# Session summary, 2026-08-30 overnight: twenty-one agents on the updated queue, no exponent moved, the map moved

<!-- ledger
id: Q-session-0830-summary
status: ANSWERED
todo: none
question: What did the 2026-08-30 overnight campaign (rolling waves of up to ten agents on the queue rewritten after the 08-29 evening wave) change, and what is still open?
verdict: Neither exponent moved; twenty-one staging notes landed, all HELD; four routes closed (the skeleton door, the K*-product and fold-composition doubling bridges, the deep-ladder Buchstab transfer), TODO item 4 deleted and items D and 8 parked, the RML route carries an adverse derived floor whose growth law survived one sealed blind test and awaits a red team only Chris can call, Z2's fifth decade rejected the exact-Mertens main term for the omega-form crossing at 13 to 24 se, three record defects were verified on independent code and applied as riders, and a coherence pass corrected 27 lags in the live layer.
-->

*(Orchestrator's summary. Records: the twenty-one notes named below, all in
`research/history/staging/`, every producer bit-honest under `embed.js
--check`; `research/history/CHANGELOG.md` 2026-08-30 entry; `TODO.md` as
rewritten; `research/REFUTED.md` rows 96-98. Gate at close: see §4.)*

## 0. The afternoon red team (added at close)

Ten red-team notes, one per group of records, all on independent code, all
verdicts applied (`CHANGELOG.md` 2026-08-30 afternoon entry). The ruling that
matters: the adverse floor's exact half is PROVEN and its growth half survived
six attacks, so item 0's REC route is a TRUTH GAP at rung
derived-and-red-teamed-once and is now `REFUTED.md`'s last row; the programme
holds no live analytic route to a two-class exponent below 4.26645 that is not
a two-class count. About forty sentences corrected across the two days'
notes, four of them the orchestrator's own live-layer sentences and one an
applied rider; one derivation under a REFUTED row refuted with its conclusion
surviving as a measurement. `TODO.md` now opens with a ranked board.

## 1. What is still open, first

- The exponent gap 4.26645 → 2 is unchanged. Every note is HELD; no red
  team ran (rule: only on Chris's ask).
- **Item 0 is closed as a route at rung derived-and-red-teamed-once** (§0);
  reopening needs a second adversarial pass to break the growth half. The
  blind slope test is ruled to test the construction, not the law.
- Box-class items awaiting Chris's go: @43 at 579 laptop-hours
  (`engine-0830-at43-bigint.md`), the exact sup at z = 41 (4.8 to 5.4 h here, eleven segments priced for the
  box, `measure-0830-rho-sup-z41.md`), @29 anchored ladder, β(37), 47#.
- Held for Chris: `paper/variance-note.md` §10's X2 correction; the two
  SEARCH-CONVENTIONS rows each of `recon-0830-rec-killrun.md` and
  `recon-0830-smooth-aps.md` propose; the HELD tag on REFUTED row 96.

## 2. What closed (REFUTED rows 96-98, plus the two of 08-29)

| route | record |
|---|---|
| Skeleton Equidistribution door (item 4 deleted) | `decide-0830-skeleton-door.md` |
| K*-product doubling bridge at every legal C₂ | `attack-0829n-doubling-bridge.md` |
| per-fold composition for the doubling kill-run (truth gap) | `attack-0830-doubling-killrun.md` |
| deep-ladder Buchstab transfer, item 8(a) | `attack-0830-buchstab-deep.md` |
| D* law in Q for the parity adversary | `attack-0829n-parity-dstar.md` |

## 3. What moved, by item (each line calibrated at its note)

- **0**: REC's cheapest point does not close (PARTIAL); adverse floor on
  file (§1); prior art finds nothing that applies and no Maier truth-gap
  result at level D (`recon-0830-rec-killrun.md`); the u_sup convention
  adjudicated, same object, correction applied to both notes
  (`verify-0830-usup-convention.md`).
- **1**: @43 priced, box-class; the "~13×" had no source (45.5 strike
  ratio, 2.23 per cell).
- **1d**: δ's sign unreadable at reach 79 by six pre-registered readers
  (`measure-0830-delta-reader.md`); all-bases form unattackable.
- **A**: first move (a) was stale (ANSWERED 08-21); @23 done under a sealed
  14-of-14 forecast (`attack-0830-anchored-ladder-17.md`).
- **D**: parked on 0c after both bridges closed.
- **Z2**: fourth decade 64 HIT / 12 MISS; fifth decade sealed, ω-form
  CONFIRMED over exact-Mertens at 13 to 24 se on 17,702 blind anchors;
  riders on `u2-engine-depth.md` and `import-rough-anatomy.md`.
- **Z4**: ensemble halves of head and tail derive exactly, anchored halves
  are HL-strength; the 5.3% head remainder is a pooling artefact (VERIFIED,
  HL now 2.5 to 3.4% below the half-decade measurement); "R + 1/2" amended
  under the tail's convention.
- **Z5**: residual of Kourbatov's b located in the under-dispersed twin-gap
  law at height, 96 to 100% in a conjectural null, not height-stationary.
- **8**: (a) closed, (b) measured per prime for the first time; parked.
- **9**: the two open steps are one; Henriot Cor. 2 applies (read at the
  text layer); the X2 defect confirmed to every digit; the smooth-numbers
  search finds nothing that applies as written, nearest Harper 2012.
- **Live layer**: 27 coherence findings applied (`coherence-0830.md`), among
  them per-fold L moved out of G2-STATE §0's PROVEN list, 67 → 71 routes,
  "8 hits" → ten of ten, "27 clauses" → 25.

## 4. Gate and custody

`node research/qc.js --full` green at each integration step and at close
(see the CHANGELOG entry for the sequence); `QUESTIONS.md` regenerated after
each landing; every landed id on its item's `Ledger:` line. One agent
disclosed a single read-only `git status`; no other git command by any
agent; nothing committed.