Revion Crew Board

Two manual drafts stood in #ach-payouts-production with no send or reconciliation after them at the 2 Sep collection: 27 Aug ($1,139.16, 9 orders) and 1 Sep 19:27Z ($358.16, 2 orders, “this Slack notification does not initiate a transfer”). New today in Linear: REV-150 (arm production payouts, [Backlog] p1 upd 27 Aug at the last read) is absent from the top-50 open list — today’s window reaches back to 19 Aug, so a still-open REV-150 would appear; it was closed or cancelled, and no 13–14 Sep commit names payout arming. In its place sits REV-221 “Automatic payouts ON by default for every merchant (enable Viltrumite now)”, [Backlog] p1 @nur, upd 13 Sep — “Viltrumite” is not one of the six known merchant channels, so who that is is itself unverified. Slack and the prod DB have failed every collection since 2 Sep: a send in the last twelve days is evidenced nowhere.

Next action

Hand Andrew both amounts with Halo’s approved banking (Mercury checking ••••2806) and the business address from app_R89kpf0lM4Jd5X_r for the BofA portal; record both bank references in Revion. Then get one line from Nur on REV-221: what REV-150 closed as, and who Viltrumite is.

31 Aug 15:54Z Jarrod: his accountant says a Washington business must collect WA tax on these orders. You answered citing agreement sections 8.5 (Revion registers, collects, remits) and 8.7 (resale certificate); 31 Aug 23:48Z: “my contract ends at section 8.4, so I never saw 8.5 or 8.7.” 1 Sep 05:57Z you conceded the text is from the master document, not the one he signed, and promised to discuss with Aadil that day — nothing readable in #rhp-rhpeptides since the 2 Sep collection. His store is live for real: RC-2026-1000243, $598, AUTHORIZED 2 Sep 05:08Z. What moved: REV-196 (“$0 FL quotes, refund collected tax, purge TaxJar test transactions”) went p2 → p1 in the 12 Sep triage — unregistered-state tax is now rated urgent, which makes the WA answer the policy, not a one-off. REV-190, whose reminder sweep skipped RHP, left the open list; whether any sweep now touches him is Nur’s armed-or-not item.

Next action

Get the executed RHP agreement from Aadil (on his list), decide addendum vs. accepting his position on WA tax, reply in-channel today — and mark the $598 while you are there. One policy answers Washington and Florida both.

Set live + banking approved 31 Aug 15:24Z (done). At the 2 Sep read he was still registered as https://www.biopepusa.com — the www/apex shape that broke RHP’s and Kiyora’s first checkouts. Casey created a session himself 31 Aug 15:37Z ($59.00, info@biopepusa.com) which expired with no payment intent: he reached our checkout and never submitted a card — why is unverified. #biopep-usa-llc held only our team, no welcome posted. Twelve days on, whether the domain got corrected or Casey retried is still unverified: the prod DB and Slack failed again in today’s 13:06Z collection.

Next action

Correct the domain to https://biopepusa.com via single-field DB update (never suspend — it wipes credentials), then email Casey directly: welcome, “run the $59 test again and we void it”, accept the Slack invite.

Banking approved 31 Aug 12:55Z; both test refunds confirmed to her and her only reply was “Thanks so much” (31 Aug–2 Sep window). Prod DB at 2 Sep 13:05Z: 0 sessions, 0 orders for mrc_zinhzQp6eyBC since 31 Aug — the twelve days since are unknown (every watch-merchant read from 7 Sep onward has come back empty). Aadil (2 Sep 05:46Z): “one transaction in five days” — Andrew is counting, and this is the merchant closest to adding some.

Next action

Ask Jenifer plainly when Revion goes on as the live payment option and offer to watch the first day; no reply by tomorrow → Aadil calls (his chase task).

app_uHyTgAc1QXxFQKBW (coreresearchpeptides.com) moved received → under review 1 Sep 21:07Z and was still there at the last successful DB read (2 Sep 13:05Z); every read since has failed, so the current state is unverified. The REV-214/218/219 cluster (verify callback → activate → then credentials), all upd 11 Sep, is absent from today’s top-50 — closed during the window — and git shipped the matching piece: “queue missing webhook setup email on provisioning” + “send webhook setup email after credentials acceptance” (#429, merged 14 Sep, promoted #430). So the sequence question this task was waiting on has an answer in code; whether callback verification itself now gates activation is Nur’s line to give. REV-169 (credentials email copy) survived the 12 Sep triage at p2 — still open, still uncorrected.

Next action

Approve → descriptor → banking → welcome; then onboarding-check.sh app_uHyTgAc1QXxFQKBW before telling him “check Slack”. The credentials email still carries the “payments disabled” line — correct it in-channel until REV-169 ships.

Chased 31 Aug 09:06Z with the $143.96 abandoned-session evidence; no reply by 2 Sep, and Slack has been unreadable in every collection since. Prod DB at 2 Sep: 0 sessions for mrc_4JFn-SyBlVNP since 31 Aug. Chris last spoke 28 Aug 04:07Z. The automated nudge that would cover him is in Nur’s armed-or-not item — REV-190 closed in Linear with no proof the sweep is on.

Next action

Stop nudging in Slack — hand to Aadil for a phone call with a booked slot (cal.com/team/revion).

Verified again today: REV-127 “Plugin: store each order’s Revion payment link in Woo order notes + meta” is [In Progress] p2 @tim, last updated 2 Sep — twelve days without an update, and one of only two tickets showing you as the active assignee (the other is REV-101, the revioncaps.com redesign, also In Progress, also upd 2 Sep). It survived a triage sweep that re-leveled a dozen other tickets — nobody touched it either way.

Next action

Either finish it this week or hand it to Nur and say so on the ticket — an In Progress p2 with no update is what Andrew reads as “nobody is on it”.

Whether the ticket has been resolved is still unverified (no billing-inquiry rows in any bundle since 2 Sep; the prod DB failed again today). On the QA side: REV-220 (the 12 Sep daily) has left the open list — the first daily ever to close — while REV-215 (11 Sep) is now the newest open one and REV-60 (the shared plan) was edited 11 Sep. The plan is live, so the “clean up after yourself” insert is worth landing now rather than after another week of leftovers. Verified at 2 Sep: the 1 Sep QA run ended US-7 + US-8 FAIL with three ZZTEST applications still “received”.

Next action

revioncaps.com/admin/billing-inquiries → Resolve, note “QA daily test — no action”. Locally: source ~/RevionCaps/creds.sh && python3 ~/RevionCaps/linear_cleanup_rev60.py (idempotent; it also comments on REV-170).

banking_approved 2 Sep 06:28Z; 0 sessions since 31 Aug at the 2 Sep read; their dev has not written since 20 Aug (“Alright, scheduled for next week”). The embed they would be testing got rebuilt 10 Sep — session bound with partitioned cookies, graceful fallback where CHIPS is absent (#419/#420, promoted #421) — and today REV-228 “PRIMARY: Embedded in-page Visa checkout box (default), redirect becomes fallback — conversion” landed as p1 @nur: the embedded path is becoming the default product. Whether any of that was ever the dev shop’s blocker is unverified, but it is a concrete reason to re-open the conversation.

Next action

The call is the test: Aadil books it by phone/email (contact is the dev shop, dev@guardedpaypro.com); run and void the first embedded order together, on the 10 Sep build.

Your question (2 Sep 05:46Z, “drop-off due to people not being able to check out with their chosen card brand”) has a measured answer from Tim (2 Sep 06:04Z): declines of $380 (31 Aug, txn 12495306934) and $361 (1 Sep, txn 12500637032), both card_declined, both non-Visa — “the Visa-only wall”. Every merchant is still told “Visa only for now”. New today: REV-228 “PRIMARY: Embedded in-page Visa checkout box (default), redirect becomes fallback — conversion” (p1, upd 14 Sep) bakes the brand restriction into the product’s next flagship surface. If Mastercard/Amex/Discover can be enabled on the NMI MID, that plan changes shape — so the brand answer is now a design input, not just recovered revenue. No brand-enablement ticket appears in today’s top-50, and no answer is visible in any source since 2 Sep.

Next action

With Andrew/Ignacio: what enables Mastercard/Amex/Discover on the NMI MID, and by when — give Tim and Nur a date before REV-228 hard-codes Visa-only into the embed.

Open since 28 Aug; Nur asked again on the morning of 29 Aug. No update in Core Engineering by 2 Sep, and TG has failed in every collection since — whether it went out is unverified. Meanwhile his 13–14 Sep output alone: encrypted server-side /apply drafts + submit freeze (REV-204, #432/#434), the NMI reversal webhook inbox (REV-172, #438/#440), the onboarding webhook-setup email (#429), a CI/CD overhaul, and six production promotions. Keep him shipping.

Next action

Send it, confirm in the thread.

Hudaifa Elahi rejoined #halo-peptides 31 Aug 15:52Z (the “restore the channel” half of REV-163 is done); nobody had written there by the 2 Sep collection, and Slack has been unreadable since. Verified today in Linear: REV-163 is still [Todo] p2 @aadil, last updated 2 Sep. Owed: $1,139.16 + $358.16. The ticket asks you to explain what was fixed (Apple Pay encoding, consent gate, order-note links) and what is coming (stable payment links), give a payout date, and get Sufyan to retest.

Next action

Call, then post the payout date in #halo-peptides so it is on record.

Jarrod’s copy ends at 8.4 (31 Aug 23:48Z); Tim: “that text is from the master document, not the one you signed… I will discuss with Aadil today” (1 Sep 05:57Z). His accountant wants WA tax collected; Tim also tagged you 31 Aug 16:36Z on Washington registration timing ($100k trailing-12-month threshold). The Florida twin, REV-196, was raised to p1 in the 12 Sep triage — the team already treats unregistered-state tax as urgent, so one policy answers both states.

Next action

Send Tim the executed RHP agreement plus your call: an addendum carrying the tax clauses, or accept his position. He transacted for real on 2 Sep ($598).

Prod DB since 31 Aug, at the last successful read (2 Sep): Kiyora 0 sessions, Zen 0, Best Peptide 0, Biopep 1 (Casey’s own $59 attempt, expired). Tim (2 Sep 06:05Z): “we have to push merchants from all sides to transact. Slack does not work very well.” You offered to do the chasing on 31 Aug. What changed in the blind window: REV-210 (the outreach campaign) was reprioritised p0 → p3 on 12 Sep, and REV-190 (the idle-reminder sweep) left the open list with no proof it is armed — so the automated demand levers were consciously parked or closed, and the personal chase is what remains.

Next action

Per merchant: Kiyora — switch real traffic on; Zen — one $-test with Tim watching; Best Peptide — book the 20-minute embed call (the embed is now the flagship: REV-228); Biopep (Casey) — retry the test and accept the Slack invite.

1 Sep 20:52Z: “can we approve Apex Peptide Supply… no payment methods at all… desperate guy.” Tim (2 Sep 06:02Z): “we do not have his application.” Today REV-226 “Webhook timeouts cancel PAID Woo orders (Apex): faster retry + plugin ack-then-process + reconcile” landed as p1 @nur. “Apex” is also the name of the long-standing sandbox demo merchant (Apex Demo), and with Slack and the prod DB dark there is no way from here to say which the ticket means — but both readings need you: if it is the prospect, he got in during the blind window and his first impression is paid orders being cancelled; if it is the demo, his application is still missing. The funnel got safer meanwhile: /apply no longer loses everything on refresh (REV-204 shipped 14 Sep), though REV-93 (scanned PDFs) is still open — his documents need to be clean exports.

Next action

Ask Nur which Apex REV-226 names. If the application is still missing, send him revioncaps.com/apply today and agree the rate (REV-189 closed silently — confirm with Nur what rates the system actually accepts before promising one).

VAPI_PRIVATE_KEY etc. are SET on revioncaps-prod; calls are held off only by VAPI_OUTBOUND_ENABLED=false. No calling-window, DNC, or consent gates exist in vapi-outbound.ts. Unchanged since 31 Aug — and today’s Linear read confirms again that no VAPI ticket appears among the 50 most recently updated open issues, so the decision still lives nowhere but this board.

Next action

Either unset the prod keys or commission the TCPA gates. Deciding is the task; whichever way it goes, file it so it stops living in a chat.

Verified today: REV-164 is [Todo] p3 @aadil, last updated 27 Aug — unchanged since the day it was filed, and it survived the 12 Sep triage untouched. Readable by anyone who scrolls Core Engineering.

Next action

Rotate it, delete the message, close the ticket.

Decided in TG 31 Aug 13:37–13:43Z: DocuSeal replaces DocuSign seats; filed as REV-187, still [Backlog] p3 @unassigned, upd 1 Sep. But today’s Linear shows fresh 13 Sep activity on the opposite path: REV-87 “Move the merchant agreement e-sign flow to DocuSign” ([Backlog] p2 @nur) and REV-144 “banking verified + agreement signed as gates to LIVE (auto MFSA via DocuSign)” ([Todo] p2 @nur). Meanwhile the five 2 Sep contracts (Pura 8.5, Planet 8.5, Fit Aminos 9, Sports Tech 10, Epic 12) still have not shown up as applications in any readable source — the sign-but-never-apply leak has no chase.

Next action

One line in Core Engineering: which e-sign direction stands, DocuSeal or DocuSign — then chase the five signed-but-never-applied merchants yourself this week (your own 31 Aug commitment).

Variants live in the “Revion Home Variants” artifact (REV-101, [In Progress] @tim, upd 2 Sep). Waiting since 27 Aug; no letter seen in any readable source. New: REV-174 was reprioritised p2 → p4 on 12 Sep. The brand-identity pass has been live since 7 Sep (#406) regardless.

Next action

Reply with a letter in the Core Engineering thread — or say the redesign is parked so REV-174 and REV-101 can close and the identity pass stands as the end state.

REV-226 “Webhook timeouts cancel PAID Woo orders (Apex): faster retry + plugin ack-then-process + reconcile”, [Backlog] p1 @nur, upd 14 Sep. Sibling filed the same day: REV-227 “checkout.expired webhooks fail platform-wide (422/session_mismatch/order_missing/too_large)”, p2 — the outbound webhook layer is failing in both directions at once. Parts already in place: outbound webhooks identify themselves (#417/#418, 9 Sep), the webhook-setup email now queues at provisioning and sends after credentials acceptance (#429, 14 Sep), and your REV-172 reversal inbox (#438/#440, promoted #439) demonstrates ack-then-process on the inbound side. Which merchant “(Apex)” is — the sandbox demo or Apex Peptide Supply — cannot be verified with the DB and Slack dark; Halo is the largest Woo merchant exposed either way.

Next action

Ack-then-process in the plugin + retry with backoff, then a reconcile sweep for orders cancelled after a successful charge — and one line to Aadil naming which Apex.

At the 12 Sep read REV-150 was [Backlog] p1, untouched since 27 Aug. Today it is absent from the top-50 open list — its 27 Aug update sits well inside today’s window (cutoff 19 Aug), so it was closed or cancelled during the blind days, and no 13–14 Sep commit names payout arming (the only payout code remains 12 Sep’s “Fix self-contained manual payout instructions”, promoted #423). In its place: REV-221 “Automatic payouts ON by default for every merchant (enable Viltrumite now)”, [Backlog] p1 @nur, upd 13 Sep. The original three gaps stand unless they were closed silently: settlement evidence for Halo’s NMI orders (2 manual_review rows), ACH generation (PAYOUT_ACH_SLACK_ENABLED="false" on prod, no ACH_* secrets), maturity math (t+4 FedACH days). Outstanding drafts: 27 Aug $1,139.16 · 1 Sep $358.16.

Next action

One paragraph on REV-221: what REV-150 closed as, which of the three gaps remain, and who “Viltrumite” is — then arm gaps 1–2 with a penny-test file to Mercury before the first real send.

REV-228 “PRIMARY: Embedded in-page Visa checkout box (default), redirect becomes fallback — conversion”, [Backlog] p1 @nur, upd 14 Sep. The foundation shipped in the blind window: the embed binds its session with partitioned cookies and degrades where CHIPS is absent (#419/#420, promoted #421, 10 Sep). Two inputs would sharpen it before code: Narboto’s PostHog replays of the ~9 stalled sessions (where exactly they stopped — price, trust, card form, error), and Aadil’s brand answer — the ticket title bakes in “Visa”, and if Mastercard/Amex arrive on the MID the box should not need a rebuild. Best Peptide’s dev-shop call is the first integration this changes.

Next action

Sequence it against REV-226 — say which p1 goes first — and state what the embed shows a non-Visa buyer today, before the box becomes the default surface.

All three (upd 11 Sep at the last read) are absent from today’s top-50 — closed or cancelled. Visible in git: “queue missing webhook setup email on provisioning” + “send webhook setup email after credentials acceptance” (#429, 14 Sep, promoted #430) — the setup-pack half. Also visible: a rev219-named branch whose commit is about the release pipeline (“exclude Woo plugin update from verification candidate”, #428, 13 Sep), which reads like a different REV-219 than the callback ticket — the mapping is unverified. Whether callback verification now gates activation, or the cluster was folded elsewhere, is one line for you to give. REV-169 (credentials email copy) survived the triage at p2, still open, still unnamed in any commit.

Next action

State what shipped vs dropped from the sequence, and whether Sean Finck goes through it or by hand — fold the REV-169 template fix into the same email path while it is open in your editor.

emails.ts:368 still carried the line at the 2 Sep read. Today: REV-169 “Fix the merchant credentials email copy” is [Backlog] p2 @nur, upd 12 Sep — touched by the triage, no state change, and no commit names it (the 14 Sep webhook-setup email is the adjacent send, not this template). It went to Biopep on 31 Aug; humans correct it in-channel each time — the template is the fix. Sean Finck plus the five signed contracts queue behind it.

Next action

One template edit, shipped with whatever remains of the REV-214 sequence: state what is live, what is pending, and what the merchant does next — no “payments disabled”.

REV-189 (“fee rate 8.5 / 9 / 10 and any %”, open at the 2 Sep read) is still absent from the top-50 today, and no commit in the 13–14 Sep window names fee rates either — what shipped, if anything, remains unverified. The defect it was filed for, at the 2 Sep code read: zod locks feeRate to 12|14 in three places, the admin types it as 12 | 14, sandbox-provisioning-client.ts:66 refuses other values, and settlement.ts guards fee_rate IN (12, 14) twice. REV-192 sent Pura 8.5, Planet 8.5, Fit Aminos 9, Sports Tech 10, Epic 12.

Next action

One line back: closed as done, or cancelled? If done — prove 8.5 round-trips to fee_cents, renders “8.5%”, and passes both settlement guards. If cancelled — Aadil has five contracts to re-paper at 12/14, and he needs to know before Apex gets quoted a rate.

Today: REV-196 “Turn OFF Florida sales-tax collection until registered: $0 FL quotes, refund collected tax, purge TaxJar test transactions”, [Backlog] p1 @nur, upd 12 Sep — up from p2. It is live, not hypothetical: TaxJar genuinely answers rather than falling back — a Miami 33131 quote returned $3.85 on $54.99 (7%, Miami-Dade incl. county surtax) per BACKLOG.md — so FL buyers are quoted and charged real tax. Same class of problem as Jarrod’s Washington question, now thirteen days open on Tim’s side.

Next action

$0 FL quotes behind a config flag today, then produce the list of FL orders already charged so the refunds can be issued, and purge the TaxJar test transactions.

support_admin_alert emails to support@send.revioncaps.com showed [failed] repeatedly from ~28 Aug in the prod email log (seen at the 31 Aug board build); earlier alerts to support@revioncaps.com delivered. No bundle since has carried email-log rows, so still-failing is unverified — and nothing in git through 14 Sep touches it. The general fix, REV-146 (“capture every checkout/gateway/plugin/webhook failure into an admin page + Slack”, Nar’s ask), survived the triage at [Todo] p2, upd 12 Sep.

Next action

Compare the sending domain on send.revioncaps.com with the delivered revioncaps.com alerts and say which one is wrong; you are the only person who can read that log while the DB is unreachable from here — and the same domain carries any future outreach send (REV-210).

REV-190 (upd 6 Sep at the last read) is absent from today’s top-50 (cutoff 19 Aug) — closed or cancelled. The code merged and was promoted 6 Sep off-by-default (#376/#402/#403). Nothing anywhere shows REVION_IDLE_REMINDER_ENABLED=true on prod, the RHP skip flag (REVION_IDLE_REMINDER_SKIP=mrc_gu6vQA07rA7D — the 8.5/8.7 dispute is still open), a rolled release, or a single merchant_idle_reminder email row. Four merchants sit at zero as of the last readable DB day, so a working sweep would be earning its keep right now.

Next action

Paste the two flag values and one emails row of kind merchant_idle_reminder (max 1 per merchant per 72h) — or say it is off and what the ticket closed as.

REV-204 closed with visible code: encrypted server-side application drafts (#432) + edit freeze through save/submit (#434), both merged 14 Sep and promoted the same day — the /apply refresh trap is dead. REV-166 (checkout link exposes buyer PII, In Review since 2 Sep) and REV-203 (storefront bootstrap fires unused requests on the hosted-checkout page, p0) are absent from today’s window with no commit naming them. REV-124 (loading overlay swallows clicks) is no longer visible either, but its last update may predate today’s 19 Aug cutoff, so it may merely have aged out rather than closed. Still verified open: REV-165 funnel/decline report (p2) · REV-146 error log + Slack (p2) · REV-142 silent-merchant alert (p2, reprioritised from p0 on 12 Sep — the alert that would have caught Halo going dark) · REV-145 plugins page (p2) · REV-173 telemetry null (p4) · REV-100 Apple Pay undiagnosable (p2) · REV-104 payment links for operators (p2).

Next action

One line each on REV-166 and REV-203 — shipped or cancelled — then REV-142: after a triage that downgraded it, say when the silent-merchant alert actually lands.

Half the wound got infrastructure today: the signed durable NMI reversal webhook inbox merged in observation mode (#438, review fixes #440, promoted #439) — gateway-side refunds and reversals stop being invisible to the platform, which is what REV-172 was filed for. The other half is unchanged: settlement.ts deducts refunded_cents but never recalculates fee_cents down, and MFSA 16.4 argues the fee share should come off too. Invisible until a merchant partially refunds a settled order — cheaper to fix while the inbox is still observing rather than acting.

Next action

Add the fee recalc before the inbox graduates from observation mode, so the first observed reversal settles right.

REV-152 completed 28 Aug 15:05Z. Prod at the 2 Sep read: mrc_roNy1mlK9Ij5 domain https://www.biopepusa.com. PR #364 folds a leading www for success_url/cancel_url and public-checkout CORS; whether “verify at registration”, “log rejected checkouts” and a correction path that does not route through suspend shipped is still unverified from any bundle since — nothing in the 13–14 Sep window names it.

Next action

Say which of the three items shipped; if any did not, file the remainder — closing the parent while merchants keep landing on www is how this reaches a fifth merchant. Tim fixes Biopep’s row by hand meanwhile.

slack_provisioning_status/_error exist on the merchant row but nothing shows them. #biopep-usa-llc still held only our team 48 hours after creation (as of 2 Sep), and only a shell query can say whether Casey’s invite failed or sits unaccepted. Welcome v2 (#412; delivery fixed #415/#416) auto-posts the integration pack once a channel exists, and the setup-pack email shipped 14 Sep (#429) — invite status in admin remains the unexposed link in that chain.

Asked in Core Engineering 31 Aug 13:48Z mid-triage of a phantom inquiry: “Would be sick if we had the customers number ready for anyone/me/andrew/pat to call them right away too.” The buyer’s phone is collected at checkout; the billing-inquiry queue and its alert email do not surface it. Verified today: no such ticket appears among the 50 most recently updated open issues.

Next action

Show the buyer’s phone on the billing-inquiry admin view + in the alert email. File it in Linear with the LLM pre-check block so it stops depending on this board to be remembered.

Today: REV-174 “Implement the revioncaps.com home page redesign (variant A/B/C from REV-101) + logos” is [Todo] p4 @nur, upd 12 Sep — reprioritised from p2. REV-101 (the variants, Tim) is still [In Progress], upd 2 Sep. Whether a letter was ever picked is unverified — TG has been dark since 2 Sep.

Next action

Nothing before Aadil’s letter; if the p4 means it is parked, say so on REV-174 so REV-101 can close too.

Tim (2 Sep 06:04Z): ~20 real sessions in the 28 Aug – 2 Sep window, 5 completed, 2 non-Visa declines, ~9 reached our page and never submitted a card (~$2k) — “that is exactly the PostHog replays — Narboto has the invite, replay review of these session IDs is the next step.” Halo sessions to pull first: cs_1MxLyGmlAnYE1uDVNOXz $340 · cs_iCIj0VavIkfhPcP5IFQG $205.20 · cs_42nR0ZkxuLfKm7MsBVBV $292.50 · cs_aBS89ts7SbcqM7oZVV2n $321 · cs_w5dSBawUW7yVwtrTWsUW $80 · cs_1da8x8iU3zmd6bY8lThA $151.20 (expired, then completed on retry) · cs_2G4sPw1W9GLgBLT_NKYX $69; plus Biopep’s cs_f5kPI4eMiSQd1xpSPnrk $59. Server-side payment_success/payment_failed events (#404, #410) let replays be joined to server truth. What changed: REV-220 (the 12 Sep daily this was to land on) has itself closed; REV-215 (11 Sep) is the newest open daily; and REV-228 now proposes rebuilding the checkout around the embedded box — the where-they-stopped distribution decides whether that is the right fix.

Next action

Watch each replay; one line per session — where they stopped (price / trust / card form / error) — on the current daily and in Core Engineering; hand the distribution to Nur before REV-228 fixes the wrong step.

Prod DB, merchant mrc_rhnnN_Yat26u (Halo), since 31 Aug: orders RC-2026-1000228, -232, -233, -235, -236, -242 — all $0.50, all AUTHORIZED, all from narbotokerimov@gmail.com (31 Aug 11:32Z → 1 Sep 16:47Z), none refunded at the 2 Sep 13:05Z read; every DB read since has failed, so a refund in the last twelve days is unverified. Nur’s expired $0.50–$3.00 sessions and Aadil’s $106 one sit on the same merchant. REV-216 (your real-QA-stores ticket, touched 13 Sep) is the structural end of this.

Next action

Say what they were; refund the six from admin today. Also the 1 Sep leftovers: US-7 + US-8 FAIL, three ZZTEST applications still “received” at the last read.

Verified today: REV-216 [Todo] p1 @narbotokerimov, upd 13 Sep. It pairs with REV-217 (Nur, “QA digest must count skipped, missing and incomplete daily tests”, now p1, upd 13 Sep) — both sides of the daily are being rebuilt. The onboarding half got more urgent, not less: the REV-214/218/219 sequence tickets closed in the blind window with only the webhook-setup email (#429) visibly shipped, and Sean Finck is next to provision — a rehearsal on your own store is the only way anyone sees what the new sequence actually does before a merchant does. A Woo QA store would also be the reproduction bench for REV-226 (webhook timeouts cancelling paid Woo orders).

Next action

Stand up one Woo store and one custom-integration store on your own account (Narboto inc / getrevion-capsule.com), move every $-test off Halo, then run one adversarial onboarding end to end and file what it breaks.

Verified today: REV-93 [Todo] p1 @narbotokerimov, last updated 31 Aug — fourteen days. The funnel around it moved: server-side encrypted drafts + edit freeze through save/submit shipped 14 Sep (#432/#434), so applicants no longer lose everything on refresh — but a phone-scanned PDF still dies with a generic error. REV-186 (“store CDR rejection codes + actionable error messages”, p2) and REV-183 (“sanitizer rejections shown as a generic error with no reason logged”, p2) are still open and unassigned. Aadil is about to send Apex Peptide Supply through this door.

Next action

Reproduce with a genuinely phone-scanned PDF, capture the CDR rejection code, and post it on REV-93 — that code is exactly what REV-186 needs in order to show an applicant something actionable.

Open dailies in today’s Linear read: REV-215 (11 Sep, p1) · REV-209 (7 Sep) · REV-207 (6 Sep) · REV-206 (5 Sep) · REV-202 (3 Sep) · REV-181 (30 Aug) · REV-180 (29 Aug) — seven, spanning two weeks; REV-220 (12 Sep) has left the open list. No 13 or 14 Sep daily is visible. REV-98 (“validate merchant dashboard production release 5565719”) has been [In Review] since 2 Sep, and REV-60 (the plan) was edited 11 Sep. REV-217, the digest counter that would make skipped/missing/incomplete runs visible, survived the triage at p1 (upd 13 Sep).

Next action

Close or explicitly mark-skipped every daily older than 9 Sep, one line each, so REV-217’s counter starts from a true baseline instead of inheriting a two-week backlog.

The 13:06Z bundle: nine Slack sections “check fly auth”, six watch-merchant reads empty (“machine/auth?”), the DB summary dead on the same raw JSONDecodeError, TG unavailable — identical to 12 Sep, and the last full read remains 2 Sep. The working half (Linear over HTTPS, /api/marks on Cloudflare Pages, local git) still needs neither binary, so the flyctl/node PATH diagnosis from the 12 Sep board stands, untested. New today: the marks fetch asked for day=2026-09-12 — the day-detection advanced past the 10 Sep board, and it also means no board published on 13 Sep. Marks for the 12 Sep board came back read and empty; the 7–9 Sep marks backfill has still never been evidenced, so three boards’ ticks remain unread.

Next action

flyctl auth whoami + machine state by hand, then an explicit PATH export (node + flyctl) at the top of collect.sh and generate.sh — or plist EnvironmentVariables — and a FORCE=1 re-collect. Then fetch marks for 2026-09-07/08/09, and log an unavailable-sections count per run so a blind bundle is loud.

Merges into production since 1 Sep: #373, #378, #402, #421, #423 — and now #426 (13 Sep) plus #430, #433, #435, #437, #439 (all 14 Sep), carrying the /apply drafts + freeze, the NMI reversal inbox and the webhook-setup email. The pipeline itself changed underneath in the same window: prod-skip decided by last released tag instead of previous commit, parallelized frontend deploys with skip-unchanged apps/clamav/cdr, parallel verify jobs behind one named gate with a timeout, clamav health-check wait matched to its 5m budget (#427), a selector test plus a documented “accepted sandbox baseline tradeoff”, and the Woo plugin update excluded from the verification candidate (#428). That much active CD work implies the pipeline is being exercised — but implication is not a release log: no bundle since 28 Aug has shown flyctl releases output or a release tag.

Next action

flyctl releases per app + release tag (the same auth the collector needs — one sitting fixes both); then the Playwright checkout E2E gate on the live sandbox channel against the CHIPS binding. Never place a payment on sandbox.revioncaps.com — prod keys.

RHP’s first real sale (2 Sep 05:08Z) and Halo’s 2 Sep morning order passed unwatched. Next: Casey’s Biopep retry, Chris (Zen), the Best Peptide embed call on the 10 Sep build, Sean Finck once provisioned — watch-merchant.sh <mrc> during the order, confirm each step, void after. Gated on the same Fly access as the collector: all six watch-merchant reads came back empty again today. If REV-226’s “Apex” turns out to be a real merchant, his first watched order joins this list at the front.

File one labelled Charge inquiry via the sandbox public support endpoint and confirm the alert email carries both admin links. (Claude’s live POST was classifier-blocked — needs a human-run curl or Tim’s go.) Worth pairing with Nur’s failing support_admin_alert item: the same send path is what would carry it.

The 31 Aug chases drew two replies (Jarrod — tax; Jenifer — thanks); Chris and Best Peptide stayed silent through 2 Sep. With Slack and the DB unreadable in every collection since, the sweep still has no fresh input — the collector fix comes first. The automated layer thinned in the meantime: REV-190 (idle reminder) closed unproven-armed and REV-210 (outreach) dropped to p3, so the manual sweep is carrying more, not less.