Morning Brief · ATLAS Standup
Monday, August 3, 2026 — published 8:08 AM ET · on schedule (7:53 AM ET cron, 15 min build)
·
all briefings →
⏰ 22 minutes until standup (8:30 AM ET)
✔ Nightly PM fired on time. 2026-08-02 11:05 PM ET — second on-schedule nightly in a row,
so the weekend sleep-burst did not recur. /todos/daily/2026-08-02.md was written
(2 shipped · 3 moved forward · 4 blocked) and todos.json was rewritten 11:40 PM ET.
No silent-failure pattern this cycle.
What shipped overnight completed 2026-08-02 ET · 2 items
- untagged
LYNCH Master Combinatorial Backtest published — 3,618 combinations, walk-forward validated.
P1 · claude
- untagged
BOAT v4.6 filter + cost sweep benched on a2 — efficiency gate, OR%, per-symbol-day cap, TZ fix.
P1 · claude
Both landed off-board — neither carries a project tag, so they show up in no
project Gantt and in no sprint. That is the same tagging gap flagged on 8/1. Sunday work, no trading session.
Decide today blocked on James · 6 days each · one is directly P&L
Blocked · P0 · james · 6 days
Demote revbreak_long from the validated set
Setup League has it confidently negative live: n=259, 18% win, −0.46 avgR, 90% lower bound −0.54.
Friday 7/31 added 6 more signals for −0.7R. This morning’s Week Ahead page still lists it as a live
play for today. Every day it stays validated it fires again and costs money — the only open item
on the board that is directly P&L.
Blocked · P0 · james · 6 days
Pick a SIP top-10 rebalance option
Catalyst scores 21.6–36.0 pts vs mover 1.5 / rrg 2.2 / liquidity ≤10 — so PVP, PBS,
GAPMA and LYNCH technical hits mathematically cannot reach the top 10. Today’s Week Ahead is the
proof: all 20 prepped names are ordered catalyst-first. Blue-line stays blocked until one of the 3 options is chosen.
Stale blockers owner ≠ claude · pending · P0/P1 · age ≥ 7d · 29 total, top 12 shown
| Age | Project | Owner | Pri | Title |
| 108d | Atlas | user | P0 | indicator_alerts DDL audit: dump schema, commit as ground truth |
| 107d | MultiTrader | user | P0 | USER: 3 steps to finish journal persistence |
| 107d | MultiTrader | user | P1 | Re-run gen_scorecard.py ON a2 (or via tunnel) and confirm enrichment populates |
| 104d | Atlas | user | P1 | Deploy z_activity_scanner.py on each Windows machine via Task Scheduler |
| 103d | MultiTrader | user | P0 | Run scientific_layer migrations 001–009 via psql tunnel — critical-path step 1 |
| 89d | Atlas | user | P1 | Cron /system/update_freshness.py to keep freshness strip auto-refreshing |
| 89d | StratLib | alex | P1 | Codify Range-Break strategy rules from Alex’s Notion (+ 2 more Range-Break tickets) |
| 89d | Classifier | roberto | P1 | Live Classifier HUD: auto-refresh classify/<SYMBOL>.json during market hours |
| 88d | Acct-v2 | jimbo | P0 | BUG: Risk Bump Progress bar ignores the selected date range (always 5-day rolling) |
| 82d | Acct-v2 | jimbo | P0 | BUG: Ticket Meter denominator doesn’t scale with date range (1901 / 30 on a 13-day MTD range) |
| 82d | Atlas | james | P0 | Compliance: delete or hide historical MGP pages 4/17–5/12 (still contain Vital Knowledge text) |
| 44d | AMS | user | P0 | Seed main.users with trader rows so identity_fn returns non-zero · + resolve decisions D1–D7 |
Unchanged from Friday and from Saturday’s brief — 29 for the 6th consecutive publishing day.
The two Acct-v2 date-range bugs are step 2 of the critical path and both sit with Jimbo.
Freshness alerts 2 of 3 monitors are themselves dead
⚠ The freshness panel cannot be trusted this morning. Two of the three monitors have not
written in weeks, so their “stale” verdicts describe May and July, not today.
| Monitor | Last written | Verdict | What it means |
/system/freshness.json | 2026-05-07 · 88d | DEAD |
Retired monitor. Its payload still claims live_classifier stale, nightly_pm stale,
backtests missing — all frozen snapshots from May 5–7. Ignore these three alerts.
Reviving it is an 89-day-old P1 (cron update_freshness.py). |
/system/app-status.html | 2026-07-08 · 26d | RED |
The stat-probe watchdog that replaced freshness.json. Last checked 7/08 12:01 PM ET on
DESKTOP-AKU8I62 and has not run since — the 30-min watchdog is not firing on the HOME box. |
atlas-flow-status.json | 2026-07-31 7:00 PM ET | RED |
The only live monitor. 12 pass · 1 fail · 3 skip of 16. Failing step: V1 morning
(V2 midday + V2 recap passed). Weekday-only cron, so the weekend gap is expected — it should
fire again tonight at 7:00 PM ET. |
Real freshness signal this morning comes from the nightly PM note, not the panels:
pipeline ran, MGP editions published, scorecards caught up 4:45 PM ET Sunday.
Today’s standup focus open P0/P1 ranked by priority × age
⚠ The formula returns a 6-way tie. Six Phase-0 foundation P0s are all
108 days old and score identically at 324. Age-weighting no longer discriminates — these six have
outranked everything else every morning for weeks and none has moved. The three below are the ones the
critical path and last night’s PM ledger both point at.
Focus 1 · score 324 · P0 · 108d · unblocks the most
Atlas 1/6 Data-freshness dashboard — /system/freshness.json + banner on hub
Owner claude · ~2h · step 3 of the critical path. This morning is the argument for it:
two of three monitors are dead and the brief had to reconstruct freshness by hand from file mtimes.
Completes the Phase 0 freeze on the property side.
Focus 2 · score 324 · P0 · 108d
Atlas 2/6 Pipeline orchestrator — run_nightly.py calls every generator, logs to one place, fails loud
Owner claude. Directly addresses the V1-morning FAIL that has been sitting RED in atlas-flow
since Friday with nobody notified.
Focus 3 · score 321 · P0 · 107d · 15 minutes of work
MultiTrader USER: run scientific_layer migrations 001–009 via psql tunnel
Owner James · 15 min. Retires 4 user 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 board
and it has been step 1 of the critical path for 103 days.
Critical path · next step from /system/sprint-plan.html
Step 1 — James runs the scientific-layer migrations 001–009 via psql tunnel
Still the first unblockable item, unmoved for 103 days. user · 15 min.
Retires 4 user P0s and unblocks p0-daily-summary-view, the auto-classifier rule-based path,
and identity_fn DB writes.
- Step 2 — Jim ships the 2 Acct-v2 date-range bugs (jimbo · 2 P0s, 88d + 82d).
Until then no metric on a multi-day view can be trusted, which blocks any team rollout.
- Step 3 — Atlas ships p0-freshness-dashboard MVP (claude · 2h) — Focus 1 above.
- Step 4 — Classifier refresh cron (roberto · 4h), 89d stale.
- Steps 5–7 — Exec Trigger Engine → Backup warm-runner → Gr8Trade adapter in prod.
Steps 1–2 are both human-owned and both untouched. Nothing downstream can start,
which is why the same three Atlas P0s keep winning the focus ranking.
Sprint reality check board vs calendar
⚠ The board still says Sprint 1. active_sprint_id = 1 — a sprint whose window
closed 2026-06-26, 38 days ago, at 0 of 8. The calendar says Sprint 5
(2026-07-27 → 2026-08-07): “Live Gr8Trade execution + StratLib expansion.”
Today is day 6 of 10 working days — 5 working days left, and this morning’s sprint-status brief
reported 0 of 6 committed items started. Sprint 5’s goal is real Gr8Trade orders behind conservative
caps, which is step 7 of a critical path still stuck at step 1.