CRITIC_SLICE-register-variety.md

pipelines/CRITIC_SLICE-register-variety.md

CRITIC VERDICT — SLICE-register-variety (fresh-context adversarial refutation)

Draft under review: scratchpad/realm_wave/SLICE-register-variety_DRAFT.md

Repo audited READ-ONLY: C:\dev\humanity-forgotten. Date 2026-07-28.

Method: every citation opened; every zero re-run with my own positive control; every MISSING

re-searched against the designed-not-missing corpus; every care-bearing recommendation run through

the devil's-advocate-against-timidity pass per the ruled calibration.

HEADLINE: 21 findings in · 1 KILLED · 20 SURVIVE · 8 AMENDED · 5 CONTESTED.

The draft's evidentiary discipline is unusually good — its registry arithmetic, its beat census and

its art-bullet arithmetic all reproduced EXACTLY under my independent parse. The defects are (a) one

finding that is a category error and self-defeating, (b) two MISSING verdicts that fail the

designed-not-missing rule, (c) one care-tier inflation, (d) one internally contradictory fixture, and

(e) a citation-precision class in the section numbers (never in the quotes).

---

0. What I re-derived independently, and what it returned

Every quoted string in the draft resolves to real text in the named file. I found no fabricated

quotation anywhere in 21 findings. That is worth stating plainly before the kill list.

Citation-precision defect class (line numbers and section numbers)

Body-prose quotes are cited to the paragraph block, not the line:

Section numbers into the REALMARC brief are wrong three times (headers re-derived by regex):

• draft "§5 (field mints)" → the quote lives at L832, inside §4.2 (*THE LOOP'S*).
• draft "§7 LENS INTERFACES" → L964, inside §6 WIRING.
• draft "§8" for the register_dominant-absent confirmation → L1011, inside §7 Verification notes. The brief has no §8.
• draft "REALMARC §5 specifies RB-ASC-REGISTER" → specified at L834 (§4.2), discussed at §5.5.

Charter section numbers, by contrast, are all correct: §4.2 L1119-1122, §5.2 L1241, §1.1 L169-172,

§2.3 V-ADEQUACY L497, §2.4 L556, §2.5 L617/L622, §2.7 L672/L694.

**Not fatal — but this document is written to be consumed by a schema author who will follow the

pointers. Fix the four REALMARC section numbers and re-anchor the five body-prose line cites before

handoff.**

---

1. KILL LIST

KILLED — F-R13, "the 'five consecutive death realms' framing is imprecise; the real run is four"

Three independent failures, any one of which is disqualifying:

(a) Category error. The charter's §4.2 phrase, read in full at L1119-1122, is:

*"a doctrine question that must be answered BEFORE the per-realm charters, not after, or five

consecutive death realms converge into the named monotone-solemn slop mode."* That is a

statement about five nodes' SUBJECT MATTER and a hypothetical failure mode — not a claim that all

five currently READ solemn. The draft refutes a register claim the charter never made. Annwn is a

death realm by subject in its own tradition (Arawn is its lord of the dead; the Cwn Annwn are

death-omen hounds — both in docs/spine/CH_73.md L9). Warm register does not remove it from the

subject run.

(b) The recommendation is self-defeating. The draft calls Annwn *"the load-bearing node for

register variety"* and then recommends *"correcting the framing in both documents so later waves do

not size four nodes of work as five."* Editing the charter to say "four" removes Annwn from the

guard's declared scope — deleting the protection the same finding says is load-bearing. The charter's

five-node count is precisely what keeps a later designer from grading Annwn's warmth down.

(c) Wrong verdict class. DIVERGENT-BY-DESIGN, per docs/DESIGN_GAP_REGISTER.md L195, means

*"we deliberately differ"* from what a comparator title teaches. F-R13 is an internal

doc-wording note about our own charter. It is not a divergence from Spiritfarer / Grim Fandango /

Edith Finch / Hades in any respect.

What survives and where it goes. One sentence: *Annwn is the road's authored break — CH_73.md

L196, "the relief after the dead-realms" — and any later pass that raises its intensity or grades

its warmth down is spending the doctrine's own reserve.* That is true, cited, and valuable. **Fold it

into F-R15 as a protective note.** Do not carry F-R13 as a finding, and do not carry its

charter-edit recommendation at all.

Revised counts after the kill: COVERED 2 · DESIGNED-UNBUILT 5 · WEAKER 11 · MISSING 0 · DIVERGENT 0

= 20 (see the two MISSING amendments below).

---

2. AMENDMENTS — findings that survive with their verdict, scope or recommendation corrected

A1 — F-R02: four of the five "probe-forced widenings" are ALREADY NAMED in the repo. Reuse-first attribution required.

The draft presents elegiac · exacting · adversarial · hospitable · ascendant as its own probe

residue. docs/pipeline_review/REALMARC_PATTERN_BRIEF_2026-07-27.md §0.3, L73-76 already reads

the same body prose and names:

*"Ch 70 elegiac-mournful · Ch 71 adversarial-trickster (the ascent's only comic register) ·

Ch 72 ascendant-moral-architectural · Ch 73 warm-hospitable · Ch 74 vast-elegiac-laboring ·

Ch 75 grief-and-war · Ch 76 grave-argument resolving to intimate-stillness · Ch 77

wonder-and-release. No two consecutive nodes share a register; every five-node window carries five

distinct ones."*

Four of the five words — elegiac, adversarial, ascendant, hospitable — appear there verbatim,

and the rotation check in the draft's §1.3 is a re-derivation of REALMARC's own sentence. **Only

exacting (Ch 69, outside §0.3's Ch 70-77 range) is genuinely new.**

The draft cites §0.3 in F-R03 for the "body prose is differentiated" conclusion, so it read the

paragraph; it simply did not credit it for the values. This is the reuse-first rule applied to

vocabulary rather than to systems.

AMEND, do not kill. The finding's substantive claim — *the seeded eight cannot carry the road,

and forcing it would manufacture in the data the monotone the doctrine exists to prevent* — is made

nowhere else and is correct. Rewrite §1.2 as: *"REALMARC §0.3 already named four of these registers

in prose; this slice's contribution is (i) promoting them from prose adjectives to closed-enum VALUES,

(ii) exacting for the Duat, and (iii) the collision evidence that shows why the seeded eight cannot

absorb them."* That is a smaller and defensible claim, and it strengthens rather than weakens the

mint, because a value two independent lanes reached from the same prose is better evidence than one.

A2 — F-R18: MISSING → WEAKER. A frame-side comparator rubric already exists at proposal tier.

The draft mints RB-REG-FRAME (sort unlabeled frames into two realms; name each realm's dominant

register; name any delivered forbidden register) and classes it MISSING. The charter's §2.7 already

proposes, at L694-700:

RB-RLM-LAWFUL (*"from ten unlabeled frames, state the realm's signature law"*) ·

RB-RLM-PROVENANCE (*"name the source tradition from frames alone; a verdict of generic fantasy is

the sharpest available slop detector"*) · **RB-RLM-POSITION (*"shown this realm beside the previous

one, state which comes later on the road"*)**

RB-RLM-POSITION is a frame-side, two-realm comparative rubric — the draft's own TASK 1 in all but

the question asked. The draft names RB-RLM-LAWFUL as the filing neighbour but never mentions

RB-RLM-POSITION, which is the nearer twin. Under the standing wave rule, a specced-but-unbuilt

neighbour makes this WEAKER, not MISSING.

AMEND: reclass WEAKER; fold TASK 1 into RB-RLM-POSITION (the frames are already interleaved and

already require a two-realm discrimination — asking "which comes later" and "which is which" off one

board is one rubric run, not two), and mint only TASK 2 + TASK 3 as the register extension.

The core insight is untouched and is the best finding in the audit: **RB-ASC-REGISTER reads body

prose, F-R03 proves the convergence is in the art bullets, so the rubric and the defect are in

different layers.** That stands and is decisive under the Josh gate.

A3 — F-R19: MISSING → WEAKER. The too-wide direction is already ruled.

The draft asserts *"Nothing in either proposal probes for this."* False for one of the two directions:

• docs/proposals/STORY_DEPTH_PROGRAM.md L932-935: *"All enum rosters are examples-not-limits —

emotional_register ×8 … Each expands by ruling, and **a never-used enum member is itself a

finding**, on the same logic that makes a never-picked pool option a balance defect."*

• Restated at STORYDEPTH_STORY_CRAFT_LENS_FAMILY_2026-07-27.md L654-655, and at

STORY_DEPTH L223.

That IS a too-wide guard, and it is the AAAAA balance doctrine's own shape re-pointed at an enum —

which makes it the cheapest possible reuse target. What the draft adds and nobody else has: the

too-narrow direction (the enum-adequacy probe) and the WIDTH GUARD (a random assignment over

the ratified enum must FAIL E1+E2+E7, or the rotation rules are ceremonial). The width guard is

genuinely new, deterministic, zero-token, and I endorse it.

AMEND: reclass WEAKER; state the recommendation as *"arm the existing never-used-member rule as

the width guard's first arm, and add the random-assignment guard as its second"* — reuse-first, one

new assertion instead of a new probe family.

A4 — F-R10: strike the §17.1 escalation. Care-tier inflation, and the citation runs the other way.

*(devil's-advocate-against-timidity pass)*

F-R10 argues the Hades-the-game warmth *"would be a §17.1-adjacent drift toward the Homeric-elegiac

care commitment."* I opened the block it points at. docs/spine/CH_70.md L256:

*"Q4 — deities-real and not-a-Christianized-Hell — RESOLVED (§17.1): Hades is the sober, just lord

of the dead by cosmic lot, never a devil; … Homeric-elegiac in register, **never a torture-hell or

a horror-country**; the Hades-as-Satan fusion is firewalled"*

The §17.1 commitment's named enemies are the DARK direction — torture-hell, horror-country,

Christianized Hell, Hades-as-Satan. It does not forbid warmth in any clause. Invoking a care line

whose whole force runs against grimness in order to close off a WARM option inverts it, and that is

exactly the care-tier inflation the ruled calibration names as a defect ("the timid option is usually

the wrong one"). Compounded in §4 contested item 2, which routes any Ch-70 warmth question *"through

the care doctrine rather than through this slice"* — it does not need to go there.

What the draft does NOT need to give up. Two arguments carry F-R10 on their own and neither is

care-tier:

1. Structural. The road is one-pass by construction; Hades-the-game's warmth accrues across dozens

of returns. That mechanism cannot be ported. Correct, and sufficient.

2. Authored craft. CH_70.md L199 forbids whimsy in as many words. Whimsy ≠ warmth. The

craft ban is on the *tonal register of whimsy*, not on the found-family tenderness the draft

itself proposes as the substitute.

AMEND: strike the §17.1 sentence from F-R10 and the care-doctrine routing from §4 item 2; keep

the structural argument, cite L199 for the craft limit, and keep the found-family substitute

(which is additive and passes the anti-timidity bar cleanly). Then note for the director that a

future Ch-70 warmth decision is a craft call the loop rules under the standing protocol, not a

care-doctrine escalation.

A5 — F-R06 / §3.4: the "live-data E1 MUST-FIRE fixture" is invalid under the draft's own assignment.

The underlying fact is real and I reproduced it: CH_67.md L201 and CH_68.md L193 both carry

awe-and-dread verbatim on consecutive nodes.

But E1 tests the column, not the prose. The draft's own §1.3 assigns Ch 67 = dread and

Ch 68 = wonder. Those do not collide, so E1 will not fire on that pair as live data — and §3.4

simultaneously lists *"E1 MUST-NOT-FIRE: the §1.3 assignment string,"* which contains that very pair.

The same two nodes cannot be both fixtures. The escape route is also shut: F-R16 establishes that

CH_67 and CH_68 have no biome_primary register clause, so a mechanical extractor would not

derive identical values from them either.

AMEND: revert Clause 3's must-fire to REALMARC §5.4's synthetic (L900), OR restate the Ch 67/68

pair as a fixture for the new E8 art-bullet agreement tooth, where it genuinely fires (two

consecutive nodes whose art bullets carry one register phrase while their declared dominants differ

is precisely E8's target). The second option is better: it gives E8 a live must-fire and keeps the

finding's evidentiary value. F-R06's WEAKER verdict is otherwise sound.

A6 — F-R16: the hand-assignment count is six, not four.

Verified by parsing every - biome_primary bullet: the explicit *"the register X and Y"* clause is

present at CH_70, 71, 72, 73, 74, 77 and **absent at CH_66, CH_67, CH_68, CH_75, CH_76 and

CH_EPILOGUE** (CH_69 carries the variant *"grave and exact"* form the draft correctly flags). The

draft lists only four absences and omits CH_76 and CH_EPILOGUE.

Low impact — §1.3 supplies all thirteen assignments anyway — but the extraction-script spec inherits

the wrong number. AMEND to six + one variant form.

A7 — F-R03: luminous spans 9 of 12 nodes, not 8.

Occurrence count 15 is exact. Node count is 9: CH_66, 68, 69, 70, 72, 74, 75, 76, 77. grave (21 /

8 nodes / CH_69-76 consecutive), warm (5 / CH_73 ×4 + CH_77 ×1) and elegiac (3 / CH_70) all

reproduce exactly. The correction makes the finding stronger, not weaker.

Also amend the method note in §0: the "corrected search returned 33 headings" figure does not

reproduce under any matcher I tried — ^## \*\*Ch \d+ = 64, ^## \*\*Ch \d+ — = 13, ^### \*\*Ch \d+

= 49. Every substantive RCM claim the number supports verifies independently, so this is a stray

figure in a positive-control paragraph rather than a bad finding — but a positive-control section is

the one place a number must be reproducible. Restate it as the matcher plus its result.

A8 — §3.1 slop-mode list: scope gore to RENDERING MODE, explicitly, or the tooth becomes a violence ban.

*(devil's-advocate-against-timidity pass)*

The proposed closed slop-mode list — `horror-gore · whimsy · spectacle · caricature · buffoonery ·

theme-park · foe-scoring · hostile-menace` — is honestly extracted; I confirmed every clause exists

verbatim in the twelve material_palette bullets, and never gore is live authored canon at

CH_72.md L203 and CH_74.md L203. The draft invented nothing.

The risk is downstream. The moment register_forbidden becomes a machine-read column feeding

negative prompts to the concept-board generator and a critic false-positive list, a bare token

gore will be applied by tools that never read the art bullet's context. This is a rated-M AAA action

RPG in which lethal violence against human NPCs is CORE (CVD content bar; memory

violence-allowed-r17-no-genocide), and the endgame includes an eschaton-dragon war front at Ch 75.

A negative prompt reading "no gore" against a realm-war concept board defangs a surface the CVD

requires to land.

AMEND docs/proposals/REALM_DESIGN_PROGRAM.md §2.3's field spec, not the extraction: declare on

the column that register_forbidden governs rendering register only — the aesthetic mode in

which a thing is depicted — and **never gates violence content, encounter lethality, or damage

depiction**, which stay governed by CVD §17 and the content bar. One sentence in the schema

docstring. This is the same construction the draft already uses correctly to keep care negations out

of the column, applied one level deeper.

---

3. FINDINGS THAT SURVIVE UNAMENDED (12)

F-R01 (DESIGNED-UNBUILT, contested). The collision is real and freeze-blocking. emotional_register

appears in STORY_DEPTH L728-733/L796 and STORYDEPTH L433/L444/L448-462; register_dominant appears in

REALMARC L284/L816/L878. Cross-check: register_dominant returns ZERO in STORY_DEPTH and

emotional_register returns ZERO in REALMARC — two lanes, one column, no awareness of each other.

Both name this slice as the enum's owner (REALMARC L832; charter §4.2 L1119). Two rotation rules

confirmed incompatible: STORYDEPTH E1 = no 3 consecutive (L451) vs REALMARC **Clause 3 = no 2

consecutive + ≥3 distinct per 5-window** (L221-222). The draft's recommendation is well-reasoned and

I endorse it: emotional_register is the name already carried into PRE_5090_BUILD_PLAN.md L120 and

FACTORY_CONTRACT.md L40 as a scheduled freeze rule, so it is the expensive name to change; the

stricter rotation costs no new machinery because REALMARC L878 states R7 is *"structurally identical

to R1/R2"*; and hosting it in gate 33 keeps the topography gate's HL_0048/HL_0049 scope statement

true. Director call, as flagged.

F-R04 (COVERED + precision correction). Verified to the cell. The correction is genuine and the

reuse target (_unpopulated at L69) is the right one. Note in support: REALMARC L878-880 does already

carry *"the same unpopulated-is-a-hard-fail discipline applies, so the register column can never be

green by vacuity"* — so both lanes have the rule and only the draft has the sentinel-FORM precision.

F-R05 (WEAKER). The strongest data finding in the audit. emotional_register is declared the

upstream source of music_mood_tags (STORYDEPTH L444) and the two vocabularies share zero terms —

verified in both directions. The mapping table in §3.2 is the right fix and is correctly seeded only

from tags the live registry already uses.

F-R07 (WEAKER). CH_76.md L212 authors both phase registers in one bullet; CH_75.md L269 names

three; P1a-P1c / P2a-P2b confirmed as real beat ranges; grief_country confirmed as a real site.

The recommendation rides funded work (REALMARC L830) and adds one field — textbook reuse-first.

F-R08 (WEAKER). Census exact in all 18 cells; every voice_and_crowd quote verbatim. The

recommendation explicitly declines to mint a populace system and routes to REALMARC's Clause-5

companion-exchange obligation (L230) — correct.

F-R09 (WEAKER, care-loaded). Passes the anti-timidity pass cleanly: the fix is additive

(a speaking Yama/Chitragupta at a weighing-place that already exists in prose at CH_72.md L26), which

is care-as-weight-and-authenticity, the opposite of omission. Charter L169-172 confirms MAXIMUM care +

LEAST attention verbatim. One scoping note for the record: CH_72.md's voice_and_crowd DOES declare

Indra, Yama and the nagas voiced *"where present"* — so the realm is not unpeopled at the declaration

layer, only at the BEAT layer. The draft's *"empty stairwell"* phrasing overstates by a shade; the

precise claim (no dialogue beat stages the declared speakers) is the one to carry.

F-R11 (WEAKER). All twelve negation clauses verified verbatim. The zero is positive-controlled

two ways: harness/*.py never reads the art bullets (control: intensity hits 3 files), and

T0_Hard_Lines (123 rows) carries none of the register-negation vocabulary. check_spine_graph.py

L348-354 parses the scene-anchor reveal TOKEN (HELD/LIFTED/NONE) but not the register text after it,

so CH_70.md L173's *"Homeric-elegiac register (HL_0001)"* is stored and unread — which corroborates

the finding rather than refuting it. Highest content-already-written-to-column-cost ratio in the audit.

(Amendment A8 applies to the column's declared scope, not to this finding.)

F-R12 (COVERED in prose / DESIGNED-UNBUILT as data). All three opening quotes real (line numbers

per the precision table). The dual verdict is honest and correctly stated.

F-R14 (DESIGNED-UNBUILT, NON-BLOCKING). Fully verified, including the Ch-73 acquisition-block gap,

and correctly narrowed: the draft claims only the block gap, never an Annwn-absence — which is

exactly the distinction REALMARC L992-997 records as the earlier draft's false zero

(*"The absence claim the earlier draft carried is false; the narrow block-gap claim is true"*). The

draft did not inherit the killed claim. The operative conclusion — **the register mint does not block

on the RCM repair** — is correct and is the most useful sentence in the finding.

F-R15 (DESIGNED-UNBUILT). CH_73.md L26's refusal of the seduction reading verified verbatim;

charter §2.4 L556 (the food consumable) and §2.5 L622 (THE SINGLE BLOW) verified. The discipline —

*do not declare a register the build does not deliver* — is sound. **Fold the surviving sentence of

the killed F-R13 in here.** Contested item 4 stands as flagged.

F-R17 (WEAKER). CH_70.md L3 (*"katabasis survived as nostos"*) and L32 verified; CH_75.md's

interrupted homecoming verified; Epilogue L181 verified. The DWELL TIME framing is the real craft

lesson and it is not made anywhere else in the corpus. Rides funded work.

F-R20 (DESIGNED-UNBUILT). Verified by direct ls with the specified positive control. Correctly

scoped as informational so no return claims an armed tooth. Corroboration: charter L655 reads

*"proposed as harness/check_realm_ruleset.py"* — proposed, never built.

F-R21 (DESIGNED-UNBUILT). 12 rows, ids exact, zero realm/region/place themes; every road node's

music_mood bullet ends [POPULATE→T0_Theme_Registry] and names motifs by name. The

CASSIUS_PHASE_1 corroborating note is verbatim. Correctly declines to claim R-A16.

---

4. RE-LITIGATION AND DIVERGENCE CHECK — CLEAN

• R-A7a / the transitions-conveyance fork (the guard named in my brief): NOT re-litigated. §5 parks

R-A7b explicitly on FORKC_ANALYSIS_DESIGN_PRECEDENT_2026-07-27.md (file confirmed present). Where

the draft touches seams (F-R12, F-R17, §3.1's register_from / register_to) it rides fields

REALMARC already proposed at L817 *"on the seam rows the Fork-C pass is already going to create"*

— supplying values to an ordered pass, never re-opening the conveyance question. Correct posture.

One watch-item only: F-R08's recommendation pulls the Clause-5 seam obligation *inside* the realm.

That is a companion-beat requirement, not a conveyance mechanic, so it clears — but the Wave-2

charter should record that the seam obligation is now claimed at two grains.

• docs/DESIGN_GAP_REGISTER.md: 464 lines, DIVERGENT-BY-DESIGN entries at L195-228 (FF7R, WoW,

GTA/RDR2, MHW, Elden Ring). R-A6, R-A12, R-A18 all return ZERO — no aspect rows exist, so

the draft's §3.6 obligation to mint them is correct and nothing is re-opened. no-scoreboard (L169)

and the markerless-navigation divergence (L207) are untouched by this audit.

• One pre-registered boundary the draft should cite and does not: charter §4.3 L1185-1186 —

*"director-ai-emergent-dread: pre-register the boundary with L-R5 so dread and grief are not audited

as the same register."* The draft carries dread as an enum value and assigns it at Ch 67 (+ three

secondaries) without acknowledging that boundary. No violation — it tags register, it does not audit

dread craft — but §3.6 should add a wiring row recording that the enum's dread value sits on a

pre-registered lens boundary, or a later L-R5/director-ai-emergent-dread collision is guaranteed.

---

5. CONTESTED — for the director

1. F-R01, the column collision. Confirmed real, confirmed freeze-blocking, confirmed mutually

unaware (each doc's zero for the other's name is positive-controlled). The draft's recommendation

is sound and its strongest objection is correctly answered. Must be settled before either mints.

2. Ch 74's register — an unflagged divergence between two live proposal-tier reads. The draft

assigns Ch 74 awe / dread. REALMARC §0.3 L75 reads the same prose as "vast-elegiac-laboring."

Under REALMARC's read Ch 70 and Ch 74 would share elegiac (non-consecutive, so both rotation rules

still pass either way — the stakes are the concept board and the audio brief, not the tooth). The

draft never records that it diverges. Director should pick, and whichever wins should be recorded

in the schema doc as an adjudicated read rather than left as two live readings.

3. The §17.1-vs-craft boundary on register commitments (amendment A4). CH_70.md L256 files

*"Homeric-elegiac in register"* inside a §17.1-RESOLVED care block. Does a register word inside a

care block make the register itself care-tier? My reading: no — the §17.1 force there runs against

torture-hell/horror-country/Satan-fusion, and treating the register word as care-bound would make

every future tonal adjustment a care escalation. Worth one ruling, because the answer governs how

register_forbidden is waived for the rest of the program.

4. F-R15, Annwn's seduction. The draft declines the charter's own §4.2 phrase

*"seductive-and-hospitable"* (L1120) on built-prose grounds, correctly citing CH_73.md L26's

explicit refusal. Defensible and disciplined — but it lets the current prose overwrite a charter

intent, and the charter is the more recent design statement. Recommend the secondary be recorded as

an OPEN slot flagged to the Wave-2 Annwn charter rather than defaulted to resolve, so the

intent is not silently retired by a data mint.

5. Citation precision before handoff. Four wrong REALMARC section numbers (incl. a cited §8 that

does not exist), five body-prose line cites off by 1-7, one non-reproducing count in the

positive-control section, luminous node count off by one, F-R16's absence list short by two.

Zero fabricated quotes across 21 findings. This is a pre-handoff cleanup pass, not a

credibility finding — but §3 is explicitly written for direct pickup by a schema author who will

follow every pointer, so it must land before the handoff is consumed.

---

6. Bound on this critic pass

I verified citations, verdict classes, zeros, reuse-first posture, divergence collisions and

care calibration. I did not independently re-run the enum-adequacy probe over Ch 2-65 — the

draft declares that bound itself (F-R19, §5) and the declaration is honest. I did not evaluate

whether the built prose DELIVERS its declared registers in play; no frame exists, the draft says so,

and no claim in this verdict inherits one.

Generated by harness/site/structure_site.py — the URL path is the repo path. review root