Open Tasks
CRITICAL
[BUILD] 5 G-fixes for dormant-subsystem silent-drift class of failure — prevention machine controls
CRITICAL
[DECISION] GDE path forward — (A) kill / (B) harvest composite_score extraction only / (C) full sweep-revised v2 plan
CRITICAL
[BUILD] SIL Output Quality + Audit Trail Closure — execute Streams 2 + 1
CRITICAL
[BUILD] Wire SHADOW gate-accuracy into Gate Alternatives composite_score (7th component)
HIGH
[DECISION] GDE — kept (HIGH-priority v2) or deprioritized? Unpack BUILD_gde_v2_remaining_gaps_20260424.md + figure out f…
HIGH
[BUILD] deploy_mobile.py DESKTOP_COPIES — derive from mobile_registry.py + add health CHECK for desktop_dashboard file p…
HIGH
[INVESTIGATE] Why does FORMAL recs (-7.9%) underperform Gate-Filtered Random (+26%) by ~33pp?
HIGH
[BUILD] Batch 2 of A22 master plan — silent-fail freshness CHECK for non-blocking pipeline steps
Latest Changelog
## 2026-05-23 — Pipeline run after ratio pull paste — exit 0; data_date=5.23.26
**What:** User pasted ratio output into Sheets, re-exported SM(56) + Holdings(53), dropped both into `data/raw/`, ran full pipeline via `py metric_discovery_engine/scripts/run_pipeline.py`. Exit 0 at 2026-05-23 19:05. Older SM(54)/H(51) auto-archived by Step 1. Mobile deploy live: https://33a4b897.claudemobiledashboards.pages.dev (49 files, 4.09s).
**Numbers:** data_date=5.23.26 | 10 buys | 31 recs | 36 metrics | 5,457 combos | 500 control | 461 watchlist tickers re-evaluated | 95 held tickers re-tiered | 28A/12B/10Dual/4Elite watchlist | sell=35 add=7 alloc=17.
**Files touched:** none (pipeline-only run).
**Closes:** none.
---
System Health
0 errors · 10 warnings
All warnings: hardcoded dist weights — expected