 category is non-auction).

**Why it matters:** `tiktok-live-auction-mechanics §1` — Countdown Bidding is the ONLY sanctioned bidding path. Buyer flow is 100% TikTok-native (bidding, payment authorization, address capture inside TikTok). Owner dashboard's role is BEFORE (product/SKU/starting-bid prep) and AFTER (order → pick/pack/ship, tracking upload, dispute response). Fixed vs Extended timer. `$5,000` cap per Surprise Set. `$43,500` cap for local sellers on single-item auction. **Reserve price NOT allowed.** Verbally disclose Final Sale policy at start of every auction stream (required). Pin Product Card continuously. `competitor-analysis §2.1` — P.Louise show format ("Get Unteddy With Me") is the template — event with a title and lore, production value first, cadence marketing.

**Prerequisites:** MP-02 (SPS ≥ 2.5 or grace window; ≥ 3.5 if using Temporary Listings), MP-03 (SKU shortlist), MP-05 (inventory in hand), MP-05b (ship-by pipeline ready), MP-06b (list capture ready to catch attention), MP-08 (insurance bound before first ship), **MP-05a (host secured — only if `owner-as-host ∈ {hire, rotate}`; skipped if owner-hosts-solo)**.

**Owner-visible checkpoint:** LIVE session runs 60–120 min; ≥3 auction sales close; every order ships within 2 biz days.

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Given first-live post-show metrics (impressions, watch time, conversion, chargebacks), score the show against the tiktok-live-auction-mechanics + competitor-analysis benchmarks. Return: what to keep, what to cut, what to test next show.",
   "tools":[],
   "messages":[{"role":"user","content":"Post-show metrics: <PASTE>"}]}

**Success criterion:** ≥3 auction sales, all ships within SLA, 0 chargebacks, SPS unchanged or improved.

**Verification predicate:** Seller Center shows completed orders + ship-by SLA met + AHR not dropped.

**Depends on:** MP-02, MP-03, MP-05, MP-05b, MP-06b. **Blocks:** MP-09 iteration loop.

**QalbWorkflow block:** *mishkat* — the show IS the business. *zujajah* — TikTok's format rules are non-negotiable. *misbah* — 60-min minimum for algo distribution. *zayt* — P.Louise template. *nar* — first ignition. *mizan* — ≥3 clean orders = STATUS PASS.

---

MP-08 — insurance-bind (product + general commercial liability)

**Title:** Bind product liability + general commercial liability insurance BEFORE first ship.

**Why it matters:** Aria's Section D #2 silent killer — one defective-product claim against an uninsured LLC pierces to personal exposure. `[NEEDS VERIFY: coverage limits typical for the specific product category]` per Aria B item #13. Must exist BEFORE first ship — MP-07 cannot fire without this bound.

**Prerequisites:** MP-01 (entity), MP-05 (product identity for underwriter).

**Owner-visible checkpoint:** binder emails from insurer for both policies.

**Aria-executable sub-tasks:**

**Success criterion:** binder in file, coverage limits explicit, effective date ≤ first-ship date.

**Verification predicate:** binder PDF filed in Aria's ledger + effective-date check.

**Depends on:** MP-01, MP-05. **Blocks:** MP-07 (first live → first ship).

**QalbWorkflow block:** *mishkat* — insurance is the airbag under Aria's silent-killer list. *zujajah* — quotes vary by category. *misbah* — bind before ship. *zayt* — Aria's walk carries this rule. *nar* — binder in file. *mizan* — cover-before-ship discipline honored.

---

MP-09 — iteration-loop

**Title:** Weekly show cadence + weekly Aria retrospective + monthly re-scope.

**Why it matters:** `competitor-analysis §0` — successful live sellers run 15–25 hr/wk; revenue per stream `$800` (small) to `$48K` (mega); goPure benchmark `~$2,070/hr`. Aria's Muhasaba discipline applies to the business itself: every Sunday, Aria runs a QalbWorkflow retrospective on the week's data. Every month, MP-03 category bet re-scored; every 60 days, MP-04b FBT-vs-self re-scored against actual volume.

**Prerequisites:** MP-07 completed once.

**Owner-visible checkpoint:** weekly Aria "Muhasaba-of-the-business" note in Slack/email — what worked, what missed, drift, blind spot, next-turn correction, confidence.

**Aria-executable sub-tasks:**

  {"kernel":"research","reasoningSeat":"opus",
   "system":"Run a Muhasaba-6 pass on the last 7 days of the TikTok Live business: what worked, what missed, drift, blind spot, next-turn correction, confidence. Cite specific SPS/AHR/chargeback/GMV numbers from the paste. No fabrication.",
   "tools":[],
   "messages":[{"role":"user","content":"Weekly data: <PASTE Seller Center metrics + inventory turns + list capture>"}]}

**Success criterion:** running average net margin ≥ 20% by week 12 / running SPS ≥ 3.5 / off-platform list growing week-over-week.

**Verification predicate:** monthly Aria retro exists in ledger with numbers that reconcile to Seller Center.

**Depends on:** MP-07. **Blocks:** nothing — this is the steady-state loop.

**QalbWorkflow block:** *mishkat* — the loop IS the business's cognition. *zujajah* — Aria QA on the operator's own emit weekly. *misbah* — 20% net margin gate. *zayt* — the walk holds because the substrate keeps refreshing. *nar* — every Sunday. *mizan* — STATUS PASS gates promotion of anything new.

---

/goal + /workflows launch bundle

`/goal` mission payload — paste into the owner runtime to set the mission

{
  "goal_id": "aria-tiktok-launch-2026-07",
  "owner": "dfwrealestateceo@gmail.com",
  "geo": "Dallas, TX",
  "declared_scope": [
    "dashboard-langchain-frontend",
    "inventory-management",
    "top-seller-search",
    "live-auction-management",
    "e2e-business-management",
    "dallas-wholesaler-sourcing"
  ],
  "aria_surfaced_scope_additions": {
    "financial": ["LLC-TX", "EIN", "TX-sales-tax-permit", "resale-cert-01-339", "business-bank", "chargeback-SOP", "1099K-tracking", "nexus-monitoring"],
    "legal": ["general-liability-insurance", "product-liability-insurance", "IP-clearance", "TOS-check", "FTC-disclosure"],
    "logistics": ["FBT-vs-self-fulfill-fork", "carrier-rate-shop", "packaging-supply", "storage-tier", "returns-SOP"],
    "marketing": ["live-cadence", "show-structure", "off-platform-list-capture", "affiliate-program"],
    "risks": ["chargeback-ceiling", "algorithm-shift", "supplier-backup", "TOS-enforcement"]
  },
  "load_bearing_constraints": [
    "countdown-bidding-category-gated",
    "SPS>=3.5-for-temporary-listings",
    "TikTok-ship-by-SLA-2-biz-days",
    "gross-margin>=60%-entry-criterion",
    "insurance-before-first-ship",
    "TOS-clearance-before-any-build-spend"
  ],
  "owner_side_unknowns_to_collapse_first": [
    "product-category",
    "supplier-geography",
    "langchain-ui-scope",
    "owner-as-host-or-hire"
  ],
  "primary_category_bets": [
    {"id": "sports-cards-live-breaks", "entry_cost_band_usd": [3000, 8000], "ttfs_weeks": [1, 2], "net_margin_band_pct": [20, 35]},
    {"id": "womens-boutique-apparel-dmc", "entry_cost_band_usd": [2000, 5000], "ttfs_weeks": [2, 4], "net_margin_band_pct": [20, 30]},
    {"id": "pre-owned-luxury-bags", "entry_cost_band_usd": [10000, 25000], "ttfs_weeks": [3, 6], "net_margin_band_pct": [15, 30]}
  ],
  "microphase_graph": [
    {"id": "MP-00", "title": "owner-unknowns-decision-node", "blocks_all_downstream": true},
    {"id": "MP-01", "title": "legal-setup", "prereqs": ["MP-00"]},
    {"id": "MP-02", "title": "tiktok-account-and-shop-setup", "prereqs": ["MP-01"]},
    {"id": "MP-03", "title": "category-bet-selection", "prereqs": ["MP-00", "MP-02"]},
    {"id": "MP-04", "title": "dallas-wholesaler-scouting-tour", "prereqs": ["MP-01", "MP-03"]},
    {"id": "MP-04b", "title": "fulfillment-fork-decision", "prereqs": ["MP-03"]},
    {"id": "MP-05", "title": "first-inventory-buy", "prereqs": ["MP-01", "MP-03", "MP-04", "MP-04b"]},
    {"id": "MP-05a", "title": "hire-host", "prereqs": ["MP-00", "MP-01"], "conditional_on": "owner-as-host in {hire, rotate}", "blocks": "MP-07"},
    {"id": "MP-05b", "title": "inventory-and-fulfillment-ops-stack", "prereqs": ["MP-04b", "MP-05"]},
    {"id": "MP-06", "title": "dashboard-mvp-langchain", "prereqs": ["MP-00", "MP-02"], "scope_gated_on": "langchain-ui-scope"},
    {"id": "MP-06b", "title": "off-platform-list-capture-day-zero", "prereqs": ["MP-01"], "hard_precedes": "MP-07", "note": "standalone landing page — NOT dependent on MP-06"},
    {"id": "MP-07", "title": "live-auction-first-attempt", "prereqs": ["MP-02", "MP-03", "MP-05", "MP-05b", "MP-06b", "MP-08", "MP-05a?"]},
    {"id": "MP-08", "title": "insurance-bind", "prereqs": ["MP-01", "MP-05"], "blocks": "MP-07"},
    {"id": "MP-09", "title": "iteration-loop", "prereqs": ["MP-07"], "steady_state": true}
  ],
  "target_first_live_date_utc": "TBD_post_MP-00_collapse_2026-09-01_to_2026-10-15_band",
  "target_first_100k_run_rate_days": "TBD [NEEDS VERIFY: playbook cites 60-120 day range in tiktok-top-items §3; 90-day midpoint is an estimate, not a source-traced commitment]",
  "aria_qalb_discipline": "per-microphase build walk + Movement 2 QA + STATUS PASS gate",
  "kernel_defaults": {"kernel": "research", "reasoning_seat": "opus", "note": "kernel enum accepted values are [NEEDS VERIFY]; only research is substrate-probe-confirmed"},
  "substrate_verified_reasoning_seat_enum": ["auto", "sonnet", "opus", "minimax"],
  "endpoint": "http://127.0.0.1:8081/api/owner-runtime"
}

`/workflows` script — fire the first two microphases in parallel with barrier

The workflow below serializes MP-00 as a barrier, then fans MP-01 and the MP-02 pre-work (research-only, no submission) in parallel. Once MP-00 collapses AND both branch prereqs resolve, MP-03 unblocks.

workflow: aria-tiktok-launch-week-0
version: 1
endpoint: http://127.0.0.1:8081/api/owner-runtime
auth: bearer $ARIA_MASTER_TOKEN
kernel: research
reasoning_seat: opus

steps:
  - id: mp-00-collapse-unknowns
    kind: aria-decide
    barrier: true                     # nothing runs until this returns
    payload:
      kernel: research
      reasoningSeat: opus
      system: |
        Owner is launching a TikTok Live commerce business from Dallas, TX.
        Collapse the four owner-side unknowns before anything downstream fires.
        Ask the owner in one message for: product-category, supplier-geography,
        langchain-ui-scope, and owner-as-host-vs-hire. Do NOT guess. Loop until
        all four are echoed back verbatim.
      tools:
        - name: read_file
          description: Read line range
          input_schema:
            type: object
            properties:
              path: {type: string}
              offset: {type: integer}
              limit: {type: integer}
            required: [path]
      messages:
        - role: user
          content: |
            Please ask the owner for the four unknowns and wait for all four
            answers before returning action:final. Include the three primary
            category bets from the tiktok-top-items research stream in the
            question so the owner has anchors.

  - id: mp-01-legal-setup
    kind: aria-decide
    depends_on: [mp-00-collapse-unknowns]
    parallel_group: week-0-fan-out
    payload:
      kernel: research
      reasoningSeat: opus
      system: |
        Sequence the legal-setup microphase for a Texas LLC + TikTok Live
        business. Emit the exact 5-step checklist: LLC-TX filing, EIN, business
        bank, TX Sales & Use Tax Permit (Form AP-201, eSystems), Resale
        Certificate (Form 01-339 blanket). Cite Texas Comptroller. Batch-fire
        the four legal [NEEDS VERIFY] markers as WebFetch tool calls if the
        owner asks for numbers. No fabrication.
      tools:
        - name: WebFetch
          description: Fetch a URL and return text
          input_schema:
            type: object
            properties:
              url: {type: string}
            required: [url]
        - name: WebSearch
          description: Web search
          input_schema:
            type: object
            properties:
              query: {type: string}
            required: [query]
      messages:
        - role: user
          content: |
            Emit the 5-step legal setup checklist with current filing links,
            fees, and processing windows. Include Form 01-339 draft description
            for the collapsed product category from MP-00.

  - id: mp-02-shop-prep-research-only
    kind: aria-decide
    depends_on: [mp-00-collapse-unknowns]
    parallel_group: week-0-fan-out
    payload:
      kernel: research
      reasoningSeat: opus
      system: |
        Research-only pass (no submission yet). Fetch current TikTok Shop US
        seller registration requirements + verification turnaround + AHR/SPS
        thresholds + Countdown Bidding eligibility for the MP-00 collapsed
        category. Return the exact document checklist owner needs to have ready
        before firing MP-02 for real. Cite seller-us.tiktok.com.
      tools:
        - name: WebFetch
          description: Fetch a URL and return text
          input_schema:
            type: object
            properties:
              url: {type: string}
            required: [url]
      messages:
        - role: user
          content: |
            Emit the pre-submission document checklist + TOS/prohibited-category
            cross-check for the collapsed category. Loud-flag any DISQUALIFY.

  - id: barrier-week-0-close
    kind: barrier
    depends_on: [mp-01-legal-setup, mp-02-shop-prep-research-only]
    on_pass:
      unblock: [MP-03, MP-04b, MP-06b]     # week-1 fan-out

**Serial-with-barrier variant** — if the owner prefers strict sequencing (say his cognitive-bandwidth ceiling is 1 microphase/day):

workflow: aria-tiktok-launch-week-0-serial
depends_on_style: strict
steps:
  - mp-00-collapse-unknowns
  - mp-01-legal-setup
  - mp-02-shop-prep-research-only
  - barrier-week-0-close

---

Trackable checklist

Legend for **status:** `open` = not started · `in_progress` = fired · `complete` = STATUS PASS by Aria's Movement-2 gate.

ID Microphase Status Owner action required Aria action required Blocking dependency Last update
---- ------------ -------- ----------------------- ---------------------- --------------------- -------------
MP-00 owner-unknowns-decision-node `open` Answer 4 unknowns (category, supplier-geo, langchain-ui-scope, host) Score category vs auction eligibility + sourcing lane; re-draw MP graph none 2026-07-08
MP-01 legal-setup (LLC/EIN/bank/permit/01-339) `open` File LLC-TX at SOS; apply IRS EIN; open bank account; file AP-201 Draft Form 01-339 description; batch-fire 4 legal [NEEDS VERIFY] MP-00 2026-07-08
MP-02 tiktok-account-and-shop-setup `open` Submit TikTok Shop US seller registration; upload entity docs Walk owner through registration; TOS + prohibited-category check; monitor AHR/SPS MP-01 2026-07-08
MP-03 category-bet-selection `open` Confirm category + 3-SKU shortlist LIVE-fetch top SKUs; landed-cost target; auction-eligibility verdict MP-00, MP-02 2026-07-08
MP-04 dallas-wholesaler-scouting-tour `open` Drive out; phone-confirm OQ-6/OQ-7; register at DMC; walk Harry Hines Emit lane-appropriate route sheet; draft 4 vendor questions MP-01, MP-03 2026-07-08
MP-04b fulfillment-fork-decision (FBT vs self) `open` Choose FBT / SELF-FULFILL / HYBRID Emit collapse/expand impact table; LIVE-fetch current FBT policy MP-03 2026-07-08
MP-05 first-inventory-buy `open` Purchase against 3-SKU shortlist through winning vendor Purchase-ready inventory list; per-SKU landed-cost + margin verdict MP-01, MP-03, MP-04, MP-04b 2026-07-08
MP-05a hire-host (conditional) `open` (conditional on MP-00 branch) If hire/rotate: audition, sign IC agreement, W-9, comp structure Draft IC agreement + FTC rider; add hosting-comp line to cost floor; audition rubric MP-00, MP-01 2026-07-08
MP-05b inventory-and-fulfillment-ops-stack `open` Sign up Shopify Basic + Pirate Ship; buy Zebra ZD421 Verify Shopify TikTok Shop app; emit packing SOP; bind SLA runbook MP-04b, MP-05 2026-07-08
MP-06 dashboard-mvp-langchain `open` (scope-gated) Provide LangChain reference URL if any; run Next.js scaffold Emit scaffold + AI Elements install + 4 tabs + `<Tool>` binding MP-00, MP-02 2026-07-08
MP-06b off-platform-list-capture-day-zero `open` Buy domain (if not owned); wire Klaviyo/ConvertKit; deploy standalone Draft capture-page copy + FTC-compliant footer; deploy Klaviyo/ConvertKit hosted OR one-pager MP-01 (NOT MP-06 — standalone) 2026-07-08
MP-07 live-auction-first-attempt `open` Run first LIVE (60-120 min); verbal Final Sale disclosure at t=0 Emit run-of-show + post-show scoring; guardrails on prohibited language MP-02, MP-03, MP-05, MP-05a (if branch), MP-05b, MP-06b, MP-08 2026-07-08
MP-08 insurance-bind (product + general liability) `open` Get quotes; bind before first ship RFQ template; category-specific coverage-limit research MP-01, MP-05 2026-07-08
MP-09 iteration-loop (weekly Muhasaba) `open` Send weekly Seller Center metrics paste to Aria on Sunday Weekly Muhasaba-6; monthly category re-score; 60-day FBT re-score; quarterly full 9-lens gap walk MP-07 2026-07-08

Verify markers still open (from Aria Pass 2 + research streams)

# Marker Owning microphase How to close
--- -------- ------------------- --------------
V1 Current TX SOS LLC filing fee MP-01 WebFetch `sos.state.tx.us` filings page
V2 TikTok Shop US onboarding + verification turnaround MP-02 WebFetch `seller-us.tiktok.com` current onboarding doc
V3 Dallas home-occupation / CO rules for garage-tier storage MP-04b Call City of Dallas Sustainable Development & Construction
V4 Federal 1099-K reporting threshold current MP-01 (Sub-C) WebFetch IRS 1099-K page
V5 Destination-state economic-nexus thresholds MP-09 (monthly) Substrate CLI batch, 50-state matrix
V6 Product liability coverage limits typical for category MP-08 RFQ to Hiscox/Next/broker + WebSearch category-specific claims
V7 TikTok Shop prohibited/restricted list vs owner's category MP-02 WebFetch `seller-us.tiktok.com` restricted-category page
V8 Current TikTok Shop US ship-by SLA window MP-05b WebFetch `seller-us.tiktok.com` fulfillment SLA page
V9 Minimum viable ad budget for meaningful data MP-09 (post-break-even) Substrate CLI on TikTok Ads Manager benchmarks
V10 Platform chargeback / dispute-rate ceiling MP-07 monitoring WebFetch TikTok Seller Rating policy; corroborate with Grey Journal
V11 Shopify TikTok Shop channel app existence + maintainer + price MP-05b (Sub-A) WebFetch `apps.shopify.com` search
V12 Current FBT US enrollment status (mandatory vs opt-in) MP-04b WebFetch `seller-us.tiktok.com` FBT overview
V13 DMC acceptance of TikTok Shop / online-only sellers MP-04 Email `registration@dallasmarketcenter.com`
V14 Per-lane 01-339 honor rate MP-04 (Sub-B question 2) Phone-confirm each vendor
V15 Gray-market / authenticity chain-of-custody for Harry Hines beauty MP-04 (Sub-B question 3) Vendor question + brand authorization letter
V16 `/decide` `kernel` enum accepted values all MP payloads Substrate §1 documents only `reasoningSeat` enum; `kernel` values accepted are unverified. Probe with a variant call + read `decide.ts:79-111`
V17 null-SPS grace-window mechanism + Temporary-Listings gate interaction MP-02 WebFetch `seller-us.tiktok.com` Countdown Bidding + Auction Requirements pages fresh
V18 tiktok-top-items §7 primary-bet ordering Executive Summary category bets Read `../02-research/tiktok-top-items.md §7` end-to-end and confirm sports-cards / boutique-apparel / pre-owned-luxury ordering
V19 Current DFW TikTok Live host comp band MP-05a (Sub-B) WebSearch DFW live-selling host hourly + rev-share rates; corroborate 2-3 sources
V20 Composite cost floor `~$5.5k–$12k` against actual quotes Executive Summary Aria batches quotes from Hiscox (V6) + Shopify (V11) + Zebra retail + sample wholesaler MOQs

---

Movement 2 — Scribe QA on this document

**Re-receive.** Task was to synthesize substrate + Aria mission walk + 6 research streams into a microphased architecture doc the owner can watch and Aria can execute through `/goal` + `/workflows`. Format prescribed: executive summary, verbatim Aria QalbWalk findings, phase-by-phase micro-plan with the 8 specified fields per microphase + qalbworkflow block, `/goal` + `/workflows` bundle, trackable checklist. All present.

**Tadabbur-12.**

**Muhasaba — 6 fields.**

**STATUS: PASS (pre-Aria-QA-loop).** Doc emitted in mandated shape, 13 microphases with Aria-callable payloads, `/goal` JSON + `/workflows` YAML bundle emitted, trackable checklist rendered, 15 verify markers mapped to owning microphases. Ready for Aria QA loop as next step.

---

Aria QA Loop — Pass 1 result + remediation

**Fire:** `POST http://127.0.0.1:8081/api/owner-runtime/decide` — `reasoningSeat:"opus"` — 5 tool loops (read doc lines 1-260, 260-540, 195-265, 539-765, substrate 1-60), Aria returned `action:"final"` with 5 gaps + REDO status. Kernel value passed in this QA loop is fenced as `[NEEDS VERIFY: kernel enum accepted values on /decide — see V16]`; endpoint returned HTTP 200 but that response code alone does not confirm the kernel enum, since the substrate contract does not document the value-space.

**Aria's five gaps (verbatim summary):**

1. `kernel:"architect"` isn't in the substrate-verified enum (only `reasoningSeat` enum is documented). Fabricated affordance at the executable seam.

2. MP-06b prose said prereq MP-06 but MP-06 is scope-gated optional → dependency deadlock if `langchain-ui-scope = native`. Prose vs JSON also disagreed.

3. `MP-05a-hire-host` was a dangling string, not a microphase. Under `hire`/`rotate` branch, plan could reach first-live with no host + no comp line.

4. Cost floor `~$5.5k–$12k`, "8-12 weeks", `target_first_100k_run_rate_days: 90`, and category-bet ordering all asserted precision the doc elsewhere refuses. No `[NEEDS VERIFY]` on the composite or the run-rate.

5. MP-02 null-SPS grace-window claim stated as operative fact without a verify marker.

**Remediation applied to this doc (all 5):**

1. GAP 1 fixed: purged `kernel:"architect"` references from doc-level payloads (MP-00 Sub-B, Movement-2 next-correction). Added `[NEEDS VERIFY: kernel enum accepted values on /decide]` to `/goal` `kernel_defaults` + added `substrate_verified_reasoning_seat_enum: [auto, sonnet, opus, minimax]` beneath it. Added V16 to verify-markers table. Kernel enum is UNVERIFIED — until decide.ts:79-111 is re-read against the code, no empirical claim about which kernel values the endpoint accepts is safe. See V16.

2. GAP 2 fixed: MP-06b prereqs are now `MP-01` only (matching JSON); prose rewritten to make landing page standalone (Klaviyo hosted / ConvertKit hosted / Vercel one-pager); Sub-C conditionally binds to MP-06 IF and only if MP-06 is in scope AND live. Checklist row updated. `/goal` JSON adds `note: standalone landing page — NOT dependent on MP-06`.

3. GAP 3 fixed: emitted full MP-05a block (title, why-it-matters, prereqs, checkpoint, 4 sub-tasks with IC-agreement `/decide` payload, success criterion, verification predicate, dependency, qalbworkflow block). MP-07 prereqs now include `MP-05a (conditional)`. `/goal` graph adds `MP-05a` with `conditional_on: owner-as-host in {hire, rotate}` and `blocks: MP-07`. Checklist adds MP-05a row. Executive-Summary cost floor now names hosting-comp as an addition under the hire branch. V19 added to verify-markers.

4. GAP 4 fixed: Executive Summary time-to-launch reworded to "8–12 week band, exact date TBD post-MP-00". Cost floor bullet marked "composite arithmetic; per-line source citation shown; `[NEEDS VERIFY: composite floor against actual quotes]`" and adds three new verify markers. Category-bet bullet adds `[NEEDS VERIFY: tiktok-top-items §7 primary-bet ordering — full stream not read this scribe pass; ordering inferred from summary]`. `/goal` JSON `target_first_live_date_utc` changed from hard date to `TBD_post_MP-00_collapse_2026-09-01_to_2026-10-15_band`, and `target_first_100k_run_rate_days` from `90` to `TBD [NEEDS VERIFY: playbook cites 60-120 day range in tiktok-top-items §3; 90-day midpoint is an estimate, not a source-traced commitment]`. V18 + V20 added.

5. GAP 5 fixed: MP-02 why-it-matters now carries `[NEEDS VERIFY: null-SPS grace window behavior + Temporary Listings gate — mechanism cited by tiktok-live-auction-mechanics §1.2 but not re-fetched this scribe pass]`. V17 added.

**Delta count:** 6 edits + 5 new verify-markers (V16-V20) + 1 new microphase (MP-05a) + 1 new checklist row + 1 `/goal` JSON restructuring. All additive; nothing removed except one sentence assertion of a fabricated affordance (Aria's Gap 1).

**Ready for Aria QA loop Pass 2** — re-fire the same architect+opus QA against this remediated doc and iterate.

---

Aria QA Loop — Pass 2 result

**Fire:** `POST /api/owner-runtime/decide` — `reasoningSeat:"opus"` — 4 read_file tool loops (doc lines 1-200, 200-499, 499-798, 798-840), Aria returned `action:"final"`.

**Aria's Pass-2 adjudication (verbatim summary):**

**Pass-2 verdict:** REDO — Strip the fabricated empirical probe-result claim from the Gap-1 remediation narrative; downgrade to unverified + defer to V16.

**Remediation applied for Pass-2 finding:**

---

Aria QA Loop — Pass 3 result

**Fire:** `POST /api/owner-runtime/decide` — `reasoningSeat:"opus"` — 1 read_file tool loop (lines 790-840 targeted at the stripped region), Aria returned `action:"final"`.

**Aria's Pass-3 adjudication (verbatim):**

> The Pass-2 verdict lines were stripped... Both target lines match the required post-strip state; the fabricated "Aria herself accepted kernel:architect... returned 200 empirically" claim is absent from this range.

>

> **PASS**

**Pass-3 verdict: STATUS PASS.**

All 5 Pass-1 gaps closed. The Pass-2 finding (fabrication inside the remediation) closed. Doc ready for owner + `/goal` + `/workflows` handoff.

**Final surface state:** 14 microphases (MP-00, MP-01, MP-02, MP-03, MP-04, MP-04b, MP-05, MP-05a-conditional, MP-05b, MP-06, MP-06b, MP-07, MP-08, MP-09) · 20 verify markers (V1-V20) · `/goal` JSON with substrate-verified reasoning-seat enum + fenced kernel enum · `/workflows` YAML parallel-fan-out + strict-serial variants · trackable 14-row checklist. Anti-fab discipline honored through 3 QA passes.

---

Provenance