CR_0177_soorasamharam_remnant.md

creatures/CR_0177_soorasamharam_remnant.md

CR_0177 — the Soorasamharam-remnant: the body under the landed encounter, and the seam it comes apart on

CANON SUBORDINATION — this document is PROPOSAL-TIER: it serves canon and never outranks it.
The canon: the CVD · the T1 foundation docs · the T0 registries (registries/) · the spine
(docs/spine/CH_*.md) · the region pages (_source/02_Tier_2_Region_Pages/). Authority order: docs/DOC_MAP.md § 0.
Canon served (scanned from this document's own citations — widen it by hand where it is thin):
CVD §17.1 · T0_Boss_Encounter_Registry · T0_Chapter_Index · T0_Creature_Roster ·
T0_Environment_Grammar_Registry · T0_Hard_Lines · T0_RNG_Drop_Table · T3_Familiars_Named ·
docs/spine/CH_07.md · _source/02_Tier_2_Region_Pages/south_india.md
READ THAT CANON FIRST — open it and derive from it before you build anything from this document.
If this document disagrees with canon, CANON WINS and this document is the defect — fix the
document, never the canon. Nothing here is applied until it is ratified into canon.

STATUS: PROPOSAL-TIER design sheet, written to docs/proposals/bestiary/BESTIARY_ELEVATION_DOCTRINE.md

and critic-gated against its ten refusal tests. It cites canon rows and writes none. It is the

CREATURE-side sheet for a row whose ENCOUNTER sheet already landed, and § 1 rules the division of

labour before it authors anything.

1. identity

Murugan demon-war)" · chapter CH_07 · region south_india

collect_verb CONTAIN_PURGE · dex_completion_class CONTAINMENT ·

cultural_depiction_required TRUE

thread_anchor Thread 18; Thread 19 · chapters_present CH_07 · region_native south_india ·

chapter_first_encounter CH_07

transformation, pre-encountered at the Palani summit and gated at the Crown-seat tier"

mesh_pipeline_type image_to_3d · generation_status pending

trophic_role, scientific_name, trade_8_tameable, temperament_prior,

familiar_slot_eligible, spawn_rate, every husbandry cell, rare_variant, extinction_threat,

and rng_drop_anchor (the payload rides BE_0007_S3's own two drop rows)

material at elevated attention [CH_07.md:294]. § 7 reads it exactly.

= FOLKLORE_FACED (SELF_EVIDENT_IN_THE_GROUND, IN_SCOPE, ruling null) and ontology_class

= CONDEMNED (RULED_IN_L11, RULING_10, IN_SCOPE). RULING_10 FIRES ON ontology_class ALONE —

the permission cell needed no ruling because it is self-evident in the ground — and the attribution

is split here rather than left reading as one parenthetical governing both

[CREATURE_DEMAND_RULED_STATE_L11.json § ruled_demand_table → CH_07 → CR_0177].

sheet's to choose — see the fork ruling below.

notes and folklore_attribution_source adjudications), and @ CR_0039, @ CR_0178, @ CR_0179

for the chapter set and the anti-conflation partners · T0_Boss_Encounter_Registry @ BE_0007_S3

and @ BE_0007 · T0_RNG_Drop_Table @ DROP_CH07_ASURA_WAR_LORE,

@ DROP_CH07_PARAVANI_PRE_ENCOUNTER · T0_Environment_Grammar_Registry @ EG_0088, @ EG_0089 ·

T0_Chapter_Index @ CH_07 :: folklore_entities_in_chapter ·

T0_Hard_Lines @ HL_0002, @ HL_0003, @ HL_0004, @ HL_0005, @ HL_0007, @ HL_0123 ·

build/3d/schema/anatomy_reference/HUMANOID.json § sub_plans.human ·

build/3d/schema/creature_archetype_map.json § archetypes.NEEDS_CANON_BODY_PLAN (the local mirror

carrying the verified content; the creature_archetype_census.json this sheet's earlier draft cited

has no file in THIS repo and lives in the game repo, so the citation is re-pointed rather than

carried) ·

docs/spine/CH_07.md L3, L12, L17, L31, L52, L53, L73, L83, L102-103, L113, L169-173, L174-178,

L208-219, L227, L247, L252, L256, L264-265, L269, L271, L275, L281, L283, L294, L301, L329, L330,

L334, L335 · south_india.md L15-16, L39, L61, L109, L203, L245-248, L546-553, L1078-1080,

L1280-1281 · docs/proposals/bestiary/sheets/BE_0007_S3_soorasamharam_remnant.md (LANDED — see

the fork ruling)

FORK 2, RULED ON THIS SHEET'S FACE — THE ENCOUNTER SHEET IS ALREADY LANDED AND IT GOVERNS

THE BRIEF'S PREMISE IS STALE AND THIS SHEET REPORTS IT RATHER THAN BUILDING ON IT. The lane was asked

to propose where the full fight grammar lives, on the understanding that "the BE_0007_S3 boss-lane

sheet, when its wave runs, re-derives phases/rewards from the beat graph citing this sheet." That wave

has already run. docs/proposals/bestiary/sheets/BE_0007_S3_soorasamharam_remnant.md is LANDED and

registered [SHEET_REGISTER.md:53], and it already authors a complete fight grammar: the division

mechanic and its bounded count, the ARENA band, the world-response oppression, four tell shapes with

one answer each, the marker read, the four escalation channels, the opener tree, three breakable parts

with silhouette fractions, and the relocation rule.

**THE RULING THIS SHEET PROPOSES: FIRST-LANDED GOVERNS, and the division of labour follows from what

the landed sheet actually contains rather than from a plan.** The precedent is DIRECTOR_RULING_L12

RULING_4 ("FIRST-LANDED GOVERNS"), re-affirmed at DIRECTOR_RULING_L13 RULING_2 on the CR_0065

inversion. Concretely:

phases, the win and death rewards, the win_condition, the drops, the tell-shape roster, the answers,

the escalation channels and the part economy are its, and this sheet quotes them verbatim where it

needs them and changes nothing. No new tell shape, no new answer, no fourth elevation axis, no

re-cut of the bounded division count.

reads, in full: "humanoid spine and limb pairs as the root chain, at ARENA scale" and "declared

ROM: humanoid baseline at the root" [BE_0007_S3:43-48]. It names no joint, no chain, no segment

count and no number, and it says "Every part of this body has a joint roster because the doctrine's

governing law of part three requires it" [:44-45] — a statement that the roster exists rather than a

roster. That is not a criticism of the sheet, which predates the amendment: doctrine § 2.1's inline

licence EXPIRED at DIRECTOR_RULING_L12 RULING_3, after which a covered row DERIVES from its

archetype bucket and cites the path. § 2 below does that, and it is the single largest thing this

sheet adds.

bilaterally symmetric humanoid — the division plane "runs the body's long axis" and "each half is a

COMPLETE articulated body with its own spine chain and its own limb pairs rather than half a rig"

[:129-131], while "the whole's reach comes from a long forelimb pair that the division plane passes

through" [:133-134]. § 5 supplies the skeletal account that makes both true at once, declared as this

sheet's own resolution of the landed sheet's two clauses rather than as a change to either.

the boss registry, the spine, T3_Familiars_Named and HL_0005 [:33-35]. It did not read CR_0177,

which did not yet carry its adjudications. This sheet reads the row's

folklore_attribution_source, cultural_care_register and notes in full and § 7 builds them.

[:72-79] but ran no control. § 3 runs both, and the negative one is decisive.

here by file:line rather than restated as new; every element this sheet builds is one the landed

sheet does not contain. Where a later BE_ pass re-derives phases and rewards from the beat graph, it

cites this sheet for the body and the seam and re-derives nothing of them.

CONSISTENCY WITH THE LIVE BOSS ROW, CHECKED CELL BY CELL AND REPORTED. @ BE_0007_S3 carries

primary_verb strike, secondary_verb endure, axis_split 35/25/40, win_condition REDUCE,

has_nonkill_path FALSE, defeat_protected FALSE, phase_trigger composure_break,

intelligence_tier sorcerous, target_type mythological_monster_folklore_combat, and

composure_nodes "direct=asura_strike_down;puzzle_routed=abode_marker_read". The spine's join index

agrees — "BE_0007_S3 the Soorasamharam-remnant (STRIKE/ENDURE, 35/25/40 patterns-and-preparation ...

the peacock pre-encounter its win reward)" [CH_07.md:113] — and so does the region page's derived

index [south_india.md:109]. The J10 reconciliation is the reason: strike/endure was RESTORED

2026-07-15 within the devotional frame, "the Soorasamharam festival's own victory reenactment IS the

moveset" [CH_07.md:329]. This sheet asserts nothing against any of those cells.

2. base_species

of the archetype census, which is where the verified content actually lives on this side — places

CR_0177 in NEEDS_CANON_BODY_PLAN (42 rows), and both anatomy tables scope those rows OUT: "The 42

NEEDS_CANON_BODY_PLAN rows and the 7 PROVISIONAL folklore rows are OUT OF SCOPE — their body plan

is a canon decision that is already boarded, and a reference table that guessed one would be

inventing creature canon" [HUMANOID.json § scope]. The row's own mesh_tags.morphology reads

"folklore", which is a provenance word and not a body plan.

The landed encounter sheet rules "a powerful humanoid frame in the register the tradition's own

asura imagery uses" [BE_0007_S3:40-42], and its plate brief specifies "a powerful humanoid

figure at arena scale in the tradition's own asura idiom" [:234-235]. Under first-landed-governs

that IS the archetype. So this sheet derives from HUMANOID.json by ARCHETYPE COVERAGE — the test

doctrine § 2.1 actually states ("or whose archetype a bucket covers") — and § 9 boards the census

assignment as the registry-side decision it is.

build/3d/schema/anatomy_reference/HUMANOID.json § sub_plans.human, whose reference_taxon is

*Homo sapiens* and whose own one-line warning is the reason this block matters: "Every range below

is clinical goniometry, which means a pose outside it is not stylised — it is an injury"

[§ one_line].

NON-OPPOSABLE hallux; a PRESENT, LOAD-BEARING CLAVICLE; and three spinal curves — cervical

lordosis, thoracic kyphosis, lumbar lordosis — with the table's own note that "a straight-spine rig

reads as a mannequin" [§ segments]

extension -60 to 180, abduction 0 to 180, axial rotation ±90; scapulothoracic elevation 0 to 60,

protraction-retraction ±30, upward rotation 0 to 60, with the note that "the last 60 deg of overhead

reach comes from HERE, not the shoulder"; elbow 0 to 150, where negative is hyperextension and

normal only to about -10; radioulnar pronation-supination ±80; wrist flexion-extension -80 to +70

and radial-ulnar deviation -20 to +30; hip flexion-extension -30 to 120, abduction-adduction -30 to

+45, axial rotation ±45; knee 0 to 135 with axial rotation ±10 AVAILABLE ONLY IN FLEXION because the

screw-home locks it at extension; ankle dorsi-plantarflexion -20 to +50; subtalar inversion-eversion

-35 to +15, which is "the joint that lets a foot sit on a slope" and is load-bearing on this

encounter's ground; cervical spine summed at ±45 flexion-extension, ±45 lateral, ±80 axial; and

thoracolumbar summed at 0-80 flexion, 0-30 extension, ±35 lateral, ±45 axial, SUMMED over the chain

because "per-vertebra it is a few degrees, which is why a two-bone spine cannot make a real bend"

[§ joints, all values AAOS1965 / NORKIN2016]

DOUBLE SUPPORT present — "its absence is what makes a walk read as a run"; run at duty factor

0.2-0.4 with a flight phase and no double support. The solver takes foot IK to ground with a

subtalar roll, pelvis counter-rotation against shoulder counter-rotation, and arm swing anti-phase

[§ gait]

of at least 4 bones each; ten hand-digit chains of at least 3; clavicle_L and clavicle_R

REQUIRED bones; biped alternating gait sequence; bilateral symmetry required [§ rig_conformance]

reaches under its own power is smaller by an amount the sources do not separate [§ unmeasured].

mechanise it. The landed A1 spend rules that "the whole form cannot move quickly on a hillside,

cannot turn sharply, and telegraphs across its whole length ... at this size the turn radius is

wide, the recovery is long, and nothing it does is quick. That is what makes it fightable by a

ten-year-old" [BE_0007_S3:46-48, :104-106]. THE MECHANISM, derived rather than asserted: THE

DEGREES ABOVE DO NOT CHANGE. Goniometric ranges are dimensionless — a scaled body's elbow still runs

0 to 150. What changes is the TIME to traverse them, because a limb's rotational inertia grows with

the fifth power of its length while the cross-sectional area that carries muscle force grows with

the square. Vril alters real physics rather than exempting a body from it, so the same 150-degree

flexion that a person spends in a fifth of a second is spent here in something the player can watch

begin, name, and answer. Every telegraph in the landed grammar is that arithmetic, and the

animation lane should read it as a rule rather than as a feel: SCALE THE CLOCK, NEVER THE ANGLES.

chase; it OPPRESSES" [BE_0007_S3:49-50]. The gait is the bucket's walk with double support always

present, which is why it never reads as running even when it covers ground fast.

large to be a person and too person-shaped to be a rock" [BE_0007_S3:51-52]. § 5 adds one thing to

it that the landed sheet's silhouette does not yet carry, and only one: the seam.

3. lore_source

paraphrased: "authentic living tamil shaiva tradition ... the asuras of the murugan war cycle are the

tradition's OWN condemned demons ... surapadman is the demon who oppressed the world until the war

god murugan was born expressly to end him ... and his brothers tarakan and singamukhan are named

with him in the same breath ... and the soorasamharam is the tradition's own commemoration of that

slaying which is ritually reenacted and is commemorated at tiruchendur"

[@ CR_0177 :: folklore_attribution_source].

oppressed the world until the god was born expressly to end him; the god struck him with a spear,

and the demon, splitting as he fell, became the peacock the god rides and the bird on his banner"

[CH_07.md:31].

restored asura-war remnant-encounter — a manifestation of the demon-host of the Soorasamharam (the

Soorasamharam itself, the ritual slaying of Surapadman, is commemorated at Tiruchendur)"

[CH_07.md:52]. The row types it the same way and says why: "spine and boss row both type the

encounter as a Surapadman-CLASS asura and a Soorasamharam-REMNANT ..., never the demon of the myth

himself, whose slaying is complete in the tradition's telling" [@ CR_0177 :: notes].

precincts guarded not desecrated, and the goddess and the deity are never targets" [CH_07.md:252];

and the region page restates it — "The asura the tradition itself condemns is the demon that is

fought, in the Soorasamharam's own demon-slaying spirit, and the god who slays him is the honored

slayer" [south_india.md:549-551].

Singamukhan) are the Tamil Shaiva tradition's OWN condemned demons, whose slaying in the

Soorasamharam is the reason for Murugan's existence (the Kandha Puranam; Clothey; Zvelebil) — faced

as the demon-war the tradition itself commemorates" [CH_07.md:256].

T0_Source_Material_Registry for "clothey", "zvelebil" and "kandha" returns ZERO ROWS. The

three named substrate authorities of this being are not on the shelf. The region page says so

itself and ranks the gap: "The Murugan tradition and the demon-war cycle — the Palani abode and

the Soorasamharam, which the spine restores and which this page carries only at the spine's own

words" [south_india.md:1280-1281], listed under MUST-ACQUIRE.

Becomes All Eyes*, 544 chunks, "the keystone ... the only source on this shelf that carries the

region's living practice at technique grain" [south_india.md:1247] — resolves and is quotable.

It simply carries nothing about this being.

block under this region slug is a parapsychology review that is a mis-sort, and a keyword scan of

all 379 of its chunks "returns zero occurrences of tamil, siddha, siddhar, kalari, kalarippayattu,

marma, murugan, boganathar, zvelebil and wootz" [south_india.md:1259-1264]. "[T]he presence of a

parapsychology corpus under a region whose canon carries a Siddha tradition is precisely the

adjacency the anti-fringe rule guards" [:1267-1269]. Nothing on this sheet touches it.

CONTROL BEHIND IT.** The Kandha Puranam's account of Surapadman's forms, his brothers' features and

the sequence of the war is not in this corpus [BE_0007_S3:72-79]. So this sheet designs the

REMNANT MANIFESTATION'S BODY and portrays no named demon; it assigns Surapadman no invented

iconography; it gives Tarakan and Singamukhan no forms, no appearances and no phases, exactly as the

row rules ("ONE COHORT ROW, NOT THREE ... carried IN PROSE as the myth holds them"

[@ CR_0177 :: notes]); and every structural claim in § 4 and § 5 is declared as the game's own.

When the specialist acquisition lands, the manifestation's forms are superseded in place.

produce an empty sheet.** HL_0002, verbatim: "Claims about a LIVING culture's tradition, ritual,

language, or sacred practice trace to primary-source record (floors 1-3). Mythic monsters, boss

encounters, dungeon/site design, and adventure content carry fantasy license — grounded in the

tradition's own material but free to extrapolate for playable spectacle; they are not bound to a

citable source" [T0_Hard_Lines @ HL_0002]. The tradition's own material here is the splitting, the

oppression, the abode and the honored slayer, all cited above. The body that carries them is ours.

account of the division, the seam interlock chain and its ROM, the seam as a pre-division telegraph,

the scale-the-clock rule, the head-as-a-join, the weight-transfer account, and the myth-anchoring of

the first division's count at two.

shapes, the marker read, the physical oppression and the hill's silence, the three breakable parts.

4. elevation

THE AXIS SET IS BE_0007_S3'S AND IS CARRIED UNCHANGED. THIS SHEET ADDS NO AXIS. Under

first-landed-governs, adding a fourth axis to a landed elevation would be exactly the drift the L12

ruling exists to stop. The three, quoted:

tradition's own most striking fact turned into a fight. The demon SPLIT as he fell. So this

manifestation's defining behaviour is that a correct strike DIVIDES it, and each division is still

it" [BE_0007_S3:88-93]. The division is earned on the answer the markers teach, never random

[:95-96]; it is bounded at twice — whole, halves, quarters [:97-99]; and each division loses

something permanently — "The whole has reach; the halves have speed and no reach; the quarters have

neither and only close" [:100-102].

at each division [:104-106]. § 2 supplies its mechanism.

ON THE HILL, not on the creature: the air over the ascent goes still and hot, the path's dust does

not settle, the scrub bends away from the line the manifestation holds, and — the strongest and

cheapest beat — THE HILL GOES QUIET. Birds first, then insects" [:108-113]. The hill's own canon

supports every clause: EG_0088 carries exposure_hazard heat and the region page records "wind

light on the plain, strong at the Palani ascent" [south_india.md:203], so a wind that STOPS on a

slope known for wind is a state change the place itself makes legible.

MECHANISM, the landed sheet's: "it is a remnant of a defeated being, and its nature is the defeat

itself. It divides because dividing is what happened to it, and it oppresses because oppression is

what it did" [:115-117].

ONE-LINE READ, the landed sheet's: "a figure too large for the hill that comes apart along a line when

it is struck properly, and both halves keep coming" [:85-86].

**THE RECIPROCAL SILHOUETTE DISCRIMINATION AGAINST CR_0179, STATED FROM THIS SHEET'S SIDE SO THE

CONTRACT BINDS BOTH WAYS.** CR_0179's § 1.2 rules that a pey must never read as a small asura at the

combat camera; the same rule from here is that THIS BEING MUST NEVER READ AS A LARGE PEY — at the

combat camera the asura is MASS, ARMATURE AND TWO-BODIES-IN-ONE (a solid arena-scale outline with a

seam flickering down its vertical centre, planted through both leg pairs, one figure that becomes

two), while the pey register is EMACIATED SINGLE FRAMES IN COHORT (thin true-scale outlines under

human mass, hanging by oversized hands off wall-lines, never singular and never planted). One is a

count that goes DOWN from one; the other is a count that was never one. Never pooled in a plate,

never pooled in a mesh set, and never lit or staged so that the two registers share a read

[CR_0179_pey_and_bhutam.md § 1.2].

EMISSION: NONE, the landed sheet's ruling, and this sheet confirms it against the row: no

defender_element is set, no emitting family is claimed, and the form reads against the hill's hard

daylight by OCCLUSION and silhouette [:119-121]. The Palani ascent is bare, hot and bright

[EG_0088 :: exposure_hazard heat; south_india.md:203, the dust-gold plain palette and the wind

strong at the Palani ascent], which is precisely the light in which a glow reads as nothing at all.

WHAT THIS SHEET ADDS TO THE ELEVATION IS NOT AN AXIS BUT THE STRUCTURE A1 AND A2 ALREADY OWE. The

landed A2 requires a body that divides into complete bodies while losing a limb pair that the plane

passes through; the landed A1 requires a body whose telegraphs read across their whole length. Neither

is buildable from "humanoid spine and limb pairs." § 5 is that debt paid.

THE ONE-MARK QUESTION, NAMED AND DELIBERATELY NOT ANSWERED HERE. The landed elevation spends no

A9, so this being's identity rests entirely on the division and the scale. That is a legitimate choice

— a creature whose signature is a structural event rather than a feature — and it is NOT this sheet's

to change. It is recorded in § 9 as a landed-sheet observation for that sheet's next touch, in the

DIRECTOR_RULING_L13 drift-boarding form, and nothing here acts on it.

5. articulation_delta

curves, two arm chains and two leg chains of at least four bones each, ten three-bone hand digits,

both clavicles, sixty deform bones minimum, ROM as declared — carried at ARENA scale with the clock

scaled and the angles untouched.

**THE DIVISION PLANE, AND THE SKELETAL ACCOUNT THAT MAKES THE LANDED SHEET'S TWO CLAUSES TRUE AT

ONCE.** The landed sheet requires the plane to run the body's long axis, each half to be a complete

articulated body rather than half a rig, and the long forelimb pair to be the thing the plane passes

through and the thing the halves lose [BE_0007_S3:129-135]. A sagittal hemisection of a person

cannot deliver that: it yields two half-bodies with one arm and one leg each. THE RESOLUTION, DECLARED

AS THIS SHEET'S OWN INVENTION: the whole is not one body that splits. **It is TWO COMPLETE BODIES

INTERLOCKED ALONG THEIR MEDIAL FACES, and the long forelimb pair is the FUSED STRUCTURE THAT SPANS THE

SEAM** — one pair of limbs belonging to both halves and to neither, rooted in the interlock rather

than in either half's shoulder girdle. That single decision makes everything the landed sheet says

literally true: the plane runs the long axis; each half is complete because each half was always a

whole; and the reach dies at the division because the limbs that carried it had no owner to survive

into.

**THE SCOPE OF THAT ACCOUNT, STATED BEFORE ANYTHING IS BUILT ON IT, BECAUSE THE LANDED BOUND IS TWO

DIVISIONS AND THIS SHEET SPECIFIES ONE.** The landed sheet bounds the mechanic at twice — "whole,

halves, quarters, and the quarters are what the player finishes" [BE_0007_S3:97-99]. Everything

below — the interlock chain, its plane, its ROM, its weight account, its telegraph — describes the

FIRST division only. **THE SECOND DIVISION, HALVES INTO QUARTERS, IS AN OPEN ARTICULATION DEBT AND

THIS SHEET DOES NOT CLOSE IT.** It is named affirmatively rather than papered over: there is no second

plane specified, no seam chain inside a half, no ROM, and no structural account of what a quarter is.

The recursive read was weighed and NOT taken — making each half itself two interlocked conformant

bodies would multiply the whole form to four bodies and its visible limb pairs past the count the

resting silhouette and § 4's landed 0.14 forelimb fraction can carry (see the limb-pair count below),

so it would buy the quarters a structure at the price of breaking the whole. What the debt's owning

lane inherits, so the next pass starts from a specification rather than a blank: the quarters must

each be a conformant body, because the landed rule is that "each division is still it"

[BE_0007_S3:90-93]; the quarters' "only close" moveset [:100-102] must fall out of what the second

plane STRUCTURALLY removes, exactly as the halves' reach loss falls out of the fused pair; and any

seam added there obeys the same law as the one below — SCALE THE CLOCK, NEVER THE ANGLES, so a

quarter's smaller interlock keeps the ±4-degree per-segment excursion and gets faster, never wider.

§ 9 boards it, and § 9's buildability claim is downgraded to one-of-two to match.

articulation law bites hardest and an unarticulated seam is exactly what it hides. Five segments,

SEAM_00 at the pelvic keel through SEAM_04 at the cranial vertex, each parented into both

halves' thoracolumbar chains. Each segment is a single-axis hinge with an excursion of ±4 degrees in

the CORONAL plane only — no flexion, no axial rotation, no translation. Summed over five segments

that is ±20 degrees of lateral play across the whole trunk, which is enough to be seen and far too

little to be a bend. WEIGHT: at rest each half's own leg pair carries its own side and the seam

carries pure shear, so nothing about the whole form's stance depends on the interlock holding.

Because the interlock has real play, the standing body carries a hairline of lateral motion running

its full height — a slow open-and-close on the idle clock, visible in black fill at combat distance

as a flicker along the vertical centre of an otherwise solid mass. The player therefore SEES THE

PLANE BEFORE IT EVER OPENS, which is what makes the division read as a mechanism the creature has

rather than an effect the game does. It also gives the landed sheet's "planted set" tell

[BE_0007_S3:146] an anatomical home: the set is the seam CLOSING to lock, which is why the strike

lands then and only then.

carries ONE head at the vertex where they meet — visibly a join rather than a skull, and the last

thing to part. This is declared as the game's own and is deliberately kept as a STRUCTURAL statement

rather than an iconographic one: no face is designed here, no invented Kandha-Puranam feature is

assigned, and the plate brief in § 8 specifies the join without specifying a physiognomy.

MESH LANE TO DERIVE: THREE ARM PAIRS — SIX ARMS — AND TWO LEG PAIRS.** Two complete interlocked

bodies contribute two ordinary arm pairs and both leg pairs between them, and the interlock-rooted

fused pair is the third. It follows from the two-bodies account and from nothing else, and a lane

that builds four arms has built a different creature. **IT DOES NOT BREAK THE LANDED RESTING

SILHOUETTE** — "a standing figure on a hill that is too large to be a person and too person-shaped

to be a rock" [BE_0007_S3:51-52] — because of how the arms are CARRIED: the four ordinary arms

hang inboard, close against the trunk mass, and at combat distance in black fill they read as the

body's own bulk rather than as limbs. Only the fused pair leaves the outline. **THAT IS ALSO THE

RECONCILIATION WITH THE LANDED 0.14 FORELIMB FRACTION** [BE_0007_S3:173-176]: 0.14 is the share of

the silhouette that arms occupy, and only the fused pair contributes to it, so six arms and one

0.14 arm-part are the same statement rather than two. THE COUNT BECOMES LEGIBLE AT THE DIVISION and

not before: as the halves part and take their corrective strides, the four inboard arms come off the

trunk and the player sees for the first time that the thing had six — a revealed count rather than

an added feature, which is the same discipline the seam's flicker follows. Declared as the game's

own structural consequence and NOT as an iconographic claim: no multi-armed asura iconography is

asserted, no invented Kandha-Puranam feature is assigned, and § 3's citation limit is untouched.

became the peacock the god rides and the bird on his banner" [CH_07.md:31] — two things. So the

whole becomes two halves, which is myth-anchored; the second division into four is the landed

sheet's own bound [BE_0007_S3:97-99] and is the game's own. The distinction is stated so no later

lane reads the quarters as a tradition claim.

legs at rest, the separation is authored as the interlock releasing and each half taking ONE

corrective stride to recover its own centre of mass — a real, animatable event with a duration. The

landed sheet's warning is exactly right and is honoured by construction: "If the halves are two

copies spawned at the seam position, the encounter has failed at the asset layer and a player will

see it in one frame" [:131-133].

radioulna are lengthened past the bucket's human proportion; the HAND IS NOT, and is kept at human

ratio to the body. A longer segment swept through the same glenohumeral excursion at the same

angular rate covers more ground in the same time, so the telegraph gets LONGER IN DISTANCE while

staying identical in duration — which is precisely the landed sheet's "telegraphs across its whole

length" [:105-106], and it is why the answer is a dodge given away by the creature's own size

[:145].

horns, no mark. Three axes, one new structure, and the structure is an interlock.

6. fight_grammar

THE GRAMMAR IS BE_0007_S3'S AND IS CITED, NOT RE-AUTHORED. Verbs strike and endure, axis 35/25/40,

counters "hold the ascent in the demon-slaying tradition's own bearing; read the abode's markers to

answer the asura's forms" [@ BE_0007_S3 :: counters; BE_0007_S3:140-141]. The four tell shapes and

their single answers are at [BE_0007_S3:143-148]; the marker read at [:153-157]; the escalation and

its four channels at [:159-168]; the opener and variants at [:170-171]; the breakable parts at

[:173-176]; the chain-meter statement at [:178-179]; relocation and turf at [:181-183].

WHAT THIS SHEET SUPPLIES IS THE BODY ORIGIN LAYER THAT DOCTRINE PART FOUR ASKS FOR AND THAT THE

LANDED TABLE CARRIES ONLY AS PROSE — which segment initiates, which chain loads, which mass moves. The

shapes and the answers are unchanged; nothing below adds, removes or re-aims one.

the landed tell shapethe body origin, derived from § 2 and § 5the landed answer, unchanged
the long reach across the paththe fused forelimb pair loads from the interlock rather than from a shoulder girdle, so the wind-up is visible at the body's CENTRE and travels outward; the thoracolumbar chain flexes toward the slope and the whole mass leans downhill into itDODGE — a whole-body commitment, given away by its own size
the planted setboth leg pairs drive into the hill through the subtalar joints that let a foot sit on a slope, the pelvis squares to the ascent, and THE SEAM CLOSES AND LOCKS — the ±4-degree play going to zero along the full height, which is the division window becoming availableSTRIKE the spear-line the markers name — this is the division window
the oppression deepeningnot on the creature at all: the hill's air stills, the sound goes out of the slope, and the only body change is that the form STOPS MOVING, which makes the seam's flicker the one thing still alive in the silhouetteENDURE — hold the climb's bearing; do not break pace
the halves closing from two sidestwo complete bodies with their own spine chains and leg pairs and no fused limb between them, each recovering from its corrective stride, converging low and fastkeep the SLOPE between them; the hill is the answer

ONE SHAPE, ONE ANSWER throughout — the landed sheet's own property, preserved because nothing was

added.

PART ECONOMY HOME: BODY, declared by name per doctrine § 2.4, and the landed sheet's three parts

are carried with their fractions unchanged — the FORELIMB PAIR at about 0.14 of silhouette (the part

the division plane passes through) and the two PLANTED LEG ANCHORS at about 0.08 each, where breaking

one costs the planted set and therefore the creature's own division window [BE_0007_S3:173-176].

THE SEAM IS EXPLICITLY NOT A FOURTH PART: it is a READ, it has no break threshold, no break event and

no material, and calling it a part would be adding to a landed economy. It is named here so no later

lane mistakes a visible feature for a breakable one.

THE MARKER READ'S CARE CONSTRAINT, RESTATED AND THEN SHARPENED WITH A LINE THE LANDED SHEET DOES NOT

CARRY. The landed rule is that "the markers are READ, never used — they are not weapons, not power-ups

and not consumed. The player learns the answer from the place and then performs it themselves"

[BE_0007_S3:155-157]. THE SHARPENING: THE VEL IS NEVER A PLAYER ITEM. The spear that slays the

asura is Thread 20 sacred-object substrate — "the Tamil kattari curved blade and the Murugan vel

carried as sacred-object substrate (Thread 20), not a roster acquisition" and "sacred_objects include

the kattari blade, the Murugan vel, and the wootz billets, referenced not reproduced"

[CH_07.md:283; :275]. So the markers tell the player where the spear-LINE falls on this form, and the

player performs it with their own weapon; no vel is picked up, equipped, crafted, dropped or wielded

at any point. That is a real constraint with a canon home and it is the difference between honouring

the myth and looting it.

TIER DISCIPLINE, per the row: reading the abode's markers is a Neophyte-register read of markers and

lore, never a tier-six vril-signature read; the substrate-perception onset waits at the Ch-15

Adeptus-Minor crossing [@ CR_0177 :: notes; CH_07.md:17; :281; HL_0028 / HL_0017].

WHERE THE FIGHT HAPPENS, AND WHERE IT DOES NOT. The encounter sits on EG_0088 Abode-Hill Pilgrim

Ascent Way (protected, CONSTRICTION, exposure_hazard heat, env_boss_verb_disabled TRUE,

alteration_permanence_default hard_blocked) and OPENS ONTO EG_0089 Abode Summit Shrine Court and

Measure Pocket, whose tactical_affordance_class is NONE with no verb, "the encounter resolves on the

way up and the summit is the opening rather than the arena" [@ EG_0089 :: notes]. Two consequences

the asset and encounter lanes should read as rules: the hill cannot be permanently altered by the

fight, and NO PART OF THE FIGHT HAPPENS IN THE SHRINE COURT.

7. care_read

Malayali communities, the Siddhas, the kalari and temple communities, the smiths and healers, the

pilgrims and the honored Raghavendra are allies and protected non-targets [CH_07.md:252;

@ CR_0177 :: cultural_care_register]. The fight's participants are the protagonist and the

manifestation, and nobody else is on the hill.

material BECAUSE THE TRADITION ITSELF CONDEMNS HIM — the spine states it as the deities-are-real

nuance in the same breath as the honored pole: "the asura is villain material because the tradition

itself condemns it (the deities-are-real nuance)" [CH_07.md:330]. has_nonkill_path FALSE is

canon's ruling, not this lane's harshness, and softening it into a banishment would substitute the

loop's squeamishness for a living tradition's own theology.

row as the most defanged in Ch 2-40; the restoration is canon: "restored WITHIN the devotional frame

— the Soorasamharam festival's own victory reenactment IS the moveset (the Sanni-Yakuma Floor-1

class: the living tradition is the combat-method, never mocked)" [CH_07.md:329]. And the golden

re-enrichment's finding: "the v2.4 build was saturated in Murugan iconography (Palani, the vel, the

peacock) yet omitted the mythic demon-war that gives all of it meaning, leaving only human and

abstract antagonists" [CH_07.md:330]. The full staged killing stands, at weight, with the oppression

real and the wrong-spirited ascent deepening it [FORK_CH07_B, CH_07.md:217].

myth's ending is a transformation into a sacred vahana, and a design that let the division's halves

read as peacock-shaped, or that dropped feathers, or that produced the bird as a combat outcome,

would turn a living faith's honored mount into loot. THE RULE: **the halves are not bird-shaped, no

plumage, feather, eyespot or avian silhouette appears anywhere in the encounter, and the peacock

never enters the fight's frame.** Canon already staged it that way and the rule simply refuses to

drift from it — the bird arrives at the SUMMIT, after, at a separate beat on separate ground

[CH07_B12, CH_07.md:174-178; EG_0089], and the pre-encounter is a REGISTRATION rather than a bond

("A PRE-ENCOUNTER REGISTRATION, NOT A BOND"

[@ DROP_CH07_PARAVANI_PRE_ENCOUNTER :: notes]). The encounter borrows the FACT of splitting; it

never stages the OUTCOME.

reason the cycle exists, never a combat target, and never taking a creature, character or boss row;

Meenakshi is honored and inviolate; no deity name is ever pinned onto a mortal NPC or onto a

controllable familiar, which is why the F2 familiar is named PARAVANI for the vahana and not for the

god [@ CR_0177 :: cultural_care_register; CH_07.md:335]. Both are typed HONORED_inviolate in the

chapter index's own folklore cell [T0_Chapter_Index @ CH_07 :: folklore_entities_in_chapter].

"[a] culturally protected space still delivers a real fight. Combat there comes from antagonist

intrusion AND/OR the tradition's own monsters ... a protected site never needs a manufactured

Velheim raid to be playable, and is never left combat-absent." This encounter is that rule's

textbook case: the abode is inviolate, the shrine is referenced not reproduced [HL_0007;

EG_0089], and the fight comes from the tradition's own condemned demon on the unconsecrated ascent

rather than from a manufactured raid.

register, which is CR_0179 and is a different thing at a different grain — the spine names the

asuras and the pey/bhutam as two distinct items on one folklore line [CH_07.md:83], gives only the

asura a BE_ row and a fork, and the roster rules that pooling them "would collapse a condemned

MYTHIC DEMON of a named war-cycle into an ambient spirit register and lose both"

[@ CR_0179 :: notes]. And it is never pooled with the honored Siddha deathless masters, who are

HUMAN and live in T0_Character_Index as CHAR_0048 Bogar.

re-opens them silently.** NOT a generic fantasy demon; NOT Hollywood horror shorthand [HL_0005];

NOT a villainized deity; NOT a manufactured or forced construct of any antagonist apparatus — the

row refuses CONSTRUCT_NOT_PERSON "decisively ... NO MAKER EXISTS — the demon is the tradition's own

and predates every operative chain" [@ CR_0177 :: notes], and importing a construct token would

drag in an apparatus the chapter's reveal binding keeps dark.

own requirement that a designed mythic being never route through a photographic-wildlife or

photographic-human template. THE CHECK THIS SHEET ADDS: the row's mesh_pipeline_type reads

image_to_3d, which is the concept-plate-then-mesh route and is compatible — provided the plate is

attended. The style cell reads mythologically-stylized and matches CH_07.md:271's exclusive list, so

there is NO style mismatch to board on this row. The hill, the path and the markers are

REFERENCE_BUILT.

(1) IS THE KILL OVER-REACHED? The landed sheet ran this pass and refused it on three grounds — the

row's own has_nonkill_path FALSE, the tradition's own central act being a slaying commemorated

annually, and the fact that the only care-shaped alternatives available are restorations of the

de-cultured abstract Palani Ascent that the golden re-enrichment replaced [BE_0007_S3:208-218].

This sheet ADOPTS that pass rather than re-running it, and adds nothing.

(2) IS THE BODY-SIDE WORK A SOFTENING? Examined and RULED NO. Nothing in § 2 or § 5 reduces the

encounter: the scale-the-clock rule makes the telegraphs longer in SPACE, not weaker; the seam adds

a read the player must earn rather than a warning the game gives; and the fused-limb account makes

the reach loss a real structural subtraction instead of a stat change. The design got harder to

build and no easier to fight.

(3) SHOULD THE HEAD-AS-A-JOIN HAVE BEEN PULLED BACK AS TOO GRUESOME FOR A TEN-YEAR-OLD

PROTAGONIST? REFUSED. This is a rated-M action RPG, the protagonist's age is an internal

designer-only fact that is never surfaced [CH_07.md:13], and the being is a condemned demon whose

defining event is coming apart. Pulling it back would be care-tier inflation applied to the one

thing the tradition itself is unambiguous about.

(4) SHOULD THE PEACOCK HAVE BEEN PUT INSIDE THE FIGHT FOR SPECTACLE — the transformation staged as

the kill's payoff? REFUSED, and this is the one place the pass runs toward MORE care and finds

something worth taking. The spectacle is real and the cost is exact: it would make a living faith's

honored vahana a combat drop. Canon already separated them onto two beats and two EG surfaces, and

the separation is better design as well as better care — the summit's quiet after the hill's

oppression is the chapter's best beat and a bird bursting out of a corpse would spend it.

account of the whole, the fused forelimb pair spanning the seam, the five-segment SEAM interlock

chain and its ±4-degree coronal ROM, the seam's idle flicker as a pre-division telegraph, the seam

closing as the planted set's anatomical home, the head as a join, the one-stride weight transfer,

the lengthened humerus-and-radioulna with a human-ratio hand, the six-arm / two-leg-pair count of the

whole form and its inboard carry, and the scale-the-clock-never-the-angles rule.

CITED: the splitting into the peacock and the banner-bird, the demon who oppressed the

world until the god was born to end him, the spear, Palani as one of six abodes, the Soorasamharam's

commemoration at Tiruchendur, Murugan as honored slayer, the Surapadman-CLASS and REMNANT typing,

and the brothers named in one breath. DERIVED: the skeleton, joint ranges, gait formalism and rig

conformance from HUMANOID.json, carried as passive-envelope bounds. CARRIED UNCHANGED FROM THE

LANDED ENCOUNTER SHEET: the three axes, the division mechanic and its bound, the four tell shapes

and answers, the marker read, the oppression and the hill's silence, the part economy, the

escalation channels, the opener tree and the relocation rule.

family or pattern; display_name is the boss row's own ("the asura of Palani"); the Grand Sage is

dark; the Sealed Letter's fairy-realm script is not identified and does not appear

[CH_07.md:115; :301]. Raghavendra's UNAWARE status is untouched — he is not in this encounter.

8. plate_brief

Written affirmatively. Every line names what IS in frame. The landed sheet's five plates

[BE_0007_S3:231-245] stand and are not restated; the four below are the BODY plates this sheet's

§ 2 and § 5 dispatch, and they exist to make the skeletal decisions unmissable to the concept lane.

front-on at fighting distance, with a single hairline of lateral offset running the full height of

the trunk from the pelvic keel to the crown — the two interlocked bodies readable as two bodies for

the first time, while the outline still reads as one figure.

CENTRE LINE in the interlock rather than at either shoulder, humerus and forearm long, the HANDS at

ordinary human proportion to the body so the reach reads as arm.

junction — the seam continuing up through it and closing — with the composition holding at half

figure and above, in the tradition's own asura idiom and with no invented facial iconography

proposed.

the seam locked to zero, and one half mid-corrective-stride, all three at one scale. JUDGED IN BLACK

FILL ALONE ON THREE CRITERIA: whether the seam's flicker survives; **whether the whole's SIX ARMS

read as trunk bulk in the standing plate and as six only in the parting plate**, which is the

inboard-carry rule of § 5 and the thing that keeps the landed "too person-shaped to be a rock"

outline true; and whether the fused pair alone accounts for the arm share of the outline, which is

the ~0.14 part fraction seen from the silhouette side.

markers. No plate depicts Murugan, the shrine or its interior, any worshipped image, or a pilgrim in

the fight. **NO PLATE IN THIS BRIEF — F, G, H OR I — CONTAINS A FEATHER, A PLUME, AN EYESPOT OR A

BIRD**, the peacock constraint of § 7 applied to the concept lane first, because a plate is what a

generator conditions on. **THE SCOPE IS THIS BRIEF AND NOT THE CHAPTER'S PLATE SET: the landed

BE_0007_S3 PLATE E STANDS AS LANDED** — "THE SUMMIT AFTER: a peacock on the high ground of the abode

in clear light, entirely itself, the hill quiet in the ordinary way again" [BE_0007_S3:241-242] —

and it is exactly what § 7's constraint protects rather than something it forbids, because the

separation of the bird onto the summit, after, on separate ground is the constraint's whole content.

A blanket no-bird rule stated without scope would have killed a landed plate, which is the defect

this line exists to prevent.

9. boarded_debts

The census places the row in NEEDS_CANON_BODY_PLAN and both tables scope those rows out because "a

reference table that guessed one would be inventing creature canon" [HUMANOID.json § scope]. The

row's mesh_tags.morphology reads "folklore", which is a provenance word standing where a body plan

belongs. THIS SHEET PROPOSES HUMANOID on the landed encounter sheet's own ruling and writes

nothing; the assignment is the registry lane's, and it is this sheet's highest-priority board because

the rig-conformance gate has nothing to check against until it lands.

this sheet's earlier draft drew was false and is corrected here rather than left standing.** The

cell is genuinely empty and the row genuinely has a body this sheet specifies, so the debt is real.

What is NOT true is that it is the chapter's lone blank: **@ BE_0007_S2 carries an empty

morphology (and an empty style) as well**, verified live, so two of the chapter's four boss rows

are blank and BE_0007 and BE_0007_S1 both read "bipedal". Nor is the creature row blank —

@ CR_0177 :: mesh_tags.morphology reads "folklore", which is a provenance word standing where a

body plan belongs (the board above) and is a different defect from an empty cell. The corrected

finding: the chapter has a blank-morphology CLASS across its boss rows, and the mythic-being row is

the one where a blank costs the most because § 5 supplies the body it should hold.

a joint roster, a ROM without numbers and a chain without names [BE_0007_S3:43-48]. It is not

edited here — first-landed governs and the block is superseded in place by § 2 above rather than

overwritten. BOARDED AS LANDED-SHEET DRIFT for that sheet's next touch, in the

DIRECTOR_RULING_L13 RULING_7 form.

event (the division) and a scale band. Recorded as an observation for the encounter sheet's next

touch, not acted on, and named here only so a later critic does not read the absence as an oversight

by this lane.

three with silhouette fractions and this sheet supplies the skeleton they attach to. The landed

sheet's own finding stands and is cited rather than re-raised: within one chapter pair the same

empty column means two opposite things on different rows [BE_0007_S3:249-253].

named why the schema may not be shaped to hold this row's answer: the escalation is on the

ENVIRONMENT and the phase voice is an ABSENCE of ambient sound [:254-256]. Cited, not re-raised.

temple polychromy, laterite, Ghats green" — which is the chapter's architecture-and-craft palette

[CH_07.md:271] pasted onto a being that canon describes as a manifestation on a bare hillside.

Wootz steel and temple polychromy are wrong nouns for this body and a generator reading the cell is

being fed the forge and the gopuram. Same contamination class as CR_0143/CH_05 and the CH_10 fauna

rows; the identical string sits on CR_0039, CR_0179 and BE_0007_S3.

chapter's era string pasted onto a being the myth places before the god's birth. It is the

chapter-paste class rather than the CH_10 BCE-for-BP unit-error class, and it also carries a U+FFFD

replacement character where an em-dash belongs — the same encoding defect as on CR_0039 and CR_0179.

HAS has_nonkill_path FALSE and win_condition REDUCE. The row already disposes it — the verb is

"taken on its PURGE side", with FIGHT weighed and refused because the two live FIGHT rows are

ordinary huntable fauna [@ CR_0177 :: notes] — so this is NOT re-litigated. It is boarded once, as

a SCHEMA observation: the roster's collect vocabulary has no token for "put down permanently", and

the row had to reach for a containment word to say a killing.

the landed sheet already names this row as the best rail candidate in its set [:257-258].

What this sheet ADDS to that debt is a PARTIAL narrowing, and the claim is stated at its true size:

**the FIRST of the mechanic's two divisions is now a buildable specification — ONE RIG CONTAINING

TWO CONFORMANT BODIES joined by a five-segment interlock — and the SECOND IS NOT.** One of two, not

the mechanic. The owning lane for the specified half is the creature-animation lane rather than

three lanes at once; the unspecified half is the next board below.

IT RATHER THAN LEAVING IT INSIDE A BUILDABILITY CLAIM.** The landed bound is TWO divisions, whole →

halves → quarters [BE_0007_S3:97-99], and § 5 supplies a plane, an interlock chain, a ROM, a

weight account and a telegraph for the first one only. The second has no plane, no seam chain inside

a half, no ROM and no structural account of what a quarter is, which also means the quarters'

"only close" moveset [:100-102] currently has no structural cause the way the halves' reach loss

does. § 5 states the constraints any resolution must satisfy — each quarter a conformant body, the

moveset loss falling out of what the second plane removes, and the ±4-degree per-segment excursion

preserved under scale-the-clock-never-the-angles — and names the recursive read as weighed and not

taken, because four interlocked bodies would break the whole's six-arm count and its landed resting

silhouette. OWNING LANE: whichever pass next touches the encounter's articulation; this creature

sheet does not close it from here.

Kandha Puranam return ZERO rows in T0_Source_Material_Registry, and the region page ranks the gap

itself [south_india.md:1280-1281]. Every structural claim in § 4 and § 5 is superseded in place

when it lands. The landed sheet already noted the pattern — three blocking acquisitions in two

chapters [:265-269] — and this is that finding's control run rather than a fourth instance.

cites CH_07.md L328-329 for the J10 bullet (live: L329-330), L333 for the honored-never-killed

bullet (live: L334), and south_india.md L1051-1052 for the Folklore Combat line (live:

L1078-1080) and L1253-1254 for the § 16 acquisition gate (live: L1280-1281). Systematic drift, and

the class is already boarded at DIRECTOR_RULING_L12 RULING_9.

10. the ten refusal tests, self-run

anything can satisfy a body that is two interlocked conformant skeletons sharing a fused limb pair.

with every joint, range, chain and conformance floor named; the one addition — the seam interlock —

names its segments, its parents, its single axis, its ±4-degree excursion, and how weight passes

through it (it does not: each half's legs carry its own side and the seam carries shear); and the

whole form's limb-pair count is stated rather than implied. THE GAP, declared here rather than

discovered by a critic: the SECOND division's structure is unspecified and is boarded in § 9, so

this test passes on one of the mechanic's two divisions and says so.

against the ascent's own hard daylight.

black fill by the seam's flicker going to zero along the body's full height, and ALL THREE breakable

parts are outline features — the fused forelimb pair at ~0.14 and the two planted leg anchors at

~0.08 each [BE_0007_S3:173-176]. PLATE I also judges the six-arm count's inboard carry, which is

the one thing this sheet added that a silhouette test could break.

different counter classes, and this sheet added no shape and re-aimed none.

the hill rather than on itself, and a division it earns rather than suffers.

two refusing a reduction, and one running toward more care and taking its finding (the peacock

constraint).

the HL_0002 fantasy licence is quoted as the ground for the designed body; positive control

(SRC_00262, the shelf keystone) and negative control (zero Clothey / Zvelebil / Kandha rows in

T0_Source_Material_Registry) both run and both reported.

string; the Sealed Letter untouched; reveal_state HELD throughout the chapter.

matching CH_07.md:271's exclusive list, image_to_3d verified as the concept-plate route, and the

photographic-template failure mode named for both the wildlife and human variants.

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