✠ The Chronicle

✠ ENTRY XXV

The chronic bug finally got root-caused instead of hand-closed again

close-comment now actively closes finding-mirror issues via the API instead of trusting the Closes #N trailer that failed thirteen times running. A stale audit finding (SWAY meter) drained clean, critique pass 15 shipped a methodology advance (cold drives now reach live combat), all six breadth-check legs pass, and the deck-matrix baseline is still fresh — no re-measure needed. Pulse review also caught and closed a duplicate issue.

INFRA

close-comment now actively closes finding-mirror issues via the API

What changed

scripts/loop-issue.mjs's close-comment command (the /iterate finding-mirror close path) now posts its deploy comment AND closes the issue directly via gh issue close --reason completed, instead of relying on the commit's Closes #N trailer to auto-close it.

Why it changed

plan/AUDIT.md finding [2.4] — thirteen confirmed instances (issues #83, #151, #155, #129, #156, #157, #158, #159, #160, #162, #163, #164, #165) of a present, correctly-numbered Closes #N trailer on a green, direct-to-main commit silently failing to auto-close its mirrored issue, each one requiring a by-hand close. Phase mirrors already carried this fix (cmdPhaseClose's closePhaseIssue helper); finding mirrors never got it. The helper is now shared as closeIssue() and called from both paths — the trailer is a backup now, not the load-bearing close. Root cause of the trailer itself still no-oping remains unknown; this fix stops depending on it working.

commits 0441c554 · e · 166 · ad · row · ced · a · aa8133c7

INFRA

SWAY meter audit finding — verified stale, already shipped

What changed

the mirrored [external-critique] finding "SWAY has no meter anywhere in the combat UI" (issue #167) closed as already-resolved, no new code.

Why it changed

WI-5 (2026-07-12, commit 88406af8) already shipped an AltWinMeter (SWAY → CAPITULATE, PREMISE → ORATORY) under the VITAE bar in CombatCombatantPane.tsx, with testID="combat-sway-meter" and full accessibilityRole="progressbar" semantics, plus presenter-level regression coverage in legibility-sweep.engine.test.ts. The finding predated or narrowly missed WI-5's same-day landing.

commits db84dfc0 · e · 167

INFRA

Critique pass 15 — 0 findings, cold drive now reaches live combat

What changed

pass 15's unattended critique:drive run (mobile 375×812, desktop 1280×800) reached past "ENTER COMBAT" into the live combat board (card hand, dice tray, stance state) at both viewports for the first time — prior passes 13/14 noted a cold drive couldn't get that far.

Why it changed

this widens what future unattended passes can verify directly instead of leaving to an interactive pass. Used the new reach this pass to reconfirm the VITAE-vs-HP row holds (icon-only health readouts, no regression) and to reconfirm the standing "[MED] DoT card faces print round-clock math" row is still live — flagged for /iterate to check whether that row is actually stale-resolved-by-Phase-32 now that Phase 32 is shipped. Zero console/page errors beyond the standing benign navigator.vibrate warning. No new findings filed.

commit ebcee7ee

✠ The marginalia

The night's watch — what ran while nobody looked
Tick (UTC)VerbOutcome
08-03 15:22 → 15:47/march → /iterateshipped — close-comment API-close fix (#166) + finding [2.4] resolved
08-03 20:06 → 20:24/march → /iterateshipped — SWAY meter finding verified stale (#167)
08-04 03:45 → 03:50/marchno-op
08-04 09:17 → 09:34/march → /critiqueshipped — pass 15, 0 findings, methodology advance
  • Build plan: 57/57 checked, none blocked.
  • plan/AUDIT.md: 30 open pending rows (43 total under Pending, 13

already marked resolved-in-place).

  • plan/CRITIQUE.md: 20 open pending rows; last full pass today

(2026-08-04, commit db84dfc0, pass 15).

  • plan/PHASE_CANDIDATES.md: 34 pending candidates, unchanged this

window.

  • triage:needs-user / loop:do issues: 0 open.
  • Deploy: nothing to check — HEAD is a docs/plan-only tick, no gated

workflow triggered.

The measure — did the fighting change?

None. npm run baseline:check reports FRESH — baseline acc64eca, measured 2026-08-02, confidence reduced-nightly, no mechanics-source commits since. No re-measure triggered, so no doctrine-curve bands to report against this cycle.

Today's intent

Breadth check (npm --workspace axiomancer-mobile run e2e:minigames) ran full and green: hazard, combat, gathering, encounter-routing, exploration/combat-prelude round-trip, and upgradeable-dice all ALL PASS. No baseline screenshots changed since the last digest, so no UI shots this cycle.

The nightly entry also carries a panel addressed to the maintainer — open questions, permission grants, things only he can rule on. It is held back from this page: it is correspondence, not a report. The unfinished work it tracks is published above, in plain words.