Sprint Status · Monday 2026-08-10

Atlas Scrum Master · pre-planning brief. All times Eastern (ET). Source: todos.json (written 8/09 11:27 PM ET) + sprint-plan.html (generated 6/20, 51 days stale) + live disk mtimes.
TIME UNTIL PLANNING: 59 MINUTES
Planning starts at 8:30 AM ET. This brief was scheduled for 7:00 AM ET and published at 7:30 AM ET — 31 minutes after schedule, 59 minutes before the meeting.
Second consecutive on-time sprint brief. Sprint 5 closed Friday; Sprint 6 opens today.
⚠ Read this first — the sprint pointer is still wrong. todos.json reads active_sprint_id: 1. Sprint 1 ended 2026-06-26 — 45 days ago. Sprints 2–6 are all still marked queued, including Sprint 5, which ended last Friday, and Sprint 6, which started yesterday. sprint-plan.html has not been regenerated since 2026-06-20.
Consequence: every sprint boundary below was reconstructed from the calendar, not read from the system. This is the 7th consecutive brief to carry the fix. It is a 2-minute edit.
38
Done in Sprint 5 (7/27–8/07)
0 / 6
Sprint 5 scope closed
22
Done last week
16
Carry-over, Sprints 1–5
41
Open P0
12
Human P0s > 14 days
1 / 6
Sprint 6 items startable
7.9h
Ledger age

1Previous sprint review — Sprint 5 closed Friday (7/27 → 8/07 ET)

Tickets marked completed with a completed_at date inside Mon–Fri of the closing week (8/03–8/07), grouped by project codename. Sprint 5 was a 2-week sprint; week 1 (16 closes) was reported in the 8/03 brief.

22 closes — the best week in the 4-week series, and the third straight increase (11 → 9 → 16 → 22). Sprint 5 total: 38 tickets Mon–Fri across two weeks (40 including weekend work). Output is genuinely and repeatably up.
And zero of them were Sprint 5 committed scope. For the second sprint running, everything that shipped was unplanned. The week went almost entirely to the Swing Playbook Engine — 14 of 22 tickets: the v5.2 port, the leg-filter gate, a 344-signal census, an 810-combo hold-rule tournament, an ATR stop sweep, and a real ATR-floor bug caught inside its own output.
That work is not junk — arguably it is “StratLib expansion,” which is literally half the Sprint 5 title. It just carries no project tag and no sprint tag, so the plan cannot see it.
Tagging discipline has decayed. Of the 22 tickets closed last week, 21 carry no project field — only hud-prev-close-api is tagged (Acct-v2). On 6/20 all 361 tasks were tagged in one pass; nothing has enforced it since. The codenames below marked StratLib* are Atlas’s inference from the work, not a field in the ledger.
ProjectClosedTickets
StratLib*14
08-03 P1 claude · BOAT v4.6 falsified — EDT hard-code + no-minimum-stop faults found, verdict flipped
boat-v46-data-faults-falsify-2026-08-03
08-03 P2 claude · Setup League rebuilt on 858 resolved signals (7/07–8/03)
setup-league-rebuild-858-signals-2026-08-03
08-04 P1 claude · Adaptive Mean + Sigma Bands v5.2 TOOL PLAYBOOK + CRWD case + v6 Pine
amsb-playbook-crwd-2026-08-04
08-05 P1 claude · Swing Playbook Engine: faithful v5.2 Python port + CRWD validation gate PASSED
swing-playbook-engine-port-2026-08-05
08-05 P1 james · Ruling closed: Play 1 stop rule becomes a tournament axis, not a fix
swing-engine-stop-rule-ruling-2026-08-05
08-05 P1 claude · CRWD 7/31 case-study playbook: Day 1 Pullback to the 50 DMA
crwd-day1-pullback-playbook-2026-08-05
08-05 P2 claude · VWAP Band Research System index page
vwap-band-research-index-2026-08-05
08-06 P1 claude · Census emits BOTH stop candidates per signal (stop_pine_literal + stop_hold_low)
swing-census-dual-stop-columns-2026-08-05
08-06 P0 james · TradingView settings export — leg filter missed the reference signal by 1.6%
swing-engine-leg-filter-tv-export-2026-08-05
08-06 P0 claude · Swing engine gate CLOSED: v5.2 as written does not fire the reference trade — port is correct
swing-leg-filter-gate-closed-2026-08-06
08-06 P0 claude · Census scaled 30 → 150 names: 21 signals rejected as a base, rebuilt to 344
swing-census-v2-scaled-2026-08-06
08-06 P0 claude · Bug found + fixed: census stored RAW structural stops without the engine’s ATR floor
swing-census-atr-floor-bug-2026-08-06
08-06 P1 claude · Hold-rule tournament: 810 combos (162 hold-rule × 5 stop) over 344 signals
swing-hold-rule-tournament-v1-2026-08-06
08-06 P1 claude · ATR stop sweep: no static stop level beats holding with no stop at all
swing-stop-sweep-2026-08-06
Acct-v23
08-04 P2 james · HUD dev-tree edits: api/trader.php + hud/js/render.js + hud/index.html
hud-dev-edits-2026-08-04
08-06 P2 james · HUD Calendar Summary modal (daily / weekly / monthly) + trader.php + render.js
hud-calendar-summary-modal-2026-08-06
08-07 P1 james · HUD prior-day-close API (api/prev-close.php) + render.js wiring, dev and prod
hud-prev-close-api-2026-08-07
Atlas2
08-03 P1 claude · Atlas Flow expanded to a 16-step end-to-end pipeline verifier — GREEN 16/16
atlas-flow-16-step-verifier-green-2026-08-03
08-03 P2 claude · First on-time sprint-status brief in six weeks — published 7:23 AM ET
sprint-status-brief-on-time-2026-08-03
SIP2
08-04 P2 claude · Stockbee 20% breadth: Pine up/down scanner pair + HOW_TO wiring TV to indicator_alerts
stockbee20-breadth-pine-2026-08-04
08-04 P2 claude · Daily SIP published as CSV (latest + permanent per-day archive)
sip-daily-csv-export-2026-08-04
Exec1
08-03 P1 claude · J3P Robot Report — 38 simulated trades, 5 gate variants, robot +$2,509 vs live +$24.11
j3p-robot-report-2026-08-03

Weekend work (8/08–8/09), outside the Mon–Fri window but inside the sprint:

DateProjectTicket
2026-08-08AtlasMGP deliverability probe — DMARC quarantine test + 3 mail-tester sends
mgp-deliverability-probe-run-2026-08-08
2026-08-09ClassifierHTF Portfolio playbook — Boat + VCET Pine forks, strategies.json registration
htf-portfolio-forks-shipped-2026-08-09
2026-08-09AtlasEmail transport migrated to authenticated SMTP on both notifiers; mailer.php demoted
notify-smtp-transport-migration-2026-08-09
2026-08-09StratLib*LYNCH Master combinatorial backtest re-run — winner unchanged across two sweeps
lynch-master-refresh-2026-08-09

2Previous sprint slippage — what carried over

Sprint 5’s six committed items, per sprint-plan.html, matched to their tickets in the ledger.

0 of 6 closed. 6 of 6 slip. Two of the six were never startable inside the sprint at all (the Gr8Trade chain, the Range Break chain), and one was vacuous by construction.
ProjectCommitted itemPriStatusOwnerWhy it did not close
ExecEnable Gr8Trade auto-execution in prod with conservative caps (1 share, 1 order/day, 3-symbol whitelist)
exec-broker-adapter-gr8trade
P1pendingclaudeNever started. Its own chain (Trigger Engine → Alert Channels → Risk Gate) is 3 unstarted tickets deep.
ExecDaily review of every executed order — calibrate caps weekly
(no ticket)
–pendingclaude+jimVacuous — zero orders were executed, so there was nothing to review.
StratLibAlex Range Break walk-forward backtest at /backtests/range-break/
alex-range-break-backtest
P1pendingalexBlocked on alex-range-break-rules (also pending, also Alex). Alex has 3 open tickets and 0 closes on record.
StratLibRegister Range Break in /strategies/ Library + canonical catalog
alex-range-break-register
P1pendingalexDownstream of the backtest above.
MultiTraderFirst Trader-of-the-Week session with Nolan
phase-4-trader-of-the-week
P1pendinguserSprint plan already deferred it “until classifier reliable”. Classifier refresh cron never shipped.
MTFVPRoberto: TradingView ticket-in handler (Path B)
mtf-vp-tradingview-handler
P1pendingclaudeOwner in plan is Roberto; owner in ledger is claude. Ownership was never reconciled.

Older carry-over — tickets still carrying a sprint tag from a sprint that has ended:

SprintWindowTaggedDoneStill openDetail
Sprint 16/22–6/2680845 days past its end date. All 8 still pending — 3 AMS, 2 Acct-v2 P0 bugs, freshness dashboard, Phase-3 doc, backup mirror.
Sprint 37/05–7/117255 open: J3P webhook smoke test, label polish, v2.3 Trend rebuild, SIP watchlist deploy + TV autosync.
Sprint 47/12–7/255322 open: exec-backtest-j3p-plays (in_progress since 6/26) and exec-hotkeys-j3p-setups.
Sprint 57/26–8/081011 open and blocked: exec-kore-live-j3p — waiting on firm-confirm Gr8Script call signatures.
16 tagged tickets are still open from sprints that already ended, on top of the 6 above. Sprint 1 is the worst: 0 of 8, 45 days past its end date. Note also that only 21 tickets in the whole 534-ticket ledger carry a sprint field at all — sprint tagging stopped on 6/29.

3This sprint’s goal

Sprint 6 · Multi-broker + AMS Phase 5 promotion
2026-08-09 → 2026-08-29 ET (3 weeks) per sprint-plan.html. todos.json states 8/10 → 8/28 — a 1-day discrepancy at each end; the two sources have never been reconciled.
Theme: scale broker coverage and complete the AMS migration. Ready for external collaborator onboarding.
Stated goal: “Scale broker coverage and complete AMS migration.”
This goal is not executable as written. Sprint 6 was scoped on 6/20 assuming Sprints 2–5 had completed. None of them did. Its first item needs a second broker; the first broker does not exist. Its AMS items need Phase 1 sign-off; that was Sprint 2 scope, 6 weeks ago.
Exactly 1 of the 6 items can be started today. The room has to choose between re-scoping the sprint and renaming it — see question 2.

4This sprint’s backlog

The six Sprint 6 items from sprint-plan.html, each matched to its ticket in todos.json with live status, owner and estimate. Readiness is Atlas’s call from the dependency chain in §5.

ProjectBacklog itemStatusPriOwnerEst.Ready?
ExecIBKR adapter — second broker, same pattern
exec-broker-adapter-ibkr
pendingP2claude5 dBLOCKED
AMSAMS Phase 5: migration + polish. 30-day green window starts
(no Phase-5 ticket exists)
pendingP0×3claude + user10 d+BLOCKED
AMSCutover: smb.a1prospect.com becomes canonical (D1 commitment honored)
ams-decisions-d1-through-d7
pendingP0user1 d after decisionBLOCKED
MultiTraderPer-trader monthly view (extend /james-monthly-reports/ to ?user=)
monthly-per-trader-view
pendingP2claude3 dRISK
AtlasNightly PM auto-updates task statuses from daily reports
p0-orchestrator
pendingP0claude2 dREADY
StratLibUser strategy upload + admin approve queue
strategy-upload-ui
pending–claude4 dBLOCKED
Estimates are Atlas’s, in ideal engineering days, and exclude blocked time. Summed as written the sprint is ~25 days of work into a 15-working-day sprint — before any of the blockers are cleared. 4 of 6 items are BLOCKED, 1 is at RISK, 1 is READY.

5Dependencies — what blocks what

No ticket in the ledger has a depends_on field — the schema does not have one. Every dependency below was read out of free-text note bodies. That is a structural weakness: dependencies are invisible to the Gantt, invisible to the cockpit, and only surface when a human reads the note.

TicketProjectReady?Blocked by
exec-broker-adapter-ibkrExecBLOCKEDTicket note is explicit: “Depends on Gr8Trade adapter being live and stable for 2+ weeks.” Gr8Trade is not written. Earliest honest start: September.
(no Phase-5 ticket exists)AMSBLOCKEDPhase 5 assumes Phase 1 sign-off, which was Sprint 2 scope and never happened. Hard blockers still open: ams-migration-002-remembered-devices (P0, 51 d), ams-seed-users-from-main (P0, user, 51 d), p0-identity-fn (P0, 115 d).
ams-decisions-d1-through-d7AMSBLOCKEDD1 is the canonical-domain decision. It has sat on James’s desk 51 days. Nothing downstream can be scheduled until D1–D7 are answered.
monthly-per-trader-viewMultiTraderRISKNote says “Depends on multi-trader-scorecards being parameterized first.” That is scorecards-multitrader-migration-2026-06-09 — in_progress since 6/09, dev only, plus phase-4-trader-selector (pending). Startable only if the parameterizer lands first.
p0-orchestratorAtlasREADYThe only Sprint 6 item with no upstream blocker. Nightly PM already runs and wrote /todos/daily/2026-08-09.html at 11:06 PM ET last night — it just does not write status back.
strategy-upload-uiStratLibBLOCKEDNote: “Blocked on phase-2_5 (add-user-role-approval).” Also needs strategy-library-ddl (main.user_strategies table) — both pending.
The longest chain in Sprint 6 is five deep and entirely unstarted:
classifier-refresh-cron → exec-trigger-engine-mvp → exec-alert-channels → exec-risk-gate → exec-broker-adapter-gr8trade → exec-broker-adapter-ibkr.
IBKR is Sprint 6’s headline item and sits at the far end of six unstarted tickets, the first of which is owned by Roberto. Its own note asks for Gr8Trade to be live and stable two weeks first — which alone pushes IBKR past the 8/29 sprint end.
Cross-cutting blockers named in notes, outside the Sprint 6 chain:

6Critical path stack — sorted by readiness

The seven-step chain from sprint-plan.html, re-sorted by what can actually start this morning, plus a step 0 that costs two minutes. READY = no upstream blocker, owner identified.

0
Advance the sprint pointer: close Sprints 1–5, set active_sprint_id=6  READY
The 7th consecutive brief carrying this item. todos.json still says active_sprint_id: 1 and all six sprints except #1 are marked queued — including the two that have already ended. Until this lands, no sprint number on any page is real, and this brief has to reconstruct the sprint from the calendar.
claude · 2 min
1
James runs scientific_layer migrations 001–009 via psql tunnel  READY
Day 110. Nothing blocks it. Retires 4 user-owned P0s and unblocks p0-daily-summary-view, the rule-based auto-classifier path, and identity_fn DB writes. Highest unblock-per-minute item on the entire board.
user · 15 min
2
Jim ships the 2 Acct-v2 date-range bugs  READY
Risk Bump Progress ignores the selected range; Ticket Meter denominator does not scale with it. Until both land, no metric on any multi-day HUD view can be trusted — which blocks every team-rollout item in MultiTrader.
jimbo · 2 P0s, 95 d + 89 d
3
Atlas ships p0-freshness-dashboard MVP (property side)  READY
Escalated since last Monday: the existing freshness monitor has itself been dead for 33 days (app-status.json last written 7/08 12:01 PM ET) and nothing alerted. Its final reading already showed 2 properties RED.
claude · 2 h
4
Backup warm-runner online (heartbeat + delay/skip + liveness)  READY
No upstream dependency — 5 open tickets, all claude-owned, none started. Sprint plan requires this before Exec goes live: if domino-1 dies, jimdo must keep alerts firing. The 4 Atlas scheduler misses on record are the argument.
claude · 3 d
5
Classifier refresh cron ships (Roberto)  READY
classify/*.json must refresh every minute during RTH before the Trigger Engine can fire on live signals. Also newly relevant: catchup-zero-regime-playbook-tags-2026-08-03 reports the classifier writing 0 regime/playbook tags into every trader-day.
roberto · 4 h
6
Exec Trigger Engine MVP + Alert Channels in DEV  RISK
Ticket note: “Depends on classifier-refresh-cron shipping first.” Becomes READY the moment step 5 lands. Alerts to humans only — no orders.
claude · 1 wk
7
Exec Gr8Trade adapter enabled in prod with caps  BLOCKED
Three-deep: needs Risk Gate, which needs Alert Channels, which needs the Trigger Engine. Separately exec-kore-live-j3p is blocked on firm-confirm Gr8Script call signatures — an external dependency with no owner and no date.
claude+jim · 2 wks
Steps 0–2 total under 20 minutes of human time and retire more blocked work than the last three sprints combined. They have been the top of this stack for seven consecutive briefs. The constraint is not capacity — 38 tickets shipped last sprint. The constraint is that the critical path is not what anyone actually works on.

7Velocity trend

Count of tickets with completed_at inside each Mon–Fri window. Weekend closes excluded for comparability.

11
7/13–7/17
9
7/20–7/24
16
7/27–7/31
22
8/03–8/07
11 → 9 → 16 → 22. Three straight increases; last week was a series high and doubled the 7/20 trough. 4-week mean 14.5/wk, last week +52% above it.
The caveat that matters: committed-scope velocity over the same four weeks is 0, 0, 0, 0. The team is fast and the plan is not steering it. A velocity number that only counts unplanned work cannot be used to forecast a sprint — do not size Sprint 6 off the 22.

8Stale alarms

P0 tickets owned by a human and open more than 14 days, then property freshness > 48 h.

12 human-owned P0s are past the 14-day alarm. Three are past 110 days. Total human P0 age on the board: ~980 ticket-days. Two of these are compliance or data-integrity items, not features.
AgeProjectOwnerTicket
115 dAtlasuserindicator_alerts DDL audit: dump schema, commit as ground truth
p0-indicator-alerts-ddl
114 dMultiTraderuserUSER: 3 steps to finish journal persistence
journal-user-steps
110 dMultiTraderuserRun scientific_layer migrations 001–009 via psql tunnel — critical-path step 1
run-scientific-layer-migrations
95 dAcct-v2jimboRisk Bump Progress bar ignores the selected date range
acct-v2-bug-risk-bump-ignores-date-range
89 dAtlasjamesCompliance: delete/hide historical MGP pages 4/17–5/12 — they still contain stripped content
mgp-historical-pages-cleanup
89 dAcct-v2jimboTicket Meter denominator does not scale with the date range
acct-v2-bug-ticket-meter-ignores-date-range
87 dAtlasJamesMGP auto-fire skipped 5/14 mid-day + recap; 5/15 morning ran ~2h45m late
mgp-cron-skipped-5-14
79 dAtlasjamesNightly PM went dark 5/20–5/22 (3 missed fires) — fix wake/launch
nightly-pm-silent-gap-2026-05-20-to-22
51 dAMSuserSeed main.users with trader rows so identity_fn returns non-zero
ams-seed-users-from-main
51 dAMSuserResolve 7 open decisions D1–D7 — gates the Sprint 6 cutover
ams-decisions-d1-through-d7
50 dPinePBuserTradingView setup guide for James
pinepb-tradingview-setup-guide
50 dPinePBuserJames pastes finalized Pine script into TradingView + configures the 3 alerts
pinepb-deploy-to-james-tradingview
Two more cross the line tomorrow: sip-scoring-rebalance-decision and revbreak-long-demotion-decision, both P0, both blocked on James, both filed 7/28 — day 13. The revbreak evidence got stronger over the weekend (n 314→364, LB-90 −0.41→−0.43): the decision is more supported now than when it was filed, and it is still open.
⚠ The freshness monitor is the stalest thing we monitor. app-status.json was last written 2026-07-08 12:01 PM ET — 33 days ago, by a probe that is supposed to run every 30 minutes. It has missed roughly 1,580 consecutive fires and nothing alerted. /system/app-status.html still renders and still returns 200, so the hub reads as healthy while showing a month-old snapshot.
This is the same silent-failure class as the 4 Atlas scheduler misses and the nightly-PM dark weeks — a monitor that fails by going quiet.
SurfaceLast writtenAgeRead
app-status.json / app-status.html
App Freshness Monitor (30-min watchdog)
2026-07-08 12:01 PM ET33 daysThe watchdog that is supposed to catch stale properties is itself the stalest thing on the property. Its last reading (7/08) already flagged playbooks RED at 38.7 h and Param Surface (DB) RED at 340.7 h — nobody has seen a fresh number since.
/system/sprint-plan.html
The document this brief is required to read
2026-06-20 7:48 PM ET51 daysStill shows Sprint 1 as ACTIVE. Every sprint boundary in this brief had to be reconstructed from the calendar.
freshness.json
Superseded property-freshness feed
2026-05-07 9:38 PM ET95 daysKnown dead and replaced by app-status — but it is still linked from the hub and still returns 200, so it reads as live.
drift-report.json
Drift report
2026-05-16 11:13 AM ET86 daysNo scheduled writer. Effectively abandoned.
focus.json
Focus surface
2026-06-17 4:58 PM ET54 daysNot refreshed since the codename registry was locked.
atlas-flow-status.json
16-step pipeline verifier
2026-08-07 8:41 PM ET2.4 daysWeekend gap only — last weekday fire was GREEN 16/16. The only live health signal on the property.
/todos/todos.json
The ledger
2026-08-09 11:27 PM ET7.9 hoursNightly PM fired on time (11:06 PM ET) and wrote /todos/daily/2026-08-09.html. Every count in this brief is trustworthy.

9Pre-meeting questions

Five questions the room should answer before committing Sprint 6. Each has a decision attached — none is rhetorical.

  1. Is the sprint model actually being used — or should we drop sprint numbering and run a rolling P0 board? Evidence for dropping it: active_sprint_id has been wrong for 45 days across 7 briefs; sprint tagging stopped on 6/29 (21 of 534 tickets); committed-scope closure over four weeks is 0, 0, 0, 0 — while 38 tickets shipped. The plan is not steering the work and has not for six weeks. Either fix the pointer this morning and commit to grooming it weekly, or stop pretending and manage the P0 board directly. The one outcome to avoid is a 7th sprint that nobody updates.
  2. Sprint 6 as written needs Sprints 2–5 finished. Do we re-scope it to the dependency-clearing work, or keep the name and accept the slip to September? 1 of its 6 items is startable today. IBKR is six unstarted tickets deep and its own note demands two weeks of stable Gr8Trade first. AMS Phase 5 needs Phase 1 sign-off from Sprint 2. The 9/11 AMS Phase 5 target assumed no sprint slipped more than a week; four have slipped entirely. Re-scoping is the honest option — but say so out loud and move the 9/11 date with it.
  3. What is actually stopping the scientific_layer migration run — access, time, or a wrong ticket? Day 110. Estimated at 15 minutes. Retires 4 user-owned P0s and unblocks the classifier enrichment path, p0-daily-summary-view, and identity_fn DB writes. It has been critical-path step 1 in every brief since June and has not moved once. A 15-minute task that survives 110 days is not a scheduling problem — the ticket is wrong, the access is missing, or it is not really the blocker we say it is. Answer which, today.
  4. Acct-v2 has 2 P0 date-range bugs at 95 and 89 days, both owned by Jim. Is Jim available this sprint, or do we reassign? Until they land, no metric on any multi-day HUD view can be trusted, which blocks every MultiTrader rollout item. Meanwhile Jim did ship three HUD tickets last week (dev edits, Calendar Summary modal, prior-close API) — so the capacity exists; it is going to new features while two P0 correctness bugs sit. Either Jim commits a day this sprint, or the bugs move to claude, or we down-grade them off P0 and say why.
  5. The freshness monitor has been dead 33 days and nothing alerted. Do we fix it before shipping anything new this sprint? Its last reading already had 2 properties RED. For 33 days we have had no ability to distinguish a stale property from a fresh one, and the hub renders green regardless. This is the fourth monitor in the system to fail silently (Atlas morning brief 8/07, nightly PM 5/20–5/22 and 5/26–5/28, MGP mid-day 5/28–5/29). The pattern is not bad luck — nothing in this system alarms on absence. Fixing p0-freshness-dashboard is 2 hours and is critical-path step 3.

Atlas recommendation for the room

#ActionOwnerCostWhy now
1Close Sprints 1–5, set active_sprint_id=6, regenerate sprint-plan.html claude2 min Seventh brief carrying it. Until it lands, no sprint number on any page is real.
2James runs scientific_layer migrations 001–009 user15 min Day 110. Retires 4 P0s and unblocks the classifier → Exec chain.
3James clears the two 13-day decisions (SIP rebalance, revbreak demotion) user10 min They hit the 14-day alarm tomorrow. The revbreak evidence strengthened over the weekend.
4Re-scope Sprint 6 to: freshness dashboard + backup warm-runner + classifier refresh cron + Trigger Engine MVP claude + robertothe sprint These are the four READY critical-path steps. They are what Sprint 6’s stated items actually depend on. Keep the multi-broker name only if the room accepts it means September.
5Add a depends_on field to the ticket schema and backfill the 6 chains in §5 claude1 h Every dependency in this brief was parsed out of prose. The Gantt cannot see any of them.
Items 1–3 are under 30 minutes of human time combined. Everything named in Sprint 6 that is not in item 4 — IBKR, AMS Phase 5, the cutover, the per-trader monthly view, the upload queue — rolls forward with an owner named out loud, or gets dropped on the record.