Task: Health review 1 — sprint 23 analysis

Mission: "Continue commissioning ores.refdata entities and codegen C++ drift correction." Two distinct threads, scored separately.

Roughly half the sprint's DONE stories sit outside either thread entirely: CRM, synthetic-data-collections, ir-rates-synthetic, consolidate_history_dialogs, badge-colour-scheme, account-default-party, data-librarian-UI, clean-up-menus, extract-quant-processes, plus 4 hotfixes. Sprint.org itself is candid about this — most are tagged "pulled in from product backlog next" — so this isn't wishful mapping, it's a genuinely broad, opportunistic sprint rather than a mission- focused one. That is a legitimate choice, but it means "goal alignment" undersells how much of the sprint's volume the two named threads actually account for.

At 137 commits/day this sprint is over 4x the commit-load target, and has been for its entire span (lowest single day was 32, on day 9 so far — a plausible wind-down, but every prior day bar one exceeded the whole-sprint target on its own). This is not a late-sprint spike; the volume has been consistently high since day 1 (84 commits on day 1 alone). Combined with running 2 days past planned close, "still adding scope" and "already over budget" are both true at once.

Throughput is healthy and cycle time is fast when PRs do merge — the one open PR (#1638) is tied to a task already marked PASSED in the worktree, so it reads as "about to land" rather than stuck. One long- open PR against an otherwise excellent record is an AMBER signal per the stated threshold, not a red flag on its own.

Overall DONE/(DONE+STARTED+BACKLOG) across all 41 sprint stories = 22/41 = 0.54 → at the GREEN threshold. Every STARTED story is decomposed into tasks (none are black boxes). The outlier is consolidate_history_dialogs, open 45 days across three sprints — that alone should disqualify a pure ratio-based GREEN read.

Eight stories in flight simultaneously, spanning four largely unrelated themes. Two of the eight (calendar, IR rates) have tasks already PASSED and are close to closing — see the Recommendations below — which would bring live focus down to 6, still at the RED boundary.

Overall: RED. Two dimensions are independently RED (sprint load at over 4x the commit target sustained all sprint, and focus with 8 concurrently STARTED stories across four unrelated themes), which is enough to call the sprint RED on its own; three more are AMBER. The single most important concern is that the sprint is both over its commit budget and two days past its own planned close without a close decision having been made — the natural forcing function that would otherwise resolve the focus and drift problems hasn't fired. The one thing that would most improve health right now is closing the sprint: land the two near-done stories (calendar, IR rates), apply the close-and-split recommendation below to the two DONE-heavy stories, and correct the five sprint.org drift rows as part of that close — rather than continuing to add scope on day 9 of a 7-day sprint.

Per request: stories where the bulk of the tasks are done should be closed now, narrowing their acceptance to what shipped, with the remaining tasks spun out into a new small "snags"/cleanup story rather than keeping the parent open indefinitely. Two clear candidates:

• Commission: party, counterparty, and party_status (12/14 tasks DONE, 86%). The 2 remaining tasks — investigate-counterparty-orchestration and generate_short_code_qt_button — are both independent follow-ups, not blockers of what already shipped (party/counterparty/ party_status Qt UI, manual chapters, all verified post-NATS). Close this story now; move the 2 tasks to a new small story.
• Retire legacy codegen profile system; add junction support (9/14 tasks DONE, 64%). The legacy-profile deletion and junction codegen (the story's stated acceptance) are done and validated end-to-end. The 5 remaining tasks — table-pathway backcompat removal, migrating remaining junctions to generated C++, retiring the table model type, and the mustache→address rename — read as a distinct "finish the migration" follow-on, not core to this story's own acceptance (sprint.org's own description already says as much for a sibling batch of tasks). Close this story now; move the 5 tasks to a new story.

Two more are borderline, not recommended yet:

• Model calendars as proper ORE Studio reference data — core entity modelling (4 tasks) is DONE and the Qt-screens task is PASSED with PR #1638 open; once that merges this becomes an 5/8-done candidate for the same split (holiday-picker widget, export mapping, DQ publish as the follow-on story).
• Improve badge colour scheme support (4/8, 50%) — not yet DONE- heavy enough to recommend closing; revisit at the next review.

Not recommended for splitting: Consolidate history dialogs — its one remaining task is the bulk of the outstanding work (Phase C template rollout across ~61 dialogs), not a small tail, so closing early would just relabel the same open work rather than shrink it.