You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[Due for payment 2026-09-30] [$250] [Exploratory] Report - Error appears on the report when Cancel Payment #96345
Preconditions: User has set up a Business Bank Account and a Personal Bank Account in the account and submitted, approved, and paid a report in Workspace chat
Open the NewDot app
Navigate to the Workspace chat
Locate the paid Report
Click to open detail -> More button -> Cancel Payment
Expected Result:
The user can cancel the payment
Actual Result:
"Auth CancelReimbursement returned an error" appears
Root Cause: This is a backend/internal issue, not an App bug. The string Auth CancelReimbursement returned an error does not exist anywhere in App source. When a user clicks More → Cancel Payment, App calls the CancelPayment API command from cancelPayment() in PayMoneyRequest.ts, passing only iouReportID, chatReportID, managerAccountID, and reportActionID — no bank account identifier. The backend (CancelPayment → Auth CancelReimbursement) fails and returns that error message, which App simply displays. A separate internal investigation is needed to find why CancelReimbursement fails for this flow.
Recommended next step: Because the fix is outside the App repo, this should be escalated to an internal engineer with access to the backend (Auth / Web-Expensify). There is no App-side code change to make here.
Likely related (same error, recurring):
Expensify/App#91804 — "Admin receive error if cancel payment for employee with BA in wallet" (labeled Internal, closed 2026-07-13 as not reproducible — never actually fixed)
Expensify/App#57715 and Expensify/App#33540 — same Auth CancelReimbursement error reported in prior years
The recurrence over multiple years, and the fact that the near-identical #91804 was labeled Internal (requires API changes), both reinforce that the root cause lives in the backend.
Investigation details & evidence
App-side (confirmed, high confidence):
grep -rni 'Auth CancelReimbursement' App/src/ → no matches. The error text is not hardcoded in App.
App's own client-side failure copy for cancel payment is the generic iou.error.other → "Unexpected error. Please try again later.", used only when no backend response is received (e.g. offline). It is not the reported string.
Backend mechanism (hypothesis, unverified): The reported precondition is that the user has both a Business Bank Account and a Personal Bank Account. Since App sends no bank-account identifier, any ambiguity in which account CancelReimbursement resolves is a backend concern. I could not confirm the exact backend failure reason (e.g. jsonCode / missing reimbursement record / bank-account state) because production log access (VictoriaLogs) was unavailable in this run. An internal engineer should query the backend logs for the CancelPayment command and the CancelReimbursement failure for the affected report to confirm the true mechanism.
Next Steps for Contributor+ team:
This is an internal/backend issue — there is no App code fix to propose. Please escalate to an internal engineer with backend access to investigate the CancelReimbursement failure. To refine this analysis, reply with @MelvinBot <your feedback>.
🚨 Edited by proposal-police: This proposal was edited at 2026-07-18 02:01:05 UTC.
Proposal
What is the root cause of that problem?
"Auth CancelReimbursement returned an error" is the backend refusing CancelPayment for an ACH reimbursement that is no longer cancellable (the daily NACHA batch is submitted at 23:45 UTC). The app has a check in isCancelPaymentAction() meant to hide Cancel payment in exactly that case, but it fails for two reasons, so the option is offered and the request errors:
1. The NACHA cutoff is computed in the device's local timezone instead of UTC.action.created is a DB datetime string in UTC with no timezone suffix (e.g. 2026-07-17 20:00:00.123 — see getDBTimeFromDate, which builds it from toISOString()):
I verified in Node that for a device in America/Los_Angeles, a payment created 2026-07-17 20:00:00 (UTC) parses to 2026-07-18T03:00:00Z, so the cutoff is computed as July 18 23:45 UTC — one day late. The real batch left July 17 at 23:45 UTC. For any device behind UTC, every ACH payment made after (24 − offset):00 UTC (after ~9:45 AM local in Pacific time) keeps Cancel payment visible through the entire next day; selecting it yields the reported Auth error. This matches the repro: the report is paid as a precondition, and the tester cancels in a later session. The existing unit tests missed this because they feed created: tomorrow.toISOString() — a Z-suffixed ISO string that new Date parses correctly — while real report actions use the suffix-less DB format, and Jest pins TZ=utc.
2. A missing pay action is treated as "paid elsewhere". While the expense report's actions are still loading (e.g. right after a fresh app open), payActions is empty, isPaidViaBankAccount is false, and an APPROVED + REIMBURSED bank-paid report takes the paid-elsewhere early return — showing Cancel payment with no processing-state or cutoff checks at all. Selecting it in that window sends CancelPayment and hits the same backend rejection:
What changes do you think we should make in order to solve the problem?
In isCancelPaymentAction():
Parse the pay action's created as UTC, following the repo's existing convention for DB datetimes (fromZonedTime(dateTime, 'UTC'), as used in DateUtils.subtractMillisecondsFromDateTime):
consthasDailyNachaCutoffPassed=payActions.some((action)=>{constpaymentDatetime=fromZonedTime(action.created,'UTC');// was: new Date(action.created)constcutoffTimeUTC=newDate(Date.UTC(paymentDatetime.getUTCFullYear(),paymentDatetime.getUTCMonth(),paymentDatetime.getUTCDate(),23,45,0));returnnowUTC.getTime()>cutoffTimeUTC.getTime();});
Return false when no pay action is loaded, instead of assuming a payment made elsewhere — until the pay action is available we cannot distinguish a bank payment from an elsewhere payment, so we should not offer an action the backend may reject:
if(payActions.length===0){returnfalse;}
This does not regress the case the current code was written for (PR #69731, showing Cancel payment for cancellable bank payments): paying from the app writes the optimistic PAY action to the expense report's actions synchronously, so payActions is populated immediately, and MoneyReportHeaderSecondaryActions recomputes reactively via usePaginatedReportActions once server actions load.
Update the tests: the paid-elsewhere test should seed a real ELSEWHERE pay action (today it passes only via the unsafe empty-actions fallback); add a test that a reimbursed report with no loaded pay action does not offer Cancel payment; and add a past-cutoff test using the real DB datetime format (yyyy-MM-dd HH:mm:ss.SSS, no Z) — no existing test covers the past-cutoff branch, and the Z-suffixed timestamps used today are what masked this bug:
I implemented and verified all of the above locally: the full ReportSecondaryActionUtilsTest suite passes (164/164) with lint and typecheck clean. Note that the timezone defect itself cannot be pinned by a Jest regression test — the runner forces TZ=utc and Jest workers ignore runtime process.env.TZ changes (verified) — which is why the new tests pin the behavior contract (past-cutoff bank payments hide the option) using realistic DB-format timestamps.
What alternative solutions did you explore? (Optional)
I considered only guarding the empty-payActions case (the earlier version of this proposal). That closes the loading-window hole but not the reported scenario: once actions load, the timezone-shifted cutoff still offers Cancel payment for up to a day after the batch left for any user behind UTC. I also considered handling the API error more gracefully (a translated message instead of the raw Auth error), but that only softens the failure; hiding the action when it cannot succeed fixes it at the source, and the raw-error cosmetics can be a separate follow-up.
Please re-state the problem that we are trying to solve in this issue.
On a report that was paid via a bank account, opening the report details and selecting More > Cancel Payment fails with Auth CancelReimbursement returned an error instead of cancelling the payment.
What is the root cause of that problem?
The Cancel Payment option is gated by isCancelPaymentAction, whose final condition is isPaymentProcessing && !hasDailyNachaCutoffPassed:
An ACH payment can only be recalled before the daily NACHA cutoff. Once the batch is submitted, Auth's CancelReimbursement refuses the request, and that refusal is what surfaces as Auth CancelReimbursement returned an error. So the error is emitted by the back end, but the reason the user ever gets to trigger it is that hasDailyNachaCutoffPassed computes the cutoff from a misparsed timestamp and reports false when the cutoff has in fact already passed.
The pay action's created is a database timestamp — a UTC time rendered as yyyy-MM-dd HH:mm:ss.SSS, with the T and the Z stripped off:
Because that string has a space separator and no timezone designator, new Date(action.created) does not parse it as UTC — the engine falls back to its lenient parser and interprets the wall-clock time as local time. The resulting Date is therefore shifted by the viewer's UTC offset, and hasDailyNachaCutoffPassed then reads getUTCFullYear() / getUTCMonth() / getUTCDate() off that shifted value to pick which day's 23:45 UTC cutoff to compare against. A shifted instant can land on a different UTC calendar day, so the gate is built against the wrong day entirely.
For any negative-offset timezone — which matches the reported Windows 10/Chrome device — a payment made late in the UTC day resolves to the next UTC day, so the cutoff is computed as 23:45tomorrow rather than 23:45 today. hasDailyNachaCutoffPassed then stays false for roughly 24 extra hours, Cancel Payment keeps being offered long after the batch went out, and every click on it is rejected by Auth. The same defect misfires in the other direction for positive-offset timezones: a payment made early in the UTC day resolves to the previous UTC day, so the cutoff reads as already passed and Cancel Payment is hidden while it is still legitimately cancellable.
Note that new Date(x.created) is used elsewhere in the codebase, but those call sites only subtract two similarly-parsed timestamps from each other, so the identical offset cancels out and the delta stays correct. This call site is different: it derives an absolute UTC calendar day from a single timestamp, which is exactly the operation the offset corrupts.
What changes do you think we should make in order to solve the problem?
In hasDailyNachaCutoffPassed in src/libs/ReportSecondaryActionUtils.ts, parse the pay action's created as the UTC timestamp it actually is, instead of letting it be interpreted as local time. DateUtils already establishes the correct pattern for reading a database timestamp — it converts through fromZonedTime(dateTime, 'UTC') rather than handing the raw string to the Date constructor:
Applying the same treatment here makes paymentDatetime the true instant of payment, so the year/month/day it contributes to cutoffTimeUTC is the real UTC day the payment happened, and the 23:45 UTC cutoff is anchored to the correct day for every viewer regardless of their timezone.
The surrounding comparison does not otherwise need to change. nowUTC is already rebuilt from now's own UTC fields and is equivalent to the current instant, so once paymentDatetime is parsed correctly the two sides of the comparison are consistent and the gate closes exactly at 23:45 UTC on the payment's own day. With the cutoff evaluated correctly, Cancel Payment stops being offered once the ACH batch has gone out, so the user no longer reaches the state where Auth rejects the request — and it remains available for the full window before the cutoff for positive-offset users who are currently losing it early.
What alternative solutions did you explore? (Optional)
Adding a general-purpose "parse a database timestamp" helper to DateUtils and routing this call site through it. This is arguably the more durable shape given that the same raw new Date(x.created) idiom appears at several other call sites, but those sites are all offset-cancelling deltas and are not currently broken, so introducing a shared helper here would widen the change well beyond the bug without fixing anything additional today.
Leaving the gate as-is and instead translating the raw Auth CancelReimbursement returned an error text into a friendlier message. This would only mask the symptom — the user would still be offered an action that cannot succeed, and would still be unable to cancel the payment, which is the actual complaint in the issue.
📣 @nikonrz001-pixel! 📣
Hey, it seems we don’t have your contributor details yet! You'll only have to do this once, and this is how we'll hire you on Upwork.
Please follow these steps:
Get the email address used to login to your Expensify account. If you don't already have an Expensify account, create one here. If you have multiple accounts (e.g. one for testing), please use your main account email.
Get the link to your Upwork profile. It's necessary because we only pay via Upwork. You can access it by logging in, and then clicking on your name. It'll look like this. If you don't already have an account, sign up for one here.
Copy the format below and paste it in a comment on this issue. Replace the placeholder text with your actual details.
Format:
Contributor details
Your Expensify account email: <REPLACE EMAIL HERE>
Upwork Profile Link: <REPLACE LINK HERE>
Activity
applause-bot commentedon Jul 16, 2026
While Applause doesn't embed attachments for this bug, here are the links for quick view:
https://platform.applause.com/services/links/v1/external/ec1fd31869b7b398427ec63e3b9fb82f8e3a6294d23e113ec93aaf637244c346
MelvinBot commentedon Jul 17, 2026
Issue Analysis
Root Cause: This is a backend/internal issue, not an App bug. The string
Auth CancelReimbursement returned an errordoes not exist anywhere in App source. When a user clicks More → Cancel Payment, App calls theCancelPaymentAPI command fromcancelPayment()inPayMoneyRequest.ts, passing onlyiouReportID,chatReportID,managerAccountID, andreportActionID— no bank account identifier. The backend (CancelPayment→ AuthCancelReimbursement) fails and returns that error message, which App simply displays. A separate internal investigation is needed to find whyCancelReimbursementfails for this flow.Recommended next step: Because the fix is outside the App repo, this should be escalated to an internal engineer with access to the backend (
Auth/Web-Expensify). There is no App-side code change to make here.Likely related (same error, recurring):
Expensify/App#91804— "Admin receive error if cancel payment for employee with BA in wallet" (labeled Internal, closed 2026-07-13 as not reproducible — never actually fixed)Expensify/App#57715andExpensify/App#33540— sameAuth CancelReimbursementerror reported in prior yearsThe recurrence over multiple years, and the fact that the near-identical #91804 was labeled Internal (requires API changes), both reinforce that the root cause lives in the backend.
Investigation details & evidence
App-side (confirmed, high confidence):
grep -rni 'Auth CancelReimbursement' App/src/→ no matches. The error text is not hardcoded in App.iou.error.other→ "Unexpected error. Please try again later.", used only when no backend response is received (e.g. offline). It is not the reported string.cancelPayment()→API.write(WRITE_COMMANDS.CANCEL_PAYMENT, ...).Backend mechanism (hypothesis, unverified): The reported precondition is that the user has both a Business Bank Account and a Personal Bank Account. Since App sends no bank-account identifier, any ambiguity in which account
CancelReimbursementresolves is a backend concern. I could not confirm the exact backend failure reason (e.g. jsonCode / missing reimbursement record / bank-account state) because production log access (VictoriaLogs) was unavailable in this run. An internal engineer should query the backend logs for theCancelPaymentcommand and theCancelReimbursementfailure for the affected report to confirm the true mechanism.Next Steps for Contributor+ team:
This is an internal/backend issue — there is no App code fix to propose. Please escalate to an internal engineer with backend access to investigate the
CancelReimbursementfailure. To refine this analysis, reply with@MelvinBot <your feedback>.mukhrr commentedon Jul 17, 2026
🚨 Edited by proposal-police: This proposal was edited at 2026-07-18 02:01:05 UTC.
Proposal
What is the root cause of that problem?
"Auth CancelReimbursement returned an error" is the backend refusing
CancelPaymentfor an ACH reimbursement that is no longer cancellable (the daily NACHA batch is submitted at 23:45 UTC). The app has a check inisCancelPaymentAction()meant to hide Cancel payment in exactly that case, but it fails for two reasons, so the option is offered and the request errors:App/src/libs/ReportSecondaryActionUtils.ts
Lines 423 to 462 in 4bafe13
1. The NACHA cutoff is computed in the device's local timezone instead of UTC.
action.createdis a DB datetime string in UTC with no timezone suffix (e.g.2026-07-17 20:00:00.123— seegetDBTimeFromDate, which builds it fromtoISOString()):App/src/libs/DateUtils.ts
Lines 356 to 368 in 4bafe13
Parsing it with
new Date(action.created)treats it as local time:App/src/libs/ReportSecondaryActionUtils.ts
Lines 453 to 459 in 4bafe13
I verified in Node that for a device in
America/Los_Angeles, a payment created2026-07-17 20:00:00(UTC) parses to2026-07-18T03:00:00Z, so the cutoff is computed as July 18 23:45 UTC — one day late. The real batch left July 17 at 23:45 UTC. For any device behind UTC, every ACH payment made after(24 − offset):00UTC (after ~9:45 AM local in Pacific time) keeps Cancel payment visible through the entire next day; selecting it yields the reported Auth error. This matches the repro: the report is paid as a precondition, and the tester cancels in a later session. The existing unit tests missed this because they feedcreated: tomorrow.toISOString()— aZ-suffixed ISO string thatnew Dateparses correctly — while real report actions use the suffix-less DB format, and Jest pinsTZ=utc.2. A missing pay action is treated as "paid elsewhere". While the expense report's actions are still loading (e.g. right after a fresh app open),
payActionsis empty,isPaidViaBankAccountisfalse, and anAPPROVED + REIMBURSEDbank-paid report takes the paid-elsewhere early return — showing Cancel payment with no processing-state or cutoff checks at all. Selecting it in that window sendsCancelPaymentand hits the same backend rejection:App/src/libs/actions/IOU/PayMoneyRequest.ts
Lines 745 to 755 in 4bafe13
What changes do you think we should make in order to solve the problem?
In
isCancelPaymentAction():createdas UTC, following the repo's existing convention for DB datetimes (fromZonedTime(dateTime, 'UTC'), as used inDateUtils.subtractMillisecondsFromDateTime):App/src/libs/DateUtils.ts
Lines 370 to 382 in 4bafe13
falsewhen no pay action is loaded, instead of assuming a payment made elsewhere — until the pay action is available we cannot distinguish a bank payment from an elsewhere payment, so we should not offer an action the backend may reject:This does not regress the case the current code was written for (PR #69731, showing Cancel payment for cancellable bank payments): paying from the app writes the optimistic PAY action to the expense report's actions synchronously, so
payActionsis populated immediately, andMoneyReportHeaderSecondaryActionsrecomputes reactively viausePaginatedReportActionsonce server actions load.ELSEWHEREpay action (today it passes only via the unsafe empty-actions fallback); add a test that a reimbursed report with no loaded pay action does not offer Cancel payment; and add a past-cutoff test using the real DB datetime format (yyyy-MM-dd HH:mm:ss.SSS, noZ) — no existing test covers the past-cutoff branch, and theZ-suffixed timestamps used today are what masked this bug:App/tests/unit/ReportSecondaryActionUtilsTest.ts
Lines 1702 to 1731 in 4bafe13
I implemented and verified all of the above locally: the full
ReportSecondaryActionUtilsTestsuite passes (164/164) with lint and typecheck clean. Note that the timezone defect itself cannot be pinned by a Jest regression test — the runner forcesTZ=utcand Jest workers ignore runtimeprocess.env.TZchanges (verified) — which is why the new tests pin the behavior contract (past-cutoff bank payments hide the option) using realistic DB-format timestamps.What alternative solutions did you explore? (Optional)
I considered only guarding the empty-
payActionscase (the earlier version of this proposal). That closes the loading-window hole but not the reported scenario: once actions load, the timezone-shifted cutoff still offers Cancel payment for up to a day after the batch left for any user behind UTC. I also considered handling the API error more gracefully (a translated message instead of the raw Auth error), but that only softens the failure; hiding the action when it cannot succeed fixes it at the source, and the raw-error cosmetics can be a separate follow-up.melvin-bot commentedon Jul 17, 2026
Triggered auto assignment to Contributor-plus team member for initial proposal review - @shubham1206agra (
External)wildan-m commentedon Jul 17, 2026
Proposal
Please re-state the problem that we are trying to solve in this issue.
On a report that was paid via a bank account, opening the report details and selecting More > Cancel Payment fails with
Auth CancelReimbursement returned an errorinstead of cancelling the payment.What is the root cause of that problem?
The
Cancel Paymentoption is gated byisCancelPaymentAction, whose final condition isisPaymentProcessing && !hasDailyNachaCutoffPassed:App/src/libs/ReportSecondaryActionUtils.ts
Lines 449 to 457 in d769fd5
An ACH payment can only be recalled before the daily NACHA cutoff. Once the batch is submitted, Auth's
CancelReimbursementrefuses the request, and that refusal is what surfaces asAuth CancelReimbursement returned an error. So the error is emitted by the back end, but the reason the user ever gets to trigger it is thathasDailyNachaCutoffPassedcomputes the cutoff from a misparsed timestamp and reportsfalsewhen the cutoff has in fact already passed.The pay action's
createdis a database timestamp — a UTC time rendered asyyyy-MM-dd HH:mm:ss.SSS, with theTand theZstripped off:App/src/libs/DateUtils.ts
Lines 356 to 358 in d769fd5
Because that string has a space separator and no timezone designator,
new Date(action.created)does not parse it as UTC — the engine falls back to its lenient parser and interprets the wall-clock time as local time. The resultingDateis therefore shifted by the viewer's UTC offset, andhasDailyNachaCutoffPassedthen readsgetUTCFullYear()/getUTCMonth()/getUTCDate()off that shifted value to pick which day's23:45UTC cutoff to compare against. A shifted instant can land on a different UTC calendar day, so the gate is built against the wrong day entirely.For any negative-offset timezone — which matches the reported
Windows 10/Chromedevice — a payment made late in the UTC day resolves to the next UTC day, so the cutoff is computed as23:45tomorrow rather than23:45today.hasDailyNachaCutoffPassedthen staysfalsefor roughly 24 extra hours,Cancel Paymentkeeps being offered long after the batch went out, and every click on it is rejected by Auth. The same defect misfires in the other direction for positive-offset timezones: a payment made early in the UTC day resolves to the previous UTC day, so the cutoff reads as already passed andCancel Paymentis hidden while it is still legitimately cancellable.Note that
new Date(x.created)is used elsewhere in the codebase, but those call sites only subtract two similarly-parsed timestamps from each other, so the identical offset cancels out and the delta stays correct. This call site is different: it derives an absolute UTC calendar day from a single timestamp, which is exactly the operation the offset corrupts.What changes do you think we should make in order to solve the problem?
In
hasDailyNachaCutoffPassedinsrc/libs/ReportSecondaryActionUtils.ts, parse the pay action'screatedas the UTC timestamp it actually is, instead of letting it be interpreted as local time.DateUtilsalready establishes the correct pattern for reading a database timestamp — it converts throughfromZonedTime(dateTime, 'UTC')rather than handing the raw string to theDateconstructor:App/src/libs/DateUtils.ts
Lines 370 to 374 in d769fd5
Applying the same treatment here makes
paymentDatetimethe true instant of payment, so the year/month/day it contributes tocutoffTimeUTCis the real UTC day the payment happened, and the23:45UTC cutoff is anchored to the correct day for every viewer regardless of their timezone.The surrounding comparison does not otherwise need to change.
nowUTCis already rebuilt fromnow's own UTC fields and is equivalent to the current instant, so oncepaymentDatetimeis parsed correctly the two sides of the comparison are consistent and the gate closes exactly at23:45UTC on the payment's own day. With the cutoff evaluated correctly,Cancel Paymentstops being offered once the ACH batch has gone out, so the user no longer reaches the state where Auth rejects the request — and it remains available for the full window before the cutoff for positive-offset users who are currently losing it early.Compare branch: https://github.com/Expensify/App/compare/main...wildan-m:App:wildan/96345-cancel-payment-nacha-cutoff-utc?expand=1
What alternative solutions did you explore? (Optional)
Adding a general-purpose "parse a database timestamp" helper to
DateUtilsand routing this call site through it. This is arguably the more durable shape given that the same rawnew Date(x.created)idiom appears at several other call sites, but those sites are all offset-cancelling deltas and are not currently broken, so introducing a shared helper here would widen the change well beyond the bug without fixing anything additional today.Leaving the gate as-is and instead translating the raw
Auth CancelReimbursement returned an errortext into a friendlier message. This would only mask the symptom — the user would still be offered an action that cannot succeed, and would still be unable to cancel the payment, which is the actual complaint in the issue.50 remaining items
iwiznia commentedon Sep 29, 2026
$425 is almost double.. does $350 sound good?
melvin-bot commentedon Sep 29, 2026
📣 @nikonrz001-pixel! 📣
Hey, it seems we don’t have your contributor details yet! You'll only have to do this once, and this is how we'll hire you on Upwork.
Please follow these steps:
Format:
melvin-bot commentedon Sep 29, 2026
Triggered auto assignment to @mallenexpensify (
Awaiting Payment)melvin-bot commentedon Sep 29, 2026
Payment Summary
Resolving PRs:
Upwork Job
BugZero Checklist (@mallenexpensify)
mukhrr commentedon Sep 30, 2026
sounds fair to me
dmkt9 commentedon Sep 30, 2026
sounds good to me as well
mallenexpensify commentedon Oct 1, 2026
Payment Summary
Contributor: @mukhrr due $350 via NewDot
Contributor+: @dmkt9 paid $350 via NewDot
@dmkt9 can you please accept the job below? Please reply here and tag me once you have. Plz ping me in Slack too, I'm triaging an Upwork bug
https://www.upwork.com/jobs/~022105663194014117665
dmkt9 commentedon Oct 1, 2026
@mallenexpensify I have submitted a Price increase request here.
mallenexpensify commentedon Oct 1, 2026
@dmkt9 paid, summary updated above. payment was as a 'bonus' cuz of Upwork bug. Price updated to $350 above.
flaviadefaria commentedon Oct 1, 2026
Approved: @mukhrr due $350 via NewDot