v0d.9 Codification + Bridge-Card-Signoff Run-3 — Handoff for Next Session

Audience: future-Claude or future-user resuming this work in a later session. Status at handoff: v0d.9 codification + bridge-card-signoff run-3 complete. HEAD: 9e8157a bridge-card-signoff run-3 (after 6d616cb v0d.9 codification). 5 of 6 decision items from the previous handoff are discharged in this session. Only Decision 6 (Lawlor/Dastur/Madison secondary-literature ingest) remains, and it requires user action — sources are not in raw/.

TL;DR

1. v0d.9 Latent-Adjacent verdict codified. New verdict-class for cross-tradition pairs where axes (i) + (ii) align in family/form but axis (iii) grounding diverges registrally. Five-condition signature including Note 5 (auto-identification). Empirical evidence base: three confirmed cross-tradition instances. Apply-mode authorization scoping: typed connections + Open Questions + motifs.md BRIDGE entries AUTHORIZED; claims.md candidate emission NOT AUTHORIZED.
2. Scan D (Schelling barbarian-principle ↔ MP wild-being) confirmed the THIRD cross-tradition Latent-Adjacent instance, discharging Scan C's "commit after third" trigger and contributing Note 5.
3. 5 Decision-1 false-friend cautions applied to enveloping-enveloped, nascent-state, unvordenklich, incorporation-of-truth, plus claims.md transtemporality-ineinander Counterpressure.
4. Bridge Card 1 revised SIGNED OFF (3/3 APPROVED high-confidence) via bridge-card-signoff run-3 — the first card the v0d.9 Latent-Adjacent verdict authorizes. 4 apply-mode writes landed on aletheia + dehiscence (typed connections + Open Questions cautions + frontmatter bridge_cards: + HTML markers + provenance register entry).
5. Bridge Card 2 RETIRED via silent-key cluster reclassification analysis (cluster-of-3 verified: haecceity + coexistence + nascent-state). PhP silent-key cluster subsection landed on the source page under What's Not Obvious.

Wiki state: 485 pages | 57 sources | 277 concepts | 129 entities | 22 questions | 158 claim entries (17 supported / 83 live / 57 candidate / 1 contested / 0 retired). Same as start of session — no new pages or claim entries; 8 concept-page edits + 1 source-page edit + 1 claims.md Counterpressure addition + the CLAUDE.md / schema-changelog / SKILL / agent updates + 13 new audit reports.

Wiki schema: v0d.8 → v0d.9 (2026-05-17).

What landed in this run

Commit 6d616cb (v0d.9 codification, 20 files changed, 507 insertions, 37 deletions)

• Scan D report: wiki/.audit/weave-pass3-2026-05-17-followup/scan-D-barbarian-wild-cross-tradition.md — the third-instance confirmation + Note 5 (auto-identification) derivation.
• Bridge Card 2 reclassification analysis: wiki/.audit/weave-pass3-2026-05-17-followup/bridge-card-2-reclassification-analysis.md — cluster-of-3 finding; PhP silent-key cluster subsection draft.
• v0d.9 schema-changelog entry appended to wiki/schema-changelog.md — failure mode addressed, three-instance evidence base, five-condition signature, apply-mode authorization scoping, cross-tradition-restriction reasoning, 6-month retrospective review clause (2026-08-17 to 2026-11-17).
• CLAUDE.md schema version 0d.8 → 0d.9; new typed connection *has cross-tradition cousin* added to Page Format → Connections closed list.
• weave SKILL Pass 3 verdict taxonomy extended from 4 verdicts to 5 (Isomorphic / Latent-Adjacent / Partial / Disanalogous / Indeterminate) with the five-condition signature; bridge-card template Bridge type field gains latent-adjacent.
• bridge-card-signoff SKILL §12e carve-out for Latent-Adjacent (Open Questions + typed connections via *shares mechanism with* / *has cross-tradition cousin* + motifs.md BRIDGE entries AUTHORIZED; claims.md candidate emission NOT AUTHORIZED).
• structural-isomorphism-scanner agent updated — description, verdict-classification rules (added Latent-Adjacent with the cross-tradition qualifier), veto criteria (updated to route to Latent-Adjacent under cross-tradition registral-divergence), report format (added Latent-Adjacent template).
• 5 Decision-1 false-friend cautions applied to: enveloping-enveloped.md Open Questions (replaced existing "worth a future scan" entry with Scan B's verdict); nascent-state.md Open Questions (new bullet mirroring Scan B); unvordenklich.md Open Questions (new bullet citing Scan C's second-instance signature); incorporation-of-truth.md Open Questions (mirror of unvordenklich); claims.md transtemporality-ineinander-as-register-pair-extensions-of-institution Counterpressure (paragraph noting Scan B's negative finding strengthens the claim's specificity).
• Scan D recommended writes applied: false-friend caution upgrades on barbarian-principle.md and wild-being.md Open Questions with auto-identification handling.
• Bridge Card 2 RETIRED: wiki/.audit/weave-pass3-2026-05-16.md Section C Card 2 updated to Status: retired (2026-05-17; reclassification analysis at ...) with retire-and-replace note.
• PhP silent-key cluster subsection landed on wiki/sources/merleau-ponty-1945-phenomenology-of-perception.md under What's Not Obvious as item 4 — references all three cluster carriers (haecceity, coexistence, nascent-state) with provenance to the 2026-04-25 silent-key scan + the reclassification analysis.
• Index.md (header comment + 7 Updated dates + frontmatter) + overview.md (Updated date) + log.md (v0d.9-codification entry).
• Baseline refresh: only the graph SHA hash moved (page/claim counts unchanged at 485 / 158).

Commit 9e8157a (bridge-card-signoff run-3, 23 files changed, 1066 insertions, 12 deletions)

• Pre-step interpretation-checker (subagent): produced wiki/.audit/bridge-card-signoff/run-3/interpretation-map.md with 3 NON-TRIVIAL alternative readings; no INTERPRETATION-AT-RISK flag.
• 3 parallel reviewers (subagents A, B, C — all Opus): all returned APPROVED with high confidence.
  • Role A (gate-applier): all 7 schema tests PASS, including v0d.7 Test 4 presence-AND-adequacy (§12e Latent-Adjacent carve-out invoked as operative blocker).
  • Role B (adversarial-attacker): zero of 17 patterns fired; closest non-fires #10, #18, #19 structurally mitigated by v0d.9 design.
  • Role C (evidence-tracer): all 10 anchors verified verbatim; Rule 18 local-context checks confirmed for the three V&I anchors.
• Adjudicator (subagent): mechanically applied §9 row 1 (3/3 APPROVED) → SIGN-OFF; cohort fingerprint at cohort-fingerprint.md.
• Main-thread ratification: all 4 procedural checks PASS. Approved by maintainer: 2026-05-17 written onto the canonical revised-card source (wiki/.audit/weave-pass3-2026-05-17-followup/bridge-card-1-revised.md) and the run-local card.
• Apply-mode writes (4 items, all in scope per Signed apply scope: "all items"):
  • aletheia.md §"Connections": typed connection *has cross-tradition cousin* dehiscence with HTML marker <!-- bridge-card: aletheia-dehiscence-cross-tradition-self-differentiation -->.
  • aletheia.md §"Open Questions": false-friend caution upgrade extending existing figuratifs / perceptual-faith cautions to dehiscence as the LATENT-ADJACENT cross-tradition cousin.
  • dehiscence.md §"Connections": mirror typed connection *has cross-tradition cousin* aletheia with HTML marker.
  • dehiscence.md §"Open Questions": mirror Open Question with sibling-instance cross-references to Scan C and Scan D.
• Both pages get frontmatter bridge_cards: [aletheia-dehiscence-cross-tradition-self-differentiation]; updated: bumped to 2026-05-17. Provenance register entry appended to wiki/.audit/bridge-cards-applied.md.
• Sign-off run-3 audit directory: cards.md, dependency-map.md, interpretation-map.md, stripped/card-1.stripped.md, role-A/B/C hypothesis-pre + reports, adjudication/per-card-recommendations.md, adjudication.md, cohort-fingerprint.md.
• Baseline refresh: home_lint_total 329 → 333 (hypothesis-pre conformance 22 → 26 from run-3's reviewer reports); graph hash updated; check_lint_baseline.py + test_paper_a_a3.py + paper_a_provenance_expected.json + provenance.py + graph fixture updated.

Subagent budget across both commits: 7 Opus subagents — 1 structural-isomorphism-scanner (Scan D) + 1 general-purpose (Bridge Card 2 silent-key analysis) + 1 interpretation-checker + 3 reviewers (A/B/C parallel fan-out) + 1 adjudicator. Plus main-thread design + writes.

The critical schema change: v0d.9 in five sentences

1. What changed: weave Pass 3's structural-isomorphism scanner gained a fifth verdict — Latent-Adjacent — sitting between Partial and Isomorphic for cross-tradition pairs that exhibit genuine structural cousinhood at the substitute axis but registral divergence at the grounding axis.

2. Why it changed: three confirmed cross-tradition instances (Card 1 BLOCKED at run-2; Scan C 2026-05-17; Scan D 2026-05-17) plus the auto-identification stress-test passed by Scan D demonstrated that the strict Iso threshold was filtering out a stable, recurrent pattern. The Card 1 BLOCK at run-2 was specifically because *shares mechanism with* over-asserted at axis (iii) for a relation that was structurally cousined but registrally divergent.

3. What it authorizes: typed connections via *shares mechanism with* or the new v0d.9 *has cross-tradition cousin*; Open Questions caution upgrades; motifs.md BRIDGE / STRUCTURAL entry creation. It does NOT authorize claims.md candidate emission (registral divergence undermines the unification-of-grounding move a structural-parallel claim requires).

4. The cross-tradition qualifier is load-bearing: intra-tradition axis-(ii)+/axis-(iii)- patterns stay at Partial. Scan A (mythical-time ↔ transtemporality, intra-MP) exhibits the signature but is NOT a Latent-Adjacent pair under v0d.9.

5. First successful sign-off: Bridge Card 1 revised, applied to aletheia + dehiscence at commit 9e8157a. Sets the apply-mode precedent for the verdict-class.

Decision items surfaced (in priority order)

Decision item 1 (user action required): Lawlor / Dastur / Madison secondary-literature ingest

The only deferred item from the previous handoff. The revised Bridge Card 1's counter-test for upgrade-beyond-Latent-Adjacent toward Iso requires ingesting:

• Lawlor on dehiscence-as-différance
• Dastur on Lichtung and clairière
• Madison's La phénoménologie de Merleau-Ponty 1981

After ingest, re-run structural-isomorphism-scanner on aletheia ↔ dehiscence. If it returns Iso (not Latent-Adjacent), the upgrade obtains and Card 1 can be promoted to a structural-parallel typed connection + a claims.md candidate. If it returns Latent-Adjacent, the current state stands. If it returns Disanalogous, retract the typed-connection writes from aletheia.md + dehiscence.md (per the card's retraction counter-test).

Estimated effort: depends on ingest volume; the three sources together are likely difficulty-2 to difficulty-3.

Decision item 2 (cheap scan, ~15 min, 1 Opus subagent): Test Latent-Adjacent generalizability on a non-MP cross-tradition pair

All three founding Latent-Adjacent instances involve MP on one side. v0d.9's schema-changelog entry flags generalizability beyond MP-involved pairs as an open empirical question (not a blocking condition for the SKILL extension). Discharging this would strengthen the verdict-class's standing for the 2026-11-17 retrospective review.

Candidate non-MP cross-tradition pairs:

• Schelling Daß-Was ↔ Heidegger Ereignis — both name the thatness-vs-whatness / granting-act register; axis (iii) may diverge (Schelling's theogonic-cosmogonic apparatus vs Heidegger's Seinsgeschichte-eventual apparatus).
• Nietzsche eternal recurrence ↔ Hegel absolute spirit's return — surface vocabulary suggests parallel; axis (iii) almost certainly diverges (genealogy-of-morals apparatus vs Phenomenology-of-Spirit dialectic).
• Schelling Ages of the World ↔ Hegel Phenomenology — both name the becoming-of-the-absolute through stages; axis (iii) may diverge (Naturphilosophie cosmogony vs phenomenological-dialectical).

Estimated effort: ~15 min, 1 Opus subagent. Expected verdict: probably Latent-Adjacent for the first two; uncertain for the third. Any Latent-Adjacent verdict on a non-MP pair discharges the generalizability limit.

Decision item 3 (low-effort cleanup, ~10 min): Address hypothesis-pre template-conformance lint noise

The lint check check_hypothesis_pre_conformance scans *.hypothesis-pre.md files for empty content directly under ## Initial confidence and ## Final confidence H2 headings. Subagents tend to put the confidence value on a new line after the heading (not directly on the heading line), producing 4 new lint findings per sign-off run. Run-3 added 4 findings (22 → 26).

Fix at the agent-prompt level: update the 5 bridge-card-* agent definitions to instruct subagents to place initial/final confidence values directly on the H2 heading line (e.g., ## Initial confidence: medium). The lint check would then return 0 findings from new runs.

This is a v0d.7 §10 weakness 6 follow-up — the failure mode was acknowledged in v0d.7 but the fix was not shipped. The fix is small (5 agent-definition edits) and would prevent baseline drift on every future sign-off run.

Estimated effort: ~10 min, main thread only.

Decision item 4 (medium-effort, ~30 min): Phase 8 seeding of PhP silent-key cluster as candidate claim

The cluster subsection landed on the PhP source page in this session is currently a meta-finding without a claims.md entry. The reclassification analysis recommended path (a) — page-level subsection — as primary, but flagged path (c) as "partially viable but redundant given (a)": a meta-claim [[claims#php-silent-key-cluster-anti-already-constituted-object]] would be contestable (claim-grade), have evidence (the 2026-04-25 scan), and have payoff (re-reads PhP's argumentative architecture under the silent-key lens).

If seeded, the claim would be:

• Status: candidate
• Confidence: medium (3-source-cluster evidence on a single primary source, with the silent-key scan as anchoring artifact)
• Claim type: gap-identification (the cluster is a gap in MP's self-presentation that the silent-key scan surfaced)
• Sources: merleau-ponty-1945-phenomenology-of-perception
• Wiki homes: haecceity, coexistence, nascent-state, merleau-ponty-1945-phenomenology-of-perception

Trade-offs: the claim would link to the PhP page subsection as its primary evidence; further cluster members surfacing in future PhP silent-key scans would extend it. Risk: redundancy with the page-level subsection.

Estimated effort: ~30 min, main thread only. Promotion to live would require the 3-test gate at a future Phase 8.

Decision item 5 (routine, ~varies): Run any pending weave or audit cycles per cadence

Per CLAUDE.md weave skill, weave triggers on at least one of:

• Major-volume single ingest just landed (5+ new concept pages from one source).
• User request to "find latent connections", "weave the wiki", etc.
• Audit Phase 3 just promoted a new HUB without a corresponding concept page.
• Routine, mid-cycle cadence between audit batches.

Last weave-survey: 2026-05-16. The 2026-05-16 → 2026-05-17 follow-up sessions discharged most of that survey's surfaced work. A fresh survey could be run if the corpus shifts (next ingest) or if the user wants to check the impact of v0d.9 on existing concept pages (e.g., does any current *shares mechanism with* connection over-assert at axis (iii) and need re-typing to *has cross-tradition cousin*?).

Audit cadence: next audit recommended once ≥10 new ingests accumulate or if thesis-coherence drift is suspected. No specific trigger at present.

Decision item 6 (heavier-effort, only if surfaced organically): A fourth Latent-Adjacent instance

The 6-month retrospective review (target 2026-08-17 to 2026-11-17) audits whether Latent-Adjacent has earned its keep. If <50% of verdicts persist over the period, the verdict folds back into Partial per the schema-changelog v0d.9 §6-month retrospective review clause. Currently the verdict has 3 founding instances + 1 successful apply-mode use (Card 1 revised); the test will be whether new instances accumulate organically over the next 6 months.

If a fourth instance surfaces (in a future weave run or via routine ingest), it would strengthen the verdict-class's standing. No specific action item; the verdict's persistence is an emergent property of corpus growth.

Quick-win immediate actions (if you have 20 minutes)

1. Run /lint to verify 333-finding baseline holds. The current EXPECTED_SECTION_SIZES is: Inventory 11, Dead wikilinks 293, Typed-connection vocabulary 3, Hypothesis-pre conformance 26 — total 333. The Bridge-card provenance check should return 0 findings (markers + register + frontmatter all aligned for the new aletheia-dehiscence-cross-tradition-self-differentiation entry).

2. Verify the new HTML markers + frontmatter bridge_cards: lists render correctly in Obsidian. Both aletheia.md and dehiscence.md now have <!-- bridge-card: aletheia-dehiscence-cross-tradition-self-differentiation --> HTML markers + frontmatter bridge_cards: [aletheia-dehiscence-cross-tradition-self-differentiation]. The frontmatter list should appear as a list in Obsidian's properties pane.

3. Spot-check the new *has cross-tradition cousin* typed connections. Both aletheia.md and dehiscence.md got bidirectional typed-connection additions; both reference the bridge-card-signoff/run-3/adjudication.md.

What NOT to redo

• Don't re-run the v0d.9 codification. The verdict-class is committed (schema-changelog v0d.9 + CLAUDE.md + SKILL files + agent definition all updated). Reverting would require both a commit revert and a coherent retirement entry in the schema-changelog.
• Don't re-run the 5 Decision-1 writes from the prior handoff. They are applied in 6d616cb with verbatim content from the scan reports.
• Don't re-submit Bridge Card 1. Sign-off run-3 is complete; the card is SIGNED OFF and applied. The card's revised file at wiki/.audit/weave-pass3-2026-05-17-followup/bridge-card-1-revised.md carries Approved by maintainer: 2026-05-17.
• Don't re-retire Bridge Card 2. It is RETIRED in wiki/.audit/weave-pass3-2026-05-16.md Section C and replaced with the PhP silent-key cluster subsection on the source page.
• Don't re-run Scan D. The third-instance confirmation is in the audit report; revisiting it would not add evidence.
• Don't refresh baselines without first checking lint findings. Per the 2026-05-17 weave-apply handoff's caveat 1: refresh_baselines.py predicts forward, not on-apply. Substantive lint findings between dry-run and apply get pinned in --apply. Run dry-run, review warnings, fix-or-pin, then --apply.

Schema-mechanics caveats discovered this session

1. The bridge-card-signoff PreToolUse hook blocks writes outside active run dirs, including main-thread writes during ratification. The hook is active for as long as the run's .active marker file exists. To write the Approved by maintainer: line onto the canonical revised-card source (which lives outside run-N/), the main-thread must first remove .active. This session's flow: completed run → deactivated .active → wrote ratification line to canonical revised-card → applied apply-mode writes. The hook also catches Bash write verbs (cp, mv, tee, >, >>) per v0d.7 D4 — defense-in-depth working as designed.

2. Subagents cannot write to .audit/ outside their assigned run subdirectory. Scan D's apply-mode-recommended writes had to be applied by main thread (the scanner subagent is read-only per its SKILL constraint). Same pattern as 2026-05-17 weave-pass3-followup. Future Scan-D-style runs should account for this in planning.

3. Hypothesis-pre template-conformance lint noise from new sign-off runs. Run-3 added 4 findings (22 → 26); each future sign-off run adds ~4 more unless the agent definitions are updated to instruct subagents to place initial/final confidence values directly on the H2 heading line. This is v0d.7 §10 weakness 6 surfacing again — decision item 3 above addresses the fix.

4. The home_lint_total baseline is the sum of all section sizes (Inventory + Dead wikilinks + Typed-connection vocabulary + Hypothesis-pre conformance + ...). When section sizes change, home_lint_total in paper_a_provenance_expected.json must move in lockstep. refresh_baselines.py handles this automatically; the maintainer just runs --apply after dry-run review.

5. The interpretation-checker's INTERPRETATION-AT-RISK flag tracks the gap between trivial-stated and non-trivial-alternative readings — not the difficulty of the card. Run-3's Card 1 has three NON-TRIVIAL alternatives but its stated reading is itself non-trivial, so the flag does not fire. This is the intended behavior: the flag is for cases where a reviewer might accept the card's stated reading without realizing a more substantive alternative exists. The starter-set framing for A/B/C is the right place for the alternative-list to live regardless of flag status.

Files to read first if resuming this work

In order:

1. This file — orientation.
2. wiki/schema-changelog.md v0d.9 entry — the new verdict definition + five-condition signature + apply-mode authorization scoping + 6-month retrospective review clause.
3. wiki/.audit/weave-pass3-2026-05-17-followup/scan-D-barbarian-wild-cross-tradition.md — the third-instance confirmation + Note 5 (auto-identification) derivation. Contains the strongest empirical anchor for the verdict-class.
4. wiki/.audit/bridge-card-signoff/run-3/adjudication.md — the first Latent-Adjacent sign-off; main-thread ratification log; documents the apply-mode writes' scope.
5. wiki/.audit/bridge-card-signoff/run-3/cohort-fingerprint.md — drift comparison runs 1/2/3; run-3's structural-improvement-not-drift framing.
6. wiki/.audit/weave-pass3-2026-05-17-followup/bridge-card-2-reclassification-analysis.md — the cluster-of-3 PhP silent-key finding; the cluster-subsection draft that landed on the source page.
7. wiki/log.md entries [2026-05-17] v0d.9-codification and [2026-05-17] bridge-card-signoff | run-3 — full inventory of this session's writes.
8. wiki/.audit/weave-pass3-followup-2026-05-17-handoff.md — predecessor handoff; lists the Decision items this session discharged.

Subagent budget reflection

This session used 7 Opus subagents total across two commits:

• 1 structural-isomorphism-scanner (Scan D)
• 1 general-purpose (Bridge Card 2 silent-key analysis)
• 1 bridge-card-interpretation-checker (run-3 pre-step)
• 3 bridge-card reviewers (A/B/C, single parallel fan-out, single-card batch)
• 1 bridge-card-adjudicator (run-3)

The parallelization pattern that worked well: spawn Scan D + Bridge Card 2 analysis in parallel at the start of the session, then use main thread for v0d.9 design while they ran. When their results returned, fold findings into the v0d.9 design (Scan D contributed Note 5; the cluster analysis contributed the retirement path for Bridge Card 2). Then commit the codification, then run the 5-subagent bridge-card-signoff fan-out, then commit again.

Compared with the 2026-05-17 weave-pass3-followup session (3 subagents, no sign-off): this session is a heavier shape (7 subagents, 2 commits) but discharges 5 decision items vs the prior session's 3. The two-commit pattern is appropriate when the second commit's content (apply-mode writes) depends on the first commit's content (v0d.9 SKILL updates that make latent-adjacent a recognized bridge type).

Suggested session-starter prompt

Resume wiki maintenance from `wiki/.audit/v0d9-codification-2026-05-17-handoff.md`
(HEAD: 9e8157a bridge-card-signoff run-3 + 6d616cb v0d.9 codification).

The 2026-05-17 v0d.9 codification + bridge-card-signoff run-3 is complete:
- v0d.9 Latent-Adjacent verdict codified (schema-changelog v0d.9; CLAUDE.md
  + weave SKILL + bridge-card-signoff SKILL + structural-isomorphism-scanner
  agent all updated)
- Scan D confirmed the third cross-tradition Latent-Adjacent instance
  (barbarian-principle ↔ wild-being), contributed Note 5 (auto-identification)
- 5 Decision-1 false-friend cautions applied (enveloping-enveloped,
  nascent-state, unvordenklich, incorporation-of-truth, claims.md
  Counterpressure)
- Bridge Card 1 revised SIGNED OFF (3/3 APPROVED high-confidence) — first
  Latent-Adjacent apply-mode precedent; aletheia + dehiscence got typed
  connections + Open Questions cautions + bridge_cards frontmatter
- Bridge Card 2 RETIRED via silent-key cluster reclassification (cluster
  of 3 verified); PhP silent-key cluster subsection landed on the source
  page under What's Not Obvious

What's available to do next (in priority order):

1. (user action, ingest) Acquire and ingest Lawlor / Dastur / Madison
   secondary literature, then re-run structural-isomorphism-scanner on
   aletheia ↔ dehiscence to test for Iso upgrade. The only deferred item
   from the previous handoff.

2. (cheap scan, ~15min, 1 Opus subagent) Test Latent-Adjacent
   generalizability on a non-MP cross-tradition pair (candidates:
   Schelling Daß-Was ↔ Heidegger Ereignis; Nietzsche eternal recurrence
   ↔ Hegel absolute spirit; Schelling Ages of the World ↔ Hegel
   Phenomenology). All three founding instances involve MP on one side;
   discharging the generalizability limit would strengthen the
   verdict-class for the 2026-11-17 retrospective.

3. (low-effort cleanup, ~10min, main thread) Address hypothesis-pre
   template-conformance lint noise at the agent-prompt level. Update the
   5 bridge-card-* agent definitions to instruct subagents to place
   initial/final confidence values directly on the H2 heading line.
   Prevents +4 lint findings on every future sign-off run.

4. (medium-effort, ~30min, main thread) Phase 8 seeding of PhP silent-key
   cluster as candidate claim. The reclassification analysis flagged
   path (c) as a viable downstream extension of the cluster subsection
   already on the PhP source page.

5. (routine) Run any pending weave or audit cycles per cadence. Last
   weave-survey 2026-05-16; substantially discharged. Next audit
   triggers on ≥10 new ingests or thesis-coherence drift suspicion.

6. (organic) Wait for a fourth Latent-Adjacent instance to surface in
   routine work. The 2026-11-17 retrospective review will assess
   whether the verdict-class has earned its keep (<50% persistence
   triggers retirement).

Tell me which to pursue (or "1+2+3" or "do all the cheap ones").