Wiki Operations Ledger

Authoritative next-work source. This file replaces the old next-actions surfaces for deciding what work is next on the wiki. Historical reports, ledgers, and wishlists remain evidence, but this file is the maintained operational ledger.

Use .claude/skills/ops-ledger/SKILL.md when refreshing this file.

Last reconciled: 2026-07-21 (targeted, second same-day — LI Vol. II ingest; OPS-018 done). Husserl's Logical Investigations Vol. II (Investigations III–VI + Appendix) ingested via the six-delegate Pass-2 pattern (896→906 pages; +1 source, +6 concepts — intentionality resolves the pinned stub —, +3 entities; 12 existing pages updated; 2 STRUCTURAL motifs minted — Fundierung, Evidenz — + the zu-den-Sachen BRIDGE entry updated). Both LI volumes are now primary sources; every Vol.-I "requires the Vol. II ingest" deferral is discharged. Six Part-D claim candidates now queue across the two LI extraction notes for the next audit Harvest (3 from Vol. I + 3 from Vol. II, one of the latter pending a Rule-18 PhP Fundierung check; plus 1 deferred material-a-priori seed). Audit-trigger check: 4 ingests since the v1.9 run (BGE, GM, LI I, LI II) — below the ≥10 bar, but the Harvest queue is now the richest since the Plato waves. No re-triage of other open items. Prior reconciliations below.

Prior reconciliation: 2026-07-21 (targeted — LI Vol. I ingest). Husserl's Logical Investigations Vol. I ingested (885→896 pages; +1 source, +7 concepts, +3 entities; 12 existing pages updated; 1 new BRIDGE motif). Three Part-D claim candidates queue in the extraction note for the next audit Harvest; OPS-018 added (LI Vol. II is in raw/ — the highest-leverage un-gated ingest); OPS-016's Husserl-leg note updated (no longer Crisis-only). No re-triage of other open items.

Prior reconciliation: 2026-07-11 (targeted — Harmonia R0/B7 harvest). The Atlas R0 truth spike's Undated-coda census (design-research/atlas-harmonia-r0-findings-2026-07-11.md) added OPS-016 (almanac-derived ingest wishlist, source-gated) and OPS-017 (Concepts-Developed under-listing backfill, execution-ready); State Snapshot counts refreshed to 885 pages / 303 claims. No re-triage of other open items.

Prior reconciliation: 2026-06-27 (OPS-sweep session). A single session discharged eight OPS items: OPS-002 (speculate-signoff batch 2 — 6 survivors adjudicated, 4 candidate claims + 2 holds), OPS-001 (non-Plato Phase-8 harvest — 12 candidate claims registered), OPS-014 (schema v0d.22 — page-body Connections may use free-form relation phrases), OPS-013 (23 non-Plato claim-status labels across 16 pages synced to register), OPS-005 (8 weave Pass-4 motif back-reference sections), OPS-010 + OPS-011 (monitoring-queue cadence review: IDL-1 sunset, IDL-2 discharged, MQ left as notes), and OPS-012 partially (CR-034 sensible-ideas reframe). OPS-006 / 007 / 008 / 009 triaged to disposition (no new work needed). OPS-003 was already done. Corpus grew 866 pp / 281 → 297 claims (4 jury + 12 Phase-8; candidate 110 → 126). Remaining open substantive work: OPS-004 (Plato silent-key integration) and the OPS-012 remainder (CR-035/045/047/048 + SA Priority-B) — both scoped below and execution-ready. Through the 2026-06-27 OPS-sweep log entries. Prior: 2026-06-27 (work session) and 2026-06-26 (deep pass).

Authority and Maintenance

This file is the single source of truth for "what action is next?"
Do not maintain wiki/.audit/next-actions.md or wiki/.audit/ingest-wishlist.md as living queues. They are historical inputs only.
Workflow-local ledgers still matter as provenance: wiki/.speculative/promotion/promotion-ledger.md, wiki/.speculative/promotion/ingest-dependency-ledger.md, and wiki/.speculative/promotion/methodological-queue.md preserve detailed source records. Their next-action status is summarized here.
Refresh this file after any ingest, audit, weave, speculate-signoff, lint cleanup, claim-status change, or explicit "what's next?" triage pass.
When a work item is completed, mark its ledger entry done, add the close date and evidence, and move a one-line note to "Recently Closed". Do not delete completion evidence during the same audit cycle.
Scope is wiki knowledge operations: ingest, audit, weave, speculate-signoff, lint/content hygiene, maintenance, gaps, and future source work. Site, build, dashboard, Atlas, and visual-check follow-ups are outside this ledger unless a repo-wide operations ledger is deliberately added.

State Snapshot

Corpus: 907 pages: 126 sources, 541 concepts, 216 entities, 24 questions (as of 2026-08-18; the 24th question page is the calibration's ascetic-priest filing, 2026-07-28).
Claims: 304 total: 23 supported, 142 live, 131 candidate, 3 contested, 5 retired (refreshed 2026-07-28 from wiki/index.md; was 303 / 144 live / 130 candidate / 1 contested, a snapshot that predated consolidation runs 3–7 — the two live→contested demotions at runs 3 and 4, and the run-7 C4 candidate). All 304 carry v0d.29 lifecycle lines (backfill 2026-07-11 plus forward emission); register 100% replay-complete.
Audit state: no active audit run. Last completed run: v1.9, 2026-06-23 (Plato-cohort comprehensive audit); a Phase-8 harvest ran 2026-06-28 (BGE+GM candidates, 297→303). Speculate-signoff: batch 2026-06-27-batch2 complete. Consolidation (OPS-015): built 2026-07-05; run 1 done 2026-07-14 (30 triaged, 0 promoted — refuter panel 12/12; 100 candidates remained); run 2 (date-key 2026-07-21, completed 2026-07-26 after a spend-limit salvage + 11-agent rerun) triaged 30 more + juried the first 5 supported-eligible live claims — 0 promotions, refuters 34/34 REFUTED across both gates; two Faul-class Rule-16 citation defects found in live entries and repaired. Runs 3–7 done (latest 2026-07-28, run 7): 0 promotions across all seven runs, refuters 146/146 REFUTED, 2 demotions (runs 3 and 4, both live→ contested); run 7 registered the C4 candidate that took the register to 304. Speculate-signoff latest batch: 2026-07-28-batch4. These markers are refreshed here only to the extent this review touched them — the per-run detail lives in the OPS-015 row and the digests at wiki/.audit/consolidation-*.md; a full snapshot rewrite is an ops-ledger pass, not a review-fix commit. Harvest queue: 6 LI Part-D candidates across the two LI extraction notes — 3 Vol. I (static-vs-instituted ideality; Prolegomena technician-seed; anti-naturalism valence-reversal) + 3 Vol. II (one pending the Rule-18 PhP Fundierung check), plus 1 deferred material-a-priori seed — clears the 4-candidate micro-Harvest bar; discharge at the next Harvest. (The formerly deferred Evidenz corpus-motif decision was discharged at the LI-Vol-II ingest: minted STRUCTURAL alongside Fundierung.) Ingests since v1.9: 4 (BGE, GM, LI-I, LI-II) — audit trigger (≥10) not met.
Phase: consolidation. The main remaining work is the OPS-015 discharge runs, the OPS-004 silent-key integration, the OPS-012 legacy-CR remainder, and ongoing capped maintenance (weave Pass-4 batches; the ~82-row Phase-8 reservoir; OPS-017 Concepts-Developed backfill).

How To Pick Work

Choose one entry from the ledger, read its named workflow skill, execute that workflow only, and update this file at close. The highest-leverage open item is OPS-015 (consolidation backlog discharge — /consolidate runs, v0d.24); the other open substantive items are OPS-004 (audit / targeted integration, ≤20-file batch) and the OPS-012 remainder (legacy CR triage). OPS-020 (the v0d.36 Case-3 supported-layer sweep + refuter-verdicts ledger) is done 2026-08-18 — run 8 is now unblocked and owes the calibration disclosure line. For a compact maintenance session, run the next OPS-005 weave Pass-4 capped batch (~140 home pages still lack the section). For the next ingest session, OPS-018 (LI Vol. II) is un-gated — the raw file is already present. This is a lightweight starting pointer, not a priority taxonomy.

Open Work

ID	Workflow	Status	Why it matters	Next action
OPS-001	audit Phase 8	done 2026-06-27	Non-Plato candidates were stranded in extraction notes.	12 D7 candidates registered; D4 re-test + ~82-row reservoir remain (see entry).
OPS-002	speculate-signoff	done 2026-06-27	Speculate survivors had no adjudication path past the v0 pilot.	Batch 2 done (4 candidate + 2 hold); ~21 v1.2-pilot + 33 run-1 survivors remain for future batches.
OPS-003	evidence maintenance	done 2026-06-27	CR-033 anchor precision.	All 6 checks resolved earlier 2026-06-27.
OPS-004	audit / integration	open	Plato Tier-B/C silent keys + L-set reciprocity remain in the report.	Execute the ≤20-file batch scoped below (Tier-B eudaimonist + new `eu-prattein.md`, 6–8 Tier-C, 6–8 L-set).
OPS-005	weave Pass 4	open (in progress)	`motifs.md` weight should be discoverable from home pages.	8 more done 2026-06-27 (102 → ~110 covered); ~140 HUB/STRUCTURAL homes remain — next capped batch.
OPS-006	weave Pass 2	done 2026-06-27	Motif residues needed home decisions.	All 4 triaged: 3 already-homed concept pages, 1 intentionally-distributed; 0 stubs needed.
OPS-007	weave Pass 3	blocked (confirmed)	Cross-tradition flux↔durée scan needs a missing page.	Still blocked: no Plato-side flux/heracliteanism page exists. Do not scan until one does.
OPS-008	motif candidates	waiting (confirmed)	BRIDGE status needs a 2nd cross-author anchor.	Still source-gated; one lead noted (Klossowski/Bataille ↔ sovereignty) for a future weave Pass 3.
OPS-009	future ingests	open (source-gated)	Many gaps need source acquisition, not maintenance.	0 ingestable now (Pinkard already done; other 10 high-rationale sources absent from `raw/`). Acquire then ingest.
OPS-010	ingest-dependency ledger	done 2026-06-27	HOLD-PENDING-INGEST entries should discharge or sunset.	IDL-1 sunset, IDL-1a strike-the-bridge codified, IDL-2 discharged.
OPS-011	methodological queue	done 2026-06-27	Process lessons should not leak into claims.	MQ-1/2/3 reviewed → left as notes (below 3-batch recurrence bar); MQ-4 logged.
OPS-012	Saint Aubert / legacy CR	partial (CR-034 done)	A few valid legacy CR items remain.	CR-034 reframed; CR-035/045/047/048 + SA Priority-B scoped below.
OPS-013	claim-label sync	done 2026-06-27	Stale inline labels misroute readers.	23 labels / 16 non-Plato pages synced to register (label-only, no `updated` bumps).
OPS-014	lint / typed-connection	done 2026-06-27	Out-of-list relation phrases were corpus-wide.	Schema amended (v0d.22): page-body Connections may use free-form phrases; closed list stays weave-Pass-3-only.
OPS-015	consolidate	open — runs 1–7 done (2026-07-14, 2026-07-21/26, 2026-07-26/27, 2026-07-27, 2026-07-27 run 5 — user-authorized second run of the date, completed 2026-07-28; run 6 2026-07-28; run 7 2026-07-28 — third run of the date, stacked on unmerged run 6)	0 untriaged candidates (C2 discharge completed at run 5 — every candidate now carries a KEEP-with-condition or block-with-objection) / 65 unjuried supported-eligible live claims (10 of the remainder precondition-screened at run 7; clean frontier starts at `hegel-arbeit-anchors-capital…` + `hegel-mature-recognition…`) / 9 remaining v1.2-pilot speculate survivors (batch 3 adjudicated 6 at run 5: 6/6 unanimous, 3 HOLD + 3 ROUTE → MQ-5..8, 0 promotions — first C4-embedded batch; batch 4 adjudicated 6 at run 7: 6/6 unanimous, 1 PROMOTE-CANDIDATE + 4 HOLD + 1 RETIRE, MQ-9..17; recount still an ops-ledger job); the register's statuses lag their evidence on the C3 side only. Runs 1–5: 130 triaged + 20 juried, 0 promotions (refuters 116/116 REFUTED under default-deny), 2 demotions (run 3: `deformation-coherente-mp-coinage` live→contested on the French Voix du silence attestation; run 4: `ambivalence-as-unconscious-replacement-predates-wahl-1951` live→contested/low on the in-corpus refutation of its priority thesis — SB 1942 raw 1686 + Inédits II raw 6577, whose extraction note had elided the lexeme; both corrective restatements in the Harvest reservoir). Run 4 was the first fully clean adjudication workflow (47/47, no salvage) and executed the Malraux locator pass. Run 5 (second run of 2026-07-27, user-authorized same-day; digest appended, artifacts `-run5`-suffixed): 27/27 clean workflow, the run-4 process-note post-apply grep adopted (caught 2 GM source-page II.18→II.19 sites beyond the reviewer's list), one Rule-16 misattributed-quotation repair (institution-as-middle-term's Halák bullet — repair-class, digest-prominent), and the first C4 speculate-signoff batch (jury notification-routing quirk documented in the digest: team-agent completions bubble to the session main thread, which relays phase transitions).	Run 6 done 2026-07-28 (branch `claude/consolidate-2026-07-28`, wf `wf_c4dced48-cbd`, 25/25 clean): C3-only as planned — 5 juried, 0 promotions (refuters 15/15 → 131/131 all-time), C1 queue all-keep (0 defects), C2 skipped (backlog dry, no state change since run 5), C4 skipped for usage conservation. Six Rule-16 citation repairs, all main-thread re-verified, incl. a false footnote attribution and commentator's-prose-as-MP; cite-back sweep fixed a stale label + a 3-site source misattribution on sein-zum-tode; new +16 pagination oracle for the Nature course-notes raw. Structural finding for the next run: all five entries' self-set promotion preconditions (recorded at creation 2026-05-05) were still unmet at 84 days, and the mechanical pre-gate does not see them — consider a pre-gate check that parses Status-History promotion conditions, which would have deprioritized all five. Run `/consolidate` again (v0d.24; run 7 — C3-only again, 70 eligible unjuried, 5-cap; or spend the run on the ~15-survivor C4 batch). Queued from runs 2–3: ~~the Malraux Voices pagination re-derivation pass~~ DONE at run 4 (per-locator via the volume's TOC/Synopsis/index; ~45 locators incl. the supported schema claim; French printed-folio annotation for raw 3762/4660 remains optional), the pbp-as-pivot revision set, the Derrida extraction-note "Corpus p. 40" hedge fix (extraction notes are outside consolidation write-scope), the corpus-wide Derrida raw-marker locator sweep — the run-3 +10 finding (raw {N} markers vs printed folios, verified against the raw TOC: §12 starts printed 265/marker {275}) generalizes beyond the repaired entry: `haptical-differance.md` cites "p. 240" ×8 (printed 230), `jacques-derrida.md`, `haptocentrism.md`, and `derrida-2000-on-touching-nancy.md` carry unconverted markers and stale "§13" refs for the §12 diagnostic — route to the next audit Citations phase as a named item (per-locator raw verification, not blanket −10) — now joined by the run-4 CPP marker-vs-folio (+16, index-corroborated) corpus sweep and the extraction-note fixes outside consolidation write-scope (Inédits II « – l'ambivalence – » elision at extraction lines 184/484; the CPP-extraction "Freud-Lacan" at line 224; the 4084 off-by-two copies), and the run-3 interpretive restatements queued per entry (death-of-god headline; bergson problem-space reformulation; onto-theo priority-tightening; ontologie-de-la-chair narrowing; inkpin Payoff carry-in). Scaffolding: verdict-conditional refuter schema + C3 apply-block emission LANDED at run 3 (the schema in field-description form — the API rejects top-level `allOf` in tool input_schema, the run-3 false-start lesson); `gen_adjudications.py` follow-ups from the PR #113 review: DONE 2026-07-27 (branch `claude/gen-adjudications-param-2026-07-27`; test-first, `scripts/tests/test_gen_adjudications.py`, 14 tests incl. a 2026-07-21 salvage byte-stability pin) — verdict-conditional C3 blocked-apply emission (discriminator = the 5-test sheet's explicit backticked restatus recommendation; fires on `deformation-coherente-mp-coinage` only across runs 2–3, emits the live→contested lifecycle apply), clean-run parameterization of `coverage_line` + the C2 "salvage" Status-History tags (salvage wording verbatim on the salvage path), coverage-sentence newline fixed (clean path), `evidence_bullet_counts` docstring refreshed, "regenerable at any time" reworded to name its actual inputs (committed verdicts JSON + date/c2/c3 slug lists — no session-dir files committed, per maintainer preference), the downstream REFUTED⇒`objection` / STANDS⇒`tried` gate validation added (generator aborts naming the seat), and `consolidation-2026-07-26-adjudications.md` regenerated so its provenance wording and deformation apply block match what was actually done; still queued: the nested-discriminator schema redesign; slot-enumerated rerun-merge coverage still queued (contingency tooling; run 3 hand-assembled its merge); speculate-survivor backlog recount queued (digest says both "~48 remaining, ledger batches 1–2" and "~54"; the run dirs' rival-pair duplicate H3s make the true count an ops-ledger job); new run-3 tooling carry-forward: add the Paper-A `dashboard_contested_claim_slugs` pin (fixture + `test_paper_a_a2.py`) to the `refresh_baselines.py` patcher list — it tripped hand-patched at run 3's first-ever contested-set change. Seat config runs 3+: all subagent seats Opus 5 — adjudicators `high`, `citation-verifier` `medium`, Fable main thread only. Run 7 done 2026-07-28 (third run of the date, branch `claude/consolidate-2026-07-28-run7` stacked on unmerged run 6, wf `wf_4199fe6a-fa7`, 25/25 clean): C3-only jury on the 5 oldest screened-clean eligibles — run 6's self-set-precondition finding implemented as a manual C0 screen (10 entries deferred with named conditions, incl. the whole 05-16 audit-14 cluster) — 0 promotions (refuters 15/15 → 146/146 all-time), headline Rule-16 find: a fabricated quotation in `bn-freedom`'s lead Evidence bullet (hero-recast of the raw's mother-analogy sentence at a fabricated extraction-note anchor; adjudicated repair-class — work real, locator real, provenance identifiable), struck and re-quoted with the reaffirming clauses restored; plus a new +12 pagination oracle for `raw/Merleau-Ponty and Contemporary Philosophy.md` (TOC-adjudicated over a dissenting verifier seat; the volume's whole extraction-note location column carries it — corpus sweep routed to next audit Citations), an edition false-friend (French Signes pp. 34–35 under the English McCleary wikilink, fixed at 7 sites incl. 2 extent-grep catches), a Bishop-inversion counterpressure, and a Foucault misattribution. First C4 batch with all seats Opus 5 incl. the lead (user directive): batch 4 = 6 survivors, 6/6 unanimous, 0 escalations — 1 PROMOTE-CANDIDATE registered (`promiscuite-as-topological-anerkennung-671`, ASCII slug, six-item pre-registration repair set executed; register 303→304), 4 HOLD, 1 RETIRE with a false-friend caution on pensee-de-survol (a second caution written to sein-zum-tode from the s2 HOLD was removed at the 2026-07-28 post-ratification review — HOLD's mapped action is a SPECULATIONS annotation, not a concept-page write; verdict retained); the C0 register-drift briefing changed an outcome (s3 near-duplicated run 6's same-morning repair); lead substantive-override near-miss on s6 caught post-shutdown by juror R, withdrawn, ledger carries a withdrawal record (no IDL-3); MQ-9…MQ-17 appended. 4 jury Harvest nominations for the next audit Harvest: s2 direct-contrast redraft (SA-2006 + Sein und Zeit); s3 subordination-with-preservation (SA raw 1507/1497, after the parent's conjunct separation); s4 reprise/lineage-placement (Madison App. II args 36–42, raw 2987–3168, extraction l. 209, at `candidate`) — the batch's most consequential loss if dropped; s6 four in-corpus Bachelard anchors (CPP ch. 2 §III; CPP chunk-B §VI.D ×2 sites; Inédits II extr. l. 486; SA-2006 raw 1618 context-only). Backlog after run 7: 65 unjuried C3 eligibles (10 precondition-screened; clean frontier starts at `hegel-arbeit-anchors-capital…` + `hegel-mature-recognition…`), C2 dry, 9 v1.2-pilot survivors for batch 5 (incl. the 2 now-unblocked Knight survivors + s4's deferred paired rival), ~82-row reservoir + the 4 nominations. Top tooling item unchanged: the self-set-precondition pre-gate check (manual screen cost ~15 entry reads this run). Run-8 pre-work landed 2026-08-18 (OPS-020): the supported tier is post-Rule-21-clean (23/23 re-verified, 0 restatuses), the refuter-verdicts ledger exists with the C5 hook live (append rows per run via `build_refuter_ledger.py --skeleton`, add the REGISTRY line same-commit, `--check` before commit), and ten new pagination oracles are registered. Run 8 owes the calibration disclosure line in its digest (jury-calibration-2026-07-28.md §7). Run `/consolidate` again (run 8).
OPS-016	ingest	open (source-gated)	The Atlas almanac's Undated coda names the corpus's own highest-leverage missing texts (B7: the wishlist IS the instrument reading).	Acquire one named text into `raw/` (top: The Gay Science), then run the ingest skill.
OPS-017	targeted maintenance	open	High-degree concepts sit Undated in ideas-time although their developing texts are already ingested — the source pages' `## Concepts Developed` sections under-list them.	Run one ≤10-file backfill batch (politics-mp, hinge, pensee-de-survol legs) where the source page's own body already asserts development.
OPS-018	ingest	done 2026-07-21	LI Vol. II was the highest-leverage un-gated ingest: intentionality, fulfilment, categorial intuition, parts/wholes, pure grammar.	Done same-day as Vol. I (branch `claude/ingest-husserl-li2-2026-07-21`): 10 new pages, 12 updates, 2 motifs minted, 3 Harvest candidates. Residual: the Rule-18 PhP Fundierung check (feeds `fundierung-mp-reciprocity-transform`) and the Crisis §§36–37 material-a-priori check — both belong to the next audit/Harvest session.
OPS-019	schema design (user-gated)	open (decision session)	The register's >2k-word tail (v0d.34 measurement: 200/303 over the retired marker, median 449w; 20 entries above 2,000w, 12 above 3,000w, topping at 4,002w) contains ~20 de-facto claim pages; the deferred `claims/` + `type: claim` extension now has its empirical case.	Paste the kickoff at `wiki/.audit/v0e-claims-extension-kickoff.md` into a fresh session with the user: decision + written design + at most a one-entry pilot. Not a build session.
OPS-020	supported-layer re-verification sweep	done 2026-08-18	Case-3 decision of the v0d.36 amendment: the 23 pre-jury supported entries swept rather than grandfathered.	Done (branch `claude/refuter-ledger-ops020-2026-08-18`, digest `wiki/.audit/reverify-supported-2026-08-18.md`): all 23 roster entries have digest rows; §6 items 2–3 discharged (16 repairs); item 4 adjudicated FIRST — the Capital formula pins to Mexico III printed p. 340 under a NEW Inédits II +4 oracle (twelve TOC anchors; the run's own first markers-as-folios derivation was caught by a second seat and superseded); 0 restatuses — every evidence-chain defect repairable, all 23 retained at supported with Updated bumps; item 6 discharged (g-w-f-hegel cite-back); item 5 (archive-bound-attestation policy) parked for a schema session. Companion record item done: `wiki/.audit/refuter-verdicts-ledger.md` (147 backfilled + 30 CAL rows = 177, zero STANDS ever; the familiar 146/146 undercounts run 1 by the rescope-harvested valery-eventuel-mode seat) + `scripts/build_refuter_ledger.py` + the consolidate-SKILL C5 forward-emission hook. Ten new pagination-oracle entries registered; corpus-wide locator sweeps (Inédits II column, GA 11 extraction, Signs) routed to the next audit Citations phase. Headline finds in the digest §3 (Chouraqui's own misquotation of V&I 179 inherited into the register; the false S&Z Brentano footnote; the AdV self-contestation quote that lives in I&P).
OPS-021	tooling / gate integrity	open	Two pre-existing fail-open predicates in `gen_adjudications.py`, surfaced by the PR #126 external deep review (2026-08-01) and deliberately kept out of that recovery PR: (a) `c2_gate`/`c3_gate` trust the reviewer's aggregate `summary: PROMOTE` without mechanically requiring every underlying per-test verdict to be PASS; (b) the C3 verifier leg checks `citations[]` row count against Evidence-bullet count, not row-to-bullet identity, so duplicate rows can satisfy the count while a bullet goes unverified. Same fail-open class as 10 of the 11 all-time Codex P1s — these predicates guard the claims register.	Small follow-up PR (~50–100 lines + tests): (a) add an all-sub-tests-PASS requirement to both gates with a new blocked-leg message; (b) default-deny approximation for row identity — require count ≥ bullets AND reject duplicate/near-identical `citation` strings (exact identity needs a bullet-index field in the workflow VERIFY schema — separate design decision, note it in the PR). Flagged risk: `TestSalvageByteStability` pins regeneration of the committed 2026-07-21 artifact byte-exactly; if any historical seat carried `summary: PROMOTE` with a CONCERNS sub-test, the tightened gate flips that claim's regenerated outcome and fails the pin — scan `wiki/.audit/consolidation-2026-07-21-salvaged-verdicts.json` for that configuration FIRST, and if it exists, adjudicate (historical outcome wrong vs. documented pin exception) before writing code. No claim status changes from this item itself; any historical-outcome correction routes through consolidation or explicit user confirmation (v0d.24).

OPS-001 — Non-Plato Phase-8 Harvest — DONE 2026-06-27

id: OPS-001
status: done
workflow: audit
closed: >
  2026-06-27. Registered 12 stranded D7 candidate claims (6 Signs + 6 Prose of
  the World) at candidate status (claims 285→297; candidate 114→126). All
  single-source → candidate-floor; NONE supported-eligible (supported HALT not
  engaged). Evidence anchored to the two MP extraction notes + raw line numbers
  (spot-checked). The two ledger-named Signs candidates
  (science-secretes-ontology-non-foundational-reciprocity,
  solus-ipse-dissolved-by-removing-its-origin) record the SPENT v1.9 Plato
  re-test in Counterpressure with explicit do-not-re-run.
residual:
  - The 4 already-registered D4 candidates need only a 2nd-source live re-test
    (expected to stay candidate; no non-Plato source since v1.8). Low priority.
  - The ~82-row corpus-wide harvest reservoir (synthetic-layer-2026-06-23.md)
    is a separate future Phase-8 pass; enumerate via harvest-claim-candidates.py
    first. Out of the named OPS-001 scope.

OPS-002 — Speculate-Signoff Batch 2 — DONE 2026-06-27

id: OPS-002
status: done
workflow: speculate-signoff
closed: >
  2026-06-27. Batch `2026-06-27-batch2` over 6 remaining v1.2-pilot survivors.
  7-role jury (interp + G/R/E/S; lead = main thread). Converged 6/6 in 2 debate
  rounds, 0 escalations, 0 lead overrides. 4 PROMOTE-CANDIDATE (claims 281→285):
  promiscuite-active-ignorance-as-the-unsaid-of-empietement,
  chiasm-constitutively-carnal-formal-extension-kills-the-concept (narrowed),
  ungrund-bestimmte-negation-coordinate-post-kantian-responses (renamed),
  flesh-three-positions-resolves-methodologically. 2 HOLD-AS-SPECULATIVE
  annotated in upstream SPECULATIONS.md. Supported HALT preserved.
residual:
  - ~21 remaining v1.2-pilot survivors + 33 run-1 survivors await future batches.
  - MQ-4 (resolve-vs-constitutive paired-rival shape) logged to methodological
    queue — review on 3rd-batch recurrence.

OPS-015 — Consolidation Backlog Discharge — OPEN (system built 2026-07-05)

id: OPS-015
status: open
workflow: consolidate
category: adjudication
source_of_truth:
  - .claude/skills/consolidate/SKILL.md
  - scripts/consolidation_pregate.py (run worksheet)
  - wiki/.audit/consolidation-*.md (per-run digests, once runs start)
why_it_matters: >
  Generation outpaced adjudication: 130 candidates (43% of register), 67
  mechanically supported-eligible live claims, ~54 speculate survivors, and
  the ~82-row Phase-8 reservoir all queue on per-decision human confirmation.
  v0d.24 (user-authorized 2026-07-05) replaced the supported HALT with the
  consolidation supported-jury + standing post-hoc user veto, so discharge
  no longer waits on the maintainer's presence.
next_action: >
  User triggers /consolidate; each run is capped (≤15 subtraction, ≤30
  triage, ≤5 supported, ≤1 C4 batch) and ends in one commit + a digest with
  per-change veto paths. Repeat until the pre-gate runs dry (~5–8 runs).
  OPS-001's D4 re-test and OPS-002's remaining survivor batches discharge
  through C2/C4 of these runs.
progress: >
  Run 1 (2026-07-14, digest wiki/.audit/consolidation-2026-07-14.md): 30
  oldest candidates triaged, 0 promoted — refuter panel 12/12 REFUTED under
  default-deny; objections written into Counterpressure; 100 candidates
  remain. C1 empty (0 defect rows). C3 empty by artifact: the v0d.29
  lifecycle-backfill markers reset the pre-gate settle clock
  (consolidation_pregate.py last_activity() reads all Status History dates,
  including "(lifecycle backfill 2026-07-11)" tails) — self-heals
  2026-07-25; the durable fix is to parse leading line dates + Updated:
  only, as claim_retirement_queue.py does. Batch-cursor rule adopted (PR
  #109 P2-A): every C2-triaged entry gets an Updated: bump, since the
  pre-gate sorts its candidate batch by Updated:.
  Run 2 (date-key 2026-07-21, digest wiki/.audit/consolidation-2026-07-21.md;
  completed 2026-07-26 after a monthly-spend-limit abort at 52/63 jury agents,
  journal salvage, and an 11-agent rerun as wf_68473faa-37c on Opus 5 @
  xhigh): pregate last_activity() fixed first (un-masking 95
  supported-eligible live claims); C1 empty again (second consecutive
  0-defect queue); C2 triaged 30 (19 gated — all REFUTED, stay candidate;
  11 single-source KEEP-with-condition); C3 juried the first 5 eligible live
  claims — all BLOCKED (no clean verifier sheet; refuters 15/15 REFUTED),
  stay live. 0 promotions across both runs. Register strengthened instead:
  34 verbatim refuter objections into Counterpressure, two Faul-class
  Rule-16 citation defects repaired in live entries, ~20 verifier-named
  locator/anchor corrections applied across claims.md + 5 pages. Queued
  follow-ups in the digest: Malraux Voices-side pagination re-derivation
  (collateral: one supported claim), pbp-as-pivot revision set.
  Run 3 (date-key 2026-07-26, digest wiki/.audit/consolidation-2026-07-26.md;
  completed 2026-07-27 after a refuter-schema false start — top-level allOf
  rejected by the API, all 28 refuter seats errored at spawn — a process
  exit, and an accidental manual kill; the 19 never-run refuter seats
  completed as wf_f4d3dab1-358, 19/19, prompts/config verbatim; jury 51/51;
  first run on the all-Opus-5 seat config): C1 empty (third consecutive
  0-defect queue); C2 triaged 30 (13 gated — refuters 13/13 REFUTED, stay
  candidate with would-promote conditions; 17 single-source
  KEEP-with-condition); C3 juried 5 — 0 promotions (no clean verifier sheet;
  refuters 15/15 REFUTED), 4 stay live, 1 DOWNGRADED live→contested
  (deformation-coherente-mp-coinage: the French Voix du silence attests
  "déformation cohérente" at raw 3762 and definitionally at 4660, refuting
  the Johnson coinage attribution; main-thread raw re-verification concurred;
  corrective restatement filed to the Harvest reservoir). 28 objections into
  Counterpressure; five Rule-16-class findings repaired (incl. the
  "Corpus p. 40" shell attribution + its 4-page propagation); 242 absence
  claims re-verified before apply (one refuter false-absence caught).

OPS-016 — Almanac-Derived Ingest Wishlist (Harmonia R0 / B7) — OPEN (source-gated)

id: OPS-016
status: open
workflow: ingest
category: future ingest
source_of_truth:
  - design-research/atlas-harmonia-r0-findings-2026-07-11.md (§Undated coda / B7 harvest)
why_it_matters: >
  The Atlas ideas-calendar's Undated coda doubles as an ingest wishlist (B7):
  each high-degree concept with no ingested developing text names, on its own
  page, the absent primary work. Unlike OPS-009 (speculate/weave-derived), this
  queue is derived from the corpus's connection topology — zero overlap with
  OPS-009's list, which is itself evidence the almanac sees different gaps.
next_action: >
  Acquire one named text into raw/, confirm no source page exists, then run the
  ingest skill. Ranked by undated-star degree (neighbors):
  1. Nietzsche, The Gay Science (1882) — lights THREE stars at once:
     death-of-god (21n; page marks GS §125/§343 "not in raw/"), nihilism (25n),
     and eternal-recurrence (§341; BGE marks it "developed elsewhere"; currently
     reception-dated 1961 via Heidegger). Opens the 1882 shelf.
  2. Nietzsche, Thus Spoke Zarathustra (1883–85) — the other BGE-named
     developing text for eternal-recurrence; opens the 1883–85 shelf.
  3. Merleau-Ponty, The Prose of the World (posth. 1969) — silence (37n, the
     TOP undated concept) already cites [[merleau-ponty-1973-prose-of-the-world]]
     ch. 2–3 with page numbers as a dead link.
  4. Husserl, Ideas I (1913) — horizon (16n); the Husserl leg is now
     Crisis + LI Vol. I (2026-07-21), but Ideas I itself (the horizon-anchor,
     the noema, the epoché's home text) remains absent.
  5. Heidegger, Gelassenheit / Country Path Conversations (1944–45) —
     letting-be (19n) is read against it only via Arendt/Carbone paraphrase.
  (Secondary lead: Picard, The World of Silence — the motif heading's namesake.)
done_when: >
  A named text is ingested and its dependent concepts gain developed anchors
  (visible as the concepts leaving the atlas Undated coda on the next build).

OPS-017 — Concepts-Developed Under-Listing Backfill — OPEN (execution-ready)

id: OPS-017
status: open
workflow: targeted maintenance
category: maintenance
source_of_truth:
  - design-research/atlas-harmonia-r0-findings-2026-07-11.md (§Undated coda, curation-gap rows)
why_it_matters: >
  Some high-degree concepts are Undated in ideas-time although their developing
  texts ARE ingested — the source pages' `## Concepts Developed` sections
  under-list them (the atlas chronology reads exactly that section, §4.2).
  The concept pages themselves already assert the development in prose.
next_action: >
  One ≤10-file batch, adding a Concepts Developed bullet ONLY where the source
  page's own body already warrants it (no new interpretation):
  - politics-mp (30n): merleau-ponty-1947-humanism-and-terror and/or
    merleau-ponty-1955-adventures-of-the-dialectic (concept page: "develops
    across" both).
  - hinge (21n): merleau-ponty-2010-institution-and-passivity ("most thematized
    in the 1954–55 course").
  - pensee-de-survol (22n): merleau-ponty-2010-child-psychology-pedagogy
    (CPP Ch 4 Piaget-survol critique; the 1961 kosmotheoros course stays
    source-gated under OPS-016's logic).
  Verify against each source page's Key Passages before adding; bump updated
  (substantive), run normalize_index_tables + refresh_baselines per Rule 19.
done_when: >
  The chosen bullets exist on the source pages and the concepts date on the
  next atlas build (leave the coda); or each row is explicitly re-deferred here
  with rationale.

OPS-004 — Plato Silent-Key and Reciprocity Backlog — OPEN (scoped, execution-ready)

id: OPS-004
status: open
workflow: audit / targeted integration
category: maintenance
source_of_truth:
  - wiki/.audit/silent-keys-2026-06-23.md
  - wiki/.audit/plato-coherence-gaps-2026-06-23.md (L-set rollup)
why_it_matters: >
  Tier-A Plato silent keys were integrated, but a second layer of load-bearing
  terms and reciprocal links remains in the report, making the Plato layer less
  navigable than the promoted claims imply.
next_action: >
  Run a single ≤20-file batch (triage-verified scope ~18 unique files):
  TIER-B (eudaimonist-premise, 5 Plato source pages + 1 NEW concept page):
    eu prattein / eudaimonia / chrēsthai / sōtēria / chrēsimon across
    plato-euthydemus, plato-charmides, plato-protagoras, plato-menexenus,
    plato-lysis; create eu-prattein.md (does not yet exist).
  TIER-C (select 6–8 from ~30): anankē (republic), auxiliary-cause (phaedo),
    hikanon/teleon (philebus), diakritikon (cratylus), ainigma (apology),
    dikaia homologia (crito), hermēneus (ion), time-undivided (laches).
  L-SET reciprocity (select 6–8): form-of-the-good ↔ statesmanship,
    elenchus ↔ collection-and-division, plato-ion/gorgias ↔ hippias-minor,
    plato-crito ↔ laches, priority-of-soul ↔ phaedo 105d.
  KEYS-PHASE CHECK (routed here 2026-07-28 from speculate-signoff batch 4; the
    methodological queue's MQ-16 now carries only the process-routing note):
    *survol absolu* was never captured by the *Institution and Passivity*
    ingest. `merleau-ponty-2010-institution-and-passivity.md` has ZERO
    occurrences of "survol", as does its extraction note; the ingested raw (an
    English edition) has zero "survol" and TWO of "absolute survey" — so the
    wiki attributes a term to that source via a third party
    (Décarie-Daigneault) which the source does not itself record. Juror E,
    verified by S. DD cites *survol absolu* to *IP* p. 136 n. 5 and NEITHER
    raw occurrence resolves at that locator (the first follows the `{168}`
    marker, the second sits near an inline `[198] (62)`): a footnote body may
    not have survived OCR, or DD may cite another edition's pagination. E
    declines to assert the locator is wrong (Rule 16 abstention) and records
    that it cuts both ways — the key also cannot be certified as checked.
    E's closing precision: this is unclosable *in-corpus*, which is NOT the
    same as unclosable — anyone holding a printed *Institution and Passivity*
    could settle it in a minute. So the action is "check against a printed
    edition", not "cannot be checked". This is the one assertion inside
    batch 4's [[pensee-de-survol]] caution that the corpus cannot verify
    as it currently stands.
done_when: >
  Chosen Tier-B/Tier-C items written to source/concept pages (or explicitly
  re-deferred with rationale here); eu-prattein.md created; index/log/baselines
  handled. Any L-set candidate claim registers at candidate and routes to
  consolidation for promotion (v0d.24/25 — no in-run supported decisions).
  AND the routed Keys-phase check above is **either executed or explicitly
  re-deferred with rationale recorded here** — an OPS-004 run may not be marked
  done while the *survol absolu* locator residue is silently dropped. Executing
  it means checking DD's *IP* p. 136 n. 5 citation against a **printed**
  *Institution and Passivity* (it is unclosable in-corpus, not unclosable) and
  recording the result on
  `wiki/sources/merleau-ponty-2010-institution-and-passivity.md` + its
  extraction note; re-deferring means saying so and why.

OPS-005 — Weave Pass-4 Motif Back-References — OPEN (in progress)

id: OPS-005
status: open
workflow: weave
progress:
  - 2026-06-03 batch + 2026-06-27 Plato batch (7) + 2026-06-27 OPS-sweep batch (8:
    hinge, anarchy-in-the-ground, ungrund, vollendung-der-metaphysik,
    onto-theo-logik, brauch, lethe-verbergung, contingency-of-the-future).
  - ~110 of 511 concept pages now carry the section; ~140 HUB/STRUCTURAL home
    pages still lack it (drafter: 409/511 lack it overall). Clean HUB *primary*
    homes are now largely exhausted; remaining batches skew STRUCTURAL / secondary.
next_action: >
  Run the next capped batch (≤10) of HUB/STRUCTURAL home pages lacking the
  section. Convention established; no first-run halt. Verify back-ref absence
  first; reference motifs.md headings verbatim inline (slash/accent rule).

OPS-006 — Weave Pass-2 Motif Residue — DONE 2026-06-27

id: OPS-006
status: done
workflow: weave
closed: >
  2026-06-27 (read-first triage; no writes needed). All 4 candidates already
  have homes — 0 of 3 stub budget used:
  - Derrida hand/hand-of-man/hand-of-God register → EXISTING-HOME
    hand-of-god.md (HUB, well-developed).
  - Nancy exscription/syncope/partage → INTENTIONALLY-DISTRIBUTED
    (exscription.md, syncope-nancy.md, partage-nancy.md + synthesis pages).
  - Christian body/chair semantics → EXISTING-HOME christian-semantics-of-chair.md
    (preliminary stub; enrich on a future Nancy/Henry ingest).
  - substance-becomes-subject → INTENTIONALLY-DISTRIBUTED (Hegel source +
    g-w-f-hegel entity + speculative-good-friday; motifs.md notes low priority).

OPS-007 — Deferred Weave Scans — BLOCKED (confirmed 2026-06-27)

id: OPS-007
status: blocked
workflow: weave
blocked_by:
  - no Plato-side concept page for flux/heracliteanism exists (verified
    2026-06-27: heracliteanism.md / flux-doctrine.md / being-becoming.md all
    absent; durée.md exists on the Bergson side). The structural-isomorphism
    scan flux/heracliteanism ↔ durée would return Indeterminate.
  - mimesis/simulacrum reciprocity already mostly captured by existing claims.
next_action: >
  Do not scan until a Plato-side flux/heracliteanism page exists (would require
  a small Cratylus/Theaetetus/Timaeus stub or a future ingest). Either create
  the page and run the scan, or retire the pair from the weave queue.

OPS-008 — Deferred Motif Candidates Awaiting Anchors — WAITING (confirmed 2026-06-27)

id: OPS-008
status: waiting
workflow: motif maintenance
waiting_on:
  - sovereignty / ipse / autoposition: needs non-Derrida cross-tradition anchor.
    LEAD (2026-06-27): Klossowski 1969 cites Bataille's souveraineté (Ch.1, raw
    160–208) — a candidate cross-author anchor; route to a future weave Pass 3 /
    motifs-delta rescan to evaluate (not promotable from OPS-008 alone).
  - survivance: intra-Derrida; no external anchor required but no entry yet.
  - person / inventory-impossible: needs Maritain/Berdyaev/Marcel primary anchor.
  - hauntology-in-MP / phantom: needs a 2nd non-M-C-2026 anchor (SA 2021 Ineinander).
next_action: >
  Recheck only when a relevant source lands; otherwise leave waiting. Evaluate
  the Klossowski/Bataille sovereignty lead at the next weave pass.

OPS-009 — Future Ingest Queue — OPEN (source-gated)

id: OPS-009
status: open
workflow: ingest
verified_2026-06-27: >
  Of the 11 high-rationale sources, 0 are ingestable now: Pinkard's Hegel
  Phenomenology is ALREADY ingested; the other 10 (Sartre B&N, Deleuze Logic of
  Sense / D&R / Nietzsche & Philosophy, Saussure CGL, Levinas TI/EE, Hegel Phil
  of Right, Heidegger Platons Lehre/Wegmarken, Derrida Khora/Dissemination,
  Hyppolite Genèse, Brandom Spirit of Trust) are absent from raw/. Stale
  wishlist rows (Heinbokel, Husserl Crisis, Bergson Creative Evolution,
  Klossowski) confirmed already ingested.
high_rationale_sources:
  - Sartre, Being and Nothingness; Deleuze, Logic of Sense / Difference and
    Repetition / Nietzsche and Philosophy; Saussure, Course in General Linguistics;
    Levinas, Totality and Infinity / Existence and Existents; Hegel, Philosophy of
    Right; Heidegger, Platons Lehre / Wegmarken; Derrida, Khora / Dissemination;
    Hyppolite, Genèse et structure; Brandom, A Spirit of Trust.
next_action: >
  Acquire one high-rationale source into raw/, confirm no source page exists,
  then run the ingest skill. Nothing actionable until a raw file lands.

OPS-010 — Ingest-Dependency Ledger Review — DONE 2026-06-27

id: OPS-010
status: done
workflow: speculate-signoff / ingest
closed: >
  2026-06-27 cadence review (see ingest-dependency-ledger.md "Review — 2026-06-27").
  IDL-1 plotinus-as-temporal-depth-substrate → SUNSET (RETIRE the HOLD,
  preserved-with-reasoning): route (c) spent, (a)/(b) inaccessible; residue flag
  added to lethe-verbergung.md + mythical-time.md Open Questions. IDL-1a →
  strike-the-bridge codified as permanent precondition (ingest dependency moot).
  IDL-2 militante → DISCHARGED (SA-2006 ingested; "un-ingested" note was stale).

OPS-011 — Methodological Queue Review — DONE 2026-06-27

id: OPS-011
status: done
workflow: workflow maintenance
closed: >
  2026-06-27 cadence review (see methodological-queue.md "Review — 2026-06-27").
  MQ-1 / MQ-2 / MQ-3 left as standing notes; NOT codified this cycle (each below
  the 3-batch recurrence bar). MQ-4 (resolve-vs-constitutive paired-rival shape)
  added 2026-06-27 — a 2nd instance of the protocol-shape family; codify on a
  3rd-batch recurrence. None has leaked into claims (verified).

OPS-012 — Saint Aubert and Legacy CR Backlog — PARTIAL (CR-034 done 2026-06-27)

id: OPS-012
status: partial
workflow: targeted audit / maintenance
done:
  - CR-034 (sensible-ideas.md): opening paragraph reframed to lead PbP-1945 as
    the ORIGIN anchor with V&I Ch.4 as canonical late elaboration, matching the
    body's "Three Readings of Vinteuil" structure. updated bumped 2026-06-27.
remaining:
  - SA Priority B: currently 4 open items (hyper-objet, generative-passivity,
    interanimality, intra-ontology) vs target ≤3 — close 1 (recommend
    intra-ontology: stale, updated 2026-05-04 before the 2026-05-07 SA-2006
    ingest; add an SA-2006 Position section) or formally re-defer.
  - CR-035 destruktion-as-the-method: SA-2006 blocker cleared but the targeted
    SA-2006 raw check (indirect-method / Destruktion chapters) is not yet done;
    if the chain is sound, promote the SuZ extraction-note candidate to a
    claims.md candidate.
  - CR-045 SA-2006 Silent Keys: the 334-line extraction note has ZERO "## Silent
    Keys" section (pre-schema ingest) — backfill 3–5 load-bearing terms
    (ontologie indirecte, médiation, démarche indirecte, …).
  - CR-047 valery-implex-genealogy: LIVE claim, audit verdict "no defect, no
    triggering ingest" — review only on a Valéry-adjacent ingest or Phase 8.6.
  - CR-048 valery-eventuel-mode: CANDIDATE; Counterpressure properly documents
    the Bergson check — no edit until Bergson secondary sources land.
next_action: >
  Run a focused legacy triage: close 1 SA Priority-B item; do the CR-035 SA-2006
  raw check; backfill CR-045 Silent Keys. CR-047/048 are proportional — leave.

OPS-013 — Non-Plato Claim-Label / Register-Status Sync — DONE 2026-06-27

id: OPS-013
status: done
workflow: targeted maintenance
closed: >
  2026-06-27. Re-derived the actual drift (the ~45/~27 estimate over-counted
  prose status-words): 23 inline label instances across 16 non-Plato pages
  mismatched the register. Relabeled each to current register status — mostly
  candidate→live (v1.9 promotions), plus brentano-as-erste-anstoss +
  sb-1942-kinetic-melody-origin + stendhal-naturalness → supported, and
  time-as-late-mp (supported)→(live). Found 2 pages beyond the original 14-file
  scope (reprise.md, merleau-ponty-2022-inedits-i); still ≤20-file gate. Label-
  only, no `updated` bumps (Rule 3 exception). The contested
  science-secrete-stiftung-chiasm cluster uses prose labels, no parenthesized
  mismatch — left as is.

OPS-014 — Out-of-Closed-List Typed-Connection Phrases — DONE 2026-06-27

id: OPS-014
status: done
workflow: lint / targeted maintenance
closed: >
  2026-06-27. User-authorized by-design schema amendment (option a). Verified
  scope: 5163 out-of-closed-list relation phrases across 815 hand-authored
  pages. Amended CLAUDE.md §Connections (schema v0d.21→v0d.22): page-body
  Connections MAY use free-form relation phrases; the closed typed-connection
  list is enforced ONLY on weave Pass-3 `Relation:` bullets (lint scope
  unchanged). The 5163 phrases recorded as accepted bespoke usage — NOT a
  mechanical-retype backlog. Permissive change: no skill/AUDIT_PLAN propagation
  needed. schema-changelog v0d.22 appended.

OPS-019 — v0e `claims/` + `type: claim` Extension — OPEN (decision session, user-gated)

id: OPS-019
status: open
workflow: schema design (user-gated)
category: deferred
source_of_truth:
  - wiki/.audit/v0e-claims-extension-kickoff.md
  - wiki/schema-changelog.md (v0d.33 + v0d.34 entries)
  - PR #119 (measurement pinned in comments)
why_it_matters: >
  v0d.34 retired the claims-register numeric outgrow marker after measuring
  it as fiction (200/303 entries over, median 449 words) and recorded the
  >2k-word tail — twenty entries at 2,054–4,002 words, twelve of them above
  3,000, the most-juried ones —
  as the standing empirical case for the extension the format has anticipated
  since v0d ("do not create dedicated claim pages under v0d"). The register
  now contains de-facto claim pages with no legal page type.
next_action: >
  Paste into a fresh session: "Read wiki/.audit/v0e-claims-extension-kickoff.md
  and run the decision session it describes." The kickoff sequences the design
  questions: go/no-go vs. the trim-downward alternative; the [[claims#slug]]
  link contract (permanent H2 stub recommended); lifecycle truth stays in the
  register (Atlas replay contract); full machinery-disposition table before
  any design; page format; migration as a newly sanctioned operation; v0e
  versioning; one-entry pilot (pbp-as-pivot) at most.
done_when: >
  User has decided go/no-go. If go: a schema-changelog design entry +
  machinery-disposition table exist, and the one-entry pilot PR is merged or
  explicitly deferred. If no-go: this item retired with the reasoning recorded
  (the trim-downward route becomes the standing answer for the tail).

Recently Closed / Do Not Re-Do

2026-08-18: OPS-020 — supported-layer re-verification sweep (23/23 rows, 0 restatuses, 16 §6-item repairs + ~60 sweep repairs) + refuter-verdicts ledger (177 rows, zero STANDS ever; 146/146 corrected to 147 historical) + build_refuter_ledger.py + consolidate C5 hook + ten pagination oracles. Do not re-sweep the supported tier; new supported promotions go through the jury as before. The Inédits II / GA 11 / Signs corpus locator sweeps are ROUTED to the next audit Citations phase, not open here.
2026-07-21: LI Vol. I ingest (branch claude/ingest-husserl-li1-2026-07-21) — Prolegomena + Investigations I–II; 11 new pages, 12 updates, 1 new BRIDGE motif (zu den Sachen selbst), 3 Harvest candidates queued in the extraction note. Do not re-extract Vol. I ranges when ingesting Vol. II; the Vol. I extraction note carries the forward-bridges (moments/pieces → Inv III; fulfilment → Inv VI; Brentano dating guard).
2026-06-27 (OPS-sweep): OPS-002 speculate-signoff batch 2 — 6 survivors, 4 candidate claims + 2 holds; OPS-001 — 12 non-Plato Phase-8 candidates registered; OPS-014 — schema v0d.22 (free-form page-body Connections); OPS-013 — 23 label syncs / 16 pages; OPS-005 — 8 weave Pass-4 sections; OPS-010 — IDL-1 sunset / IDL-2 discharged; OPS-011 — MQ left as notes; OPS-012 CR-034 — sensible-ideas reframed PbP-primary. OPS-006/007/008/009 triaged to disposition. Claims 281→297.
2026-06-27: OPS-003 (CR-033 evidence anchors) fully discharged — all 6 remaining checks resolved; three were material citation errors (IP "36/7" → 4; V&I particle-physics page-pairs; perceptual-faith footnote p.95 → p.123). CR-023's p.95 finding superseded → p.123. Bonus: absolute-as-moral- catastrophe locus corrected S&NS p.95 → p.124 (raw l.1087).
2026-06-27: OPS-005 named seed batch (bildung, hand-of-god, schritt-zurueck, sache-selbst) confirmed ALREADY covered (2026-06-03 batch) — do not re-add.
2026-06-23: Plato v1.9 Step 8.9 cite-back for 22 promoted claims / 33 pages; Tier-A Plato silent-key subsections; homoiosis-theoi page-grade; ion-enthousiasmos-phaedrus-mania candidate; 4 relation-only Plato cousins retired/relocated; PR #80/#81 review fixups.
2026-06-26: Old ingest-wishlist.md "Heinbokel un-ingested" row is stale; Heinbokel is ingested at wiki/sources/heinbokel-2021-johann-to-maurice.md.

Refresh Checklist

Read NEXT.md first and preserve its OPS-### IDs.
Read wiki/index.md for counts.
Read git status --short --branch.
Read the last 80-120 lines of wiki/log.md.
Read the completed-runs ledger at the top of AUDIT_HISTORY.md plus the newest wiki/.audit/audit-*.md digest (if any), and the newest relevant run block in AUDIT_HISTORY.md.
Inspect newest .audit/ reports since the last reconciliation.
Inspect .speculative/promotion/ ledgers only for status changes affecting OPS-002, OPS-010, or OPS-011.
Check whether old trackers contain stale claims before importing them.
Update this file, append a concise wiki/log.md entry if the operation changes the wiki operating system, and run the relevant validation.