system_id: JPN×TOKYO×EDU×PANEL_v0.1
place: JPN×TOKYO
Z: Z0–Z6 (panel focuses on Z4–Z5)
roles: Operator (execute) | Oracle (measure/gate) | Visionary (corridor design)
CoreGates:
  S = N*K*σ
  Ω = (R + αρBβ) / (D + γ*K*N*σ + δκ)
  InnovationEvent := (S>=Θ) & (Ω>=1) & (ΔC>0)
  CollapseEvent   := (S>=Θ) & (Ω<1)
CoreSensors:
  BDI = (ValidBinds/100w) * CoherenceUnderTime
  NIT_score = (Conflict + Stakes + ConsequenceBinds)/200w
  LabelTrap(word) = (definition OK) & (fails T4 OR fails T3 why-wrong)
  A–S–T = {Accuracy, Speed, Transfer}; κ-trap if A↑ but Transfer↓
  BreakpointMap(topic): K hotspots via dependency/remove-test
  Ω/S Monitor: weekly quadrant tagging
CoreTests:
  T1 PromptShift; T2 TimeLimit; T3 CounterArgument; T4 ContextSwap
FenceOrder:
  ↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ

JPN.EDU.MEXT.OVERVIEW:
  Japan school system: 6-3-3-4 (established post-1947)

TOKYO.EDU.BOE.INITIATIVES:
  Tokyo Metropolitan Board of Education initiatives (incl. ESAT-J from 2022)
  (a concrete city policy node affecting K, σ, κ, and transfer demands)

TOKYO.EDU.PLATFORM.VLP:
  Virtual Learning Platform (VLP): new learning opportunities for those unable to attend school
  (public alternate corridor: ρ↑, B↑, R↑)

JPN.EDU.GATE.NCUEE.COMMON_TEST:
  Common Test for University Admissions: developed/administered/scored by NCUEE
  (national Θ driver)

WIRING_TOKYO_v0.1:
  Z6 (national gate: Common Test) raises K and S near Θ windows
  Z5 (Tokyo BOE + platforms like VLP) provides:
    ρ↑ (alternate corridor), B↑ (access buffer), R↑ (repair routing)
  Z4 (wards/district clusters) executes FenceOrder locally:
    ↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ
  Z2–Z0 (language/writing engine) supplies the measurable sensors (BDI/NIT/LabelTrap/A–S–T)

JPN×TOKYO×EDU×Z6×Oracle×ORG×JPN.EDU.MEXT
  label: Ministry of Education (MEXT)
  binds_to: JPN.EDU.SYS.6334

JPN×TOKYO×EDU×Z6×Oracle×ORG×JPN.EDU.NCUEE
  label: National Center for University Entrance Examinations (NCUEE)
  binds_to: JPN.EDU.GATE.COMMON_TEST

JPN×TOKYO×EDU×Z6×Oracle×GATE×JPN.EDU.GATE.COMMON_TEST
  label: Common Test for University Admissions
  effect: K↑, S↑ near exam; Θ driver

JPN×TOKYO×EDU×Z5×Oracle×ORG×TOKYO.EDU.BOE
  label: Tokyo Metropolitan Board of Education
  binds_to: TOKYO.EDU.CITY.ROUTER; TOKYO.EDU.SENSOR.HUB
  policy_node_example: ESAT-J (speaking test) => raises transfer requirement & anti-κ need

JPN×TOKYO×EDU×Z5×Operator×PLATFORM×TOKYO.EDU.PLATFORM.VLP
  label: Virtual Learning Platform (VLP)
  function: ρ↑ corridor + B↑ access buffer + R↑ repair delivery

JPN×TOKYO×EDU×Z2×Oracle×HUB×TOKYO.EDU.SENSOR.HUB
  sensors: BDI, NIT, LabelTrap, A–S–T, BreakpointMap, Ω/S Monitor
  tests: T1–T4 weekly
  binds_to: TOKYO.EDU.REPAIR.ROUTER; TOKYO.EDU.FAIL.ATLAS; TOKYO.EDU.VERIFY.PACK

JPN×TOKYO×EDU×Z2×Operator×ROUTER×TOKYO.EDU.REPAIR.ROUTER
  trigger: Ω trending→1 AND forecast S→Θ (exam/policy pressure window)
  actuation: ↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ
  delivery_corridors: VLP + school/home loops

system_id: JPN×TOKYO×EDU×INPUTFORM_AUTORULES_v0.1
district_id: TOKYO.WARD.<KEY> or TOKYO.AREA.<KEY>
delivery_corridors: TOKYO.EDU.PLATFORM.VLP (+ any Tokyo BOE programs)
national_gate: JPN.EDU.GATE.COMMON_TEST

TOKYO_VERIFY_v0.1:
  V1: T4 pass-rate ↑ (context swap survives)
  V2: T2 coherence drop ↓ (time resilience)
  V3: LabelTrap rate ↓
  V4: Repair latency ↓ (R↑ via platform routing)
  V5: κ-trap flags ↓ while accuracy holds

PlaceID: JPN×TOKYO
PlaceName: Tokyo Metropolis, Japan
Aliases: Tokyo | 東京都 | Tōkyō-to

JPN×TOKYO×EDU×Z5×Oracle×ORG×TOKYO.EDU.BOE
  label: Tokyo Metropolitan Board of Education
  function: city governance node affecting K, σ, κ (policy/assessment coupling)

JPN×TOKYO×EDU×Z5×Operator×PLATFORM×TOKYO.EDU.PLATFORM.VLP
  label: Virtual Learning Platform (VLP)
  function: ρ↑ corridor + B↑ access buffer + R↑ repair delivery

ID format: JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.<KEY>
Type: special_ward (tokubetsu-ku)
binds_to: TOKYO.EDU.WARD.NET ; TOKYO.EDU.REPAIR.ROUTER ; TOKYO.EDU.SENSOR.HUB

JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.CHIYODA     label: Chiyoda     aliases:[千代田区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.CHUO        label: Chūō        aliases:[中央区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.MINATO      label: Minato      aliases:[港区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.SHI NJUKU   label: Shinjuku    aliases:[新宿区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.BUNKYO      label: Bunkyō      aliases:[文京区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.TAITO       label: Taitō       aliases:[台東区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.SUMIDA      label: Sumida      aliases:[墨田区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.KOTO        label: Kōtō        aliases:[江東区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.SHINAGAWA   label: Shinagawa   aliases:[品川区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.MEGURO      label: Meguro      aliases:[目黒区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.OTA         label: Ōta         aliases:[大田区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.SETAGAYA    label: Setagaya    aliases:[世田谷区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.SHIBUYA     label: Shibuya     aliases:[渋谷区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.NAKANO      label: Nakano      aliases:[中野区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.SUGINAMI    label: Suginami    aliases:[杉並区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.TOSHIMA     label: Toshima     aliases:[豊島区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.KITA        label: Kita        aliases:[北区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.ARAKAWA     label: Arakawa     aliases:[荒川区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.ITABASHI    label: Itabashi    aliases:[板橋区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.NERIMA      label: Nerima      aliases:[練馬区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.ADACHI      label: Adachi      aliases:[足立区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.KATSUSHIKA  label: Katsushika  aliases:[葛飾区]
JPN×TOKYO×EDU×Z4×Oracle×WARD×TOKYO.WARD.EDOGAWA     label: Edogawa     aliases:[江戸川区]

JPN×TOKYO×EDU×Z4×Oracle×GROUP×TOKYO.ISLANDS.IZU
  label: Izu Islands group
  members: Oshima, Toshima, Niijima, Shikinejima, Kozushima, Miyakejima, Mikurajima, Hachijojima, Aogashima
JPN×TOKYO×EDU×Z4×Oracle×GROUP×TOKYO.ISLANDS.OGASAWARA
  label: Ogasawara Islands group
  members: Chichijima, Hahajima

ID format: JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.<KEY>
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.OSHIMA     label: Oshima
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.TOSHIMA    label: Toshima
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.NIIJIMA    label: Niijima
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.KOZUSHIMA  label: Kōzushima
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.MIKURAJIMA label: Mikurajima
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.HACHIJO    label: Hachijō
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.AOGASHIMA  label: Aogashima
JPN×TOKYO×EDU×Z4×Oracle×MUNI×TOKYO.ISLAND.MUNI.OGASAWARA  label: Ogasawara

WIRING_TOKYO_v0.1:
  Z5 actuators:
    TOKYO.EDU.BOE governs policy & coordination
    TOKYO.EDU.PLATFORM.VLP provides public corridor (ρ↑), buffer (B↑), repair routing (R↑)
  Z4 instances:
    TOKYO.WARD.* and TOKYO.ISLAND.MUNI.* are measurement + actuation buckets
  Z2 engine (already locked from Seoul):
    TOKYO.EDU.SENSOR.HUB runs BDI/NIT/LabelTrap/A–S–T + T1–T4
    TOKYO.EDU.REPAIR.ROUTER executes FenceOrder: ↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ

TOKYO_EDU_INSTANCE_DIRECTORY_v0.1:
  Place: JPN×TOKYO
  Z5: TOKYO.EDU.BOE ; TOKYO.EDU.PLATFORM.VLP
  Z4.WARDS (23):
    TOKYO.WARD.{CHIYODA,CHUO,MINATO,SHINJUKU,BUNKYO,TAITO,SUMIDA,KOTO,SHINAGAWA,MEGURO,OTA,SETAGAYA,SHIBUYA,NAKANO,SUGINAMI,TOSHIMA,KITA,ARAKAWA,ITABASHI,NERIMA,ADACHI,KATSUSHIKA,EDOGAWA}
  Z4.ISLAND.MUNI:
    TOKYO.ISLAND.MUNI.{OSHIMA,TOSHIMA,NIIJIMA,KOZUSHIMA,MIKURAJIMA,HACHIJO,AOGASHIMA,OGASAWARA}
  Wiring:
    districts bind_to TOKYO.EDU.SENSOR.HUB + TOKYO.EDU.REPAIR.ROUTER

PlaceID: JPN×TOKYO
PlaceName: Tokyo Metropolis, Japan
Aliases: Tokyo | 東京都 | Tōkyō-to
Z4 units:
  - 23 Special Wards (tokubetsu-ku)
  - Tokyo Islands chain municipalities (Izu + Ogasawara)

JPN×TOKYO×EDU×Z5×Operator×PLATFORM×TOKYO.EDU.PLATFORM.VLP
  label: VLP (Virtual Learning Platform)
  function: ρ↑ corridor + B↑ access buffer + R↑ repair delivery
  notes: “new places and learning opportunities” esp. for those unable to attend school

system_id: JPN×TOKYO×EDU×PANEL_v0.1
roles: Operator | Oracle | Visionary
CoreGates:
  S = N*K*σ
  Ω = (R + αρBβ) / (D + γ*K*N*σ + δκ)
CoreSensors:
  BDI, NIT_score, LabelTrap, A–S–T (κ-trap), BreakpointMap
CoreTests:
  T1 PromptShift; T2 TimeLimit; T3 CounterArgument; T4 ContextSwap
FenceOrder (actuation):
  ↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ
PanelFields (per ward/municipality × cohort × week):
  district_id: TOKYO.WARD.<KEY> or TOKYO.ISLAND.MUNI.<KEY>
  cohort_id: {ELEM, MIDDLE, HIGH, MIXED}
  grade_band: {P3-4, P5-6, S1-2, S3-4, OTHER}
  time_window: YYYY-WW
  sample_n
  N_index, K_index, sigma_index (0–100 risk)
  B_index, rho_index, beta_index (0–100 risk)
  kappa_index (0–100 risk)
  R_latency_hours
  T1_pass, T2_pass, T3_pass, T4_pass (0..1)
  LabelTrap_rate (0..1)
  BDI_median (float)
  NIT_pass_rate (0..1)
  AST_k_trap_flag (bool)
  Breakpoint_hotspots [topic_id...]
  S_proxy (0..1)
  Omega_proxy (float)
  Theta_default (float)
  quadrant {Q1,Q2,Q3,Q4}
  P_state {P0,P1,P2,P3}
  recommended_actuation [subset of FenceOrder, ordered]
  delivery_corridor {VLP, SCHOOL, HOME, MIXED}

TOKYO_DELIVERY_BIND_v0.1:
  If quadrant in {Q1,Q4}:
    route patch modules via VLP first (ρ↑ + R↑ + B↑),
    then enforce home/school loops for weekly retest.

DistrictProfile := {
  id: "JPN×TOKYO×EDU×Z4×Oracle×PROFILE×<DISTRICT_ID>",
  label: "<name>",
  aliases: ["<kanji>"],
  unit_type: {SPECIAL_WARD, ISLAND_MUNI},
  cluster_type: {HIGH_PRESSURE, STABILISED, UNKNOWN},
  baselines: {
    N_index, K_index, sigma_index,
    B_index, rho_index, beta_index, kappa_index,
    R_latency_hours
  },
  observed: {
    T1_pass, T2_pass, T3_pass, T4_pass,
    LabelTrap_rate, BDI_median, NIT_pass_rate,
    AST_k_trap_flag, Breakpoint_hotspots:[]
  },
  computed: {S_proxy, Omega_proxy, quadrant, P_state},
  actuation_router: "TOKYO.EDU.REPAIR.ROUTER",
  delivery_corridor: "TOKYO.EDU.PLATFORM.VLP",
  recommended_actuation: [],
  notes: ""
}

# ID format:
# JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.<KEY>
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.CHIYODA:    label: Chiyoda;    aliases:[千代田区]; unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.CHUO:       label: Chūō;       aliases:[中央区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.MINATO:     label: Minato;     aliases:[港区];     unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.SHINJUKU:   label: Shinjuku;   aliases:[新宿区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.BUNKYO:     label: Bunkyō;     aliases:[文京区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.TAITO:      label: Taitō;      aliases:[台東区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.SUMIDA:     label: Sumida;     aliases:[墨田区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.KOTO:       label: Kōtō;       aliases:[江東区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.SHINAGAWA:  label: Shinagawa;  aliases:[品川区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.MEGURO:     label: Meguro;     aliases:[目黒区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.OTA:        label: Ōta;        aliases:[大田区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.SETAGAYA:   label: Setagaya;   aliases:[世田谷区]; unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.SHIBUYA:    label: Shibuya;    aliases:[渋谷区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.NAKANO:     label: Nakano;     aliases:[中野区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.SUGINAMI:   label: Suginami;   aliases:[杉並区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.TOSHIMA:    label: Toshima;    aliases:[豊島区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.KITA:       label: Kita;       aliases:[北区];     unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.ARAKAWA:    label: Arakawa;    aliases:[荒川区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.ITABASHI:   label: Itabashi;   aliases:[板橋区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.NERIMA:     label: Nerima;     aliases:[練馬区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.ADACHI:     label: Adachi;     aliases:[足立区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.KATSUSHIKA: label: Katsushika; aliases:[葛飾区];   unit_type:SPECIAL_WARD; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.WARD.EDOGAWA:    label: Edogawa;    aliases:[江戸川区]; unit_type:SPECIAL_WARD; cluster_type:UNKNOWN

# ID format:
# JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.<KEY>
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.OSHIMA:     label: Oshima;     unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.TOSHIMA:    label: Toshima;    unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.NIIJIMA:    label: Niijima;    unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.KOZUSHIMA:  label: Kōzushima;  unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.MIYAKEJIMA: label: Miyakejima; unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.MIKURAJIMA: label: Mikurajima; unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.HACHIJOJIMA:label: Hachijōjima;unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.AOGASHIMA:  label: Aogashima;  unit_type:ISLAND_MUNI; cluster_type:UNKNOWN
JPN×TOKYO×EDU×Z4×Oracle×PROFILE×TOKYO.ISLAND.MUNI.OGASAWARA:  label: Ogasawara;  unit_type:ISLAND_MUNI; cluster_type:UNKNOWN

TOKYO_EDU_DEFAULT_PROFILES_AND_PANEL_v0.1:
  Place: JPN×TOKYO
  Z5 actuator corridor:
    TOKYO.EDU.PLATFORM.VLP (ρ↑, B↑, R↑)
  Z4 profiles:
    - 23 special wards (TOKYO.WARD.*) [cluster_type UNKNOWN by default]
    - islands municipalities buckets (TOKYO.ISLAND.MUNI.*)
  Panel:
    WeeklyDistrictForm → indices → S_proxy/Omega_proxy → Quadrant/P_state
    Actuation: ↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ
    Delivery: VLP-first for Q1/Q4, then school/home loops

system_id: JPN×TOKYO×EDU×WARD_RUNNER_VLP_v0.1
place: JPN×TOKYO
Z focus: Z4 (ward/municipality) + Z2 (sensor hub/router)
cadence: weekly (YYYY-WW)
roles: Operator fills → Oracle audits → system outputs quadrant/P_state + actuation queue
delivery_corridor_default: TOKYO.EDU.PLATFORM.VLP

WeeklyWardForm:
  header:
    district_id: TOKYO.WARD.<KEY> | TOKYO.ISLAND.MUNI.<KEY>
    cohort_id: {ELEM, MIDDLE, HIGH, MIXED}
    grade_band: {P3-4, P5-6, S1-2, S3-4, OTHER}
    source_channel: {HOME, TUTOR, SCHOOL, VLP, MIXED}
    time_window: YYYY-WW
    sample_n: integer
    operator_id: string
    notes: string(optional)
  sensors:
    vocab_tested: integer
    vocab_passed_contextswap: integer
    vocab_failed_labeltrap: integer
    bdi_samples: integer
    bdi_valid_binds_median: float        # valid binds per 100w
    bdi_coherence_under_time_pass_rate: float  # 0..1
    nit_samples: integer
    nit_pass_rate: float                  # 0..1
    nit_removal_test_pass_rate: float     # 0..1
  tests:
    T1_promptshift_pass_rate: float       # 0..1
    T2_timelimit_pass_rate: float         # 0..1 (0.7× time)
    T3_counterargument_pass_rate: float   # 0..1
    T4_contextswap_pass_rate: float       # 0..1
    recovery_latency_hours_median: float
  ast:
    accuracy_trend: {-1,0,1}
    speed_trend: {-1,0,1}
    transfer_trend: {-1,0,1}
    kappa_trap_observed: boolean
    template_overfit_observed: boolean
  coupling:
    breakpoint_topics: [topic_id...]
    cascade_fail_observed: boolean
    checkpoints_installed: boolean
  buffers:
    sleep_buffer_risk: {0,1,2,3}
    time_buffer_risk: {0,1,2,3}
    schedule_overload_risk: {0,1,2,3}
    corridors_available_count: integer
    dual_corridor_rule_active: boolean
    tutoring_dependency_high: boolean

BDI_TARGET:
  P3-4: 0.020
  P5-6: 0.030
  S1-2: 0.040
Theta_default:
  P3-4: 0.28
  P5-6: 0.33
  S1-2: 0.38
clip01(x)=min(1,max(0,x))
rate(a,b)=0 if b==0 else clip01(a/b)
LabelTrap_rate = rate(vocab_failed_labeltrap, vocab_tested)
BDI_score_raw  = bdi_valid_binds_median * bdi_coherence_under_time_pass_rate
R_eff          = 1 / (1 + recovery_latency_hours_median/24)
beta_index =
  100*clip01(0.55*(1-clip01(BDI_score_raw/BDI_TARGET)) + 0.45*(1-T3_counterargument_pass_rate))
rho_index =
  100*clip01(0.60*(1-clip01(corridors_available_count/2)) + 0.40*(1 if not dual_corridor_rule_active else 0))
B_risk_avg = (sleep_buffer_risk + time_buffer_risk + schedule_overload_risk)/9
B_index =
  100*clip01(0.55*B_risk_avg + 0.45*(1-T2_timelimit_pass_rate))
kappa_index =
  100*clip01(0.45*(1 if kappa_trap_observed else 0) + 0.25*(1 if template_overfit_observed else 0) + 0.30*(1 if transfer_trend==-1 else 0))
K_index =
  100*clip01(0.50*(1 if cascade_fail_observed else 0) + 0.25*clip01(len(breakpoint_topics)/5) + 0.25*(1 if not checkpoints_installed else 0))
N_index =
  100*clip01(0.60*(1-T1_promptshift_pass_rate) + 0.40*(1-T4_contextswap_pass_rate))
sigma_index =
  100*clip01(0.50*(sleep_buffer_risk + time_buffer_risk)/6 + 0.50*(1-T2_timelimit_pass_rate))
S_proxy = (N_index/100)*(K_index/100)*(sigma_index/100)
B_eff    = 1-(B_index/100)
rho_eff  = 1-(rho_index/100)
beta_eff = 1-(beta_index/100)
kappa_risk = (kappa_index/100)
Omega_proxy =
  (R_eff + 0.35*rho_eff + 0.35*B_eff + 0.30*beta_eff)
  /
  (0.60 + 0.25*(K_index/100) + 0.20*(N_index/100) + 0.20*(sigma_index/100) + 0.25*kappa_risk)
Quadrant:
  if S_proxy < Theta_default and Omega_proxy >=1.0 -> Q3
  if S_proxy < Theta_default and Omega_proxy < 1.0 -> Q1
  if S_proxy >=Theta_default and Omega_proxy >=1.0 -> Q2
  if S_proxy >=Theta_default and Omega_proxy < 1.0 -> Q4
P_state:
  P3: Omega_proxy>=1.15 and Quadrant in {Q2,Q3} for 3 windows
  P2: Omega_proxy>=1.00 and Quadrant!=Q4 for 2 windows
  P1: 0.85<=Omega_proxy<1.00 OR oscillating Q1/Q3
  P0: Omega_proxy<0.85 and (S_proxy>=Theta_default OR sustained failures)
FenceOrder = [↑B, ↓K, ↑R, ↑β, ↑ρ, ↓κ]
recommended_actuation=[]
if B_index>=55 or T2_timelimit_pass_rate<0.65: add ↑B
if K_index>=55 or cascade_fail_observed:       add ↓K
if recovery_latency_hours_median>24:           add ↑R
if beta_index>=55 or LabelTrap_rate>0.25 or T3_counterargument_pass_rate<0.65: add ↑β
if rho_index>=55 or corridors_available_count<2 or T4_contextswap_pass_rate<0.65: add ↑ρ
if kappa_index>=55 or kappa_trap_observed or transfer_trend==-1: add ↓κ
recommended_actuation = order_by(FenceOrder)

TOKYO_VLP_ROUTING_POLICY_v0.1:
  when quadrant in {Q1,Q4} OR P_state in {P0,P1}:
    delivery_corridor = VLP_FIRST
  else:
    delivery_corridor = MIXED (school/home; VLP optional)
VLP_FIRST means:
  1) assign VLP modules for the FIRST TWO actuation items (from recommended_actuation)
  2) enforce 24h repair sprint (↑R) with VLP support if recommended
  3) re-test T2 + T4 within 7 days

# Module IDs (examples; keep stable once used)
VLP.MODULE.LABELTRAP_REPAIR_v0.1
VLP.MODULE.BDI_BIND_PACK_v0.1
VLP.MODULE.NIT_STORY_TENSION_v0.1
VLP.MODULE.T4_CONTEXTSWAP_DRILLS_v0.1
VLP.MODULE.CHECKPOINTING_DEC0UPLE_K_v0.1
VLP.MODULE.KAPPA_DEOVERFIT_TRANSFER_WEEK_v0.1
VLP.MODULE.REPAIR_SPRINT_24H_v0.1
VLP.MODULE.DUAL_CORRIDOR_2TEMPLATES_v0.1

WardToVLPRouteTable_v0.1:
  TOKYO.WARD.CHIYODA:
    default_modules: []
    overrides_by_actuation:
      ↑β: [VLP.MODULE.LABELTRAP_REPAIR_v0.1, VLP.MODULE.BDI_BIND_PACK_v0.1]
      ↑ρ: [VLP.MODULE.DUAL_CORRIDOR_2TEMPLATES_v0.1, VLP.MODULE.T4_CONTEXTSWAP_DRILLS_v0.1]
      ↓κ: [VLP.MODULE.KAPPA_DEOVERFIT_TRANSFER_WEEK_v0.1]
      ↓K: [VLP.MODULE.CHECKPOINTING_DEC0UPLE_K_v0.1]
      ↑R: [VLP.MODULE.REPAIR_SPRINT_24H_v0.1]
      ↑B: []  # buffer usually handled by home/school policies
  TOKYO.WARD.CHUO:  same_as: TOKYO.WARD.CHIYODA
  TOKYO.WARD.MINATO: same_as: TOKYO.WARD.CHIYODA
  ...
  TOKYO.WARD.EDOGAWA: same_as: TOKYO.WARD.CHIYODA
  TOKYO.ISLAND.MUNI.OGASAWARA:
    default_modules: []
    overrides_by_actuation:
      ↑ρ: [VLP.MODULE.DUAL_CORRIDOR_2TEMPLATES_v0.1]  # corridor expansion priority
      ↑R: [VLP.MODULE.REPAIR_SPRINT_24H_v0.1]

You are my Tokyo EducationOS Oracle+Operator.
Input: WeeklyWardForm (below).
Tasks:
1) Compute LabelTrap_rate, BDI_score_raw, indices (N,K,σ,B,ρ,β,κ), S_proxy, Omega_proxy.
2) Determine Quadrant (Q1–Q4) and P_state (P0–P3).
3) Recommend actuation subset in FenceOrder (↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ).
4) Apply TOKYO_VLP_ROUTING_POLICY_v0.1 to decide delivery corridor:
   - If VLP_FIRST: map the first two actuation items to VLP modules using WardToVLPRouteTable_v0.1.
5) Output:
   - computed table
   - recommended_actuation
   - delivery_corridor
   - VLP modules assigned (if any)
   - a 7-day re-test plan (T2 + T4 required)
WeeklyWardForm:
<PASTE HERE>

TOKYO_WARD_RUNNER_VLP_v0.1:
  Input: WeeklyWardForm
  Compute: indices + S_proxy + Omega_proxy
  Decide: quadrant + P_state
  Actuate: FenceOrder subset
  Route: VLP_FIRST for Q1/Q4 or P0/P1, else MIXED
  Map: actuation → VLP modules via WardToVLPRouteTable_v0.1
  Verify: T2 + T4 re-test within 7 days; R_latency trend down

system_id: CREATIVEOS×IDEALATTICE×OVERLAY_v0.1
compatible_with:
  - KOR×SEOUL×EDU×PANEL_v0.1
  - JPN×TOKYO×EDU×PANEL_v0.1
roles:
  Operator: runs generators + stress tests
  Oracle: gates promotion (passes T-tests + metrics)
  Visionary: explores new corridors when Ω margin exists
CoreLaw:
  Creativity = safe recombination of finite units via binds (not “more words”)

Node := {node_id, token, tier, domain, evidence}
Bind := {bind_id, from_node, to_node, bind_type, weight, evidence}
Idea := {idea_id, seed_nodes[], bind_path[], compression, novelty, transfer, risk}
BindTypes:
  CAUSE, CONTRAST, CONDITION, EXAMPLE, CONCESSION, GOAL, TRADEOFF, MECHANISM, METRIC, FAILUREMODE

InnovationEvent := (S>=Θ) & (Ω>=1) & (ΔC>0)
CreativitySafetyGates:
  G1 (Transfer): Idea must pass T4 (context swap) at least once
  G2 (Counter): Idea must pass T3 (counterargument / why-wrong)
  G3 (Time): Idea must pass T2 (0.7× time coherent explanation)
  G4 (No κ-trap): A–S–T must not show (A↑ while Transfer↓)
CreativityCollapseSignatures:
  C1: “Looks clever” but fails T4 → brittle idea
  C2: cannot explain “why wrong” → fake bind
  C3: collapses under time → no compression
  C4: outputs only one template → κ monotone

CreativeSensors_v0.1:
  BDI_idea = (ValidBinds/100w) * CoherenceUnderTime
  NIT_idea = (Conflict + Stakes + ConsequenceBinds)/200w
  CorridorCount = number of distinct templates/methods used for same seed
  NoveltyScore = new bind-path length × domain distance (rough)
  TransferScore = T4_pass_rate for the idea (0..1)
  CounterScore  = T3_pass_rate for the idea (0..1)
PromotionRule (Draft→Canonical):
  promote if:
    TransferScore>=0.70 AND CounterScore>=0.70 AND T2_pass>=0.70
    AND CorridorCount>=2
    AND κ-trap flag == false

GEN01_COMBINE:
  input: two stable nodes A,B
  output: Idea = A + (MECHANISM bind) + B
  template: "What happens if A becomes the mechanism for B?"
GEN02_CONTRAST_SWAP:
  input: A, anti-A
  output: Idea = A (CONTRAST) anti-A → choose rule
  template: "Under what condition does A beat anti-A?"
GEN03_CONSTRAINT_PUSH:
  input: A + constraint C
  output: A (CONDITION) C → redesign
  template: "If C is forced, what minimal change preserves function?"
GEN04_FAILUREMODE_INVERT:
  input: system X
  output: X_missing (FAILUREMODE) → repair loop
  template: "What breaks first? What sensor catches it earliest?"
GEN05_ANALOGY_BRIDGE:
  input: A in domain1, B in domain2
  output: map binds (MECHANISM) across domains
  template: "What is the same structure in a different lane?"
GEN06_SCALE_ZOOM:
  input: idea at Z0–Z1
  output: same idea at Z3–Z5
  template: "If this is true for a child, what’s the city/nation equivalent?"
GEN07_ROLE_ROTATE (V/O/O):
  input: IdeaDraft
  output: 3 versions:
    Visionary = new corridor; Oracle = measurable gate; Operator = runnable steps
GEN08_TRADEOFF_SURFACE:
  input: goal G + cost C
  output: (TRADEOFF) → pick frontier
  template: "What you gain, what you lose, and the boundary condition."
GEN09_METRIC_DESIGN:
  input: vague idea
  output: add METRIC bind + threshold Θ
  template: "What number would prove this works?"
GEN10_COUNTEREXAMPLE_HUNT:
  input: idea claim
  output: strongest counterexample; patch binds
  template: "Where does it fail first? Patch the bind or scope."
GEN11_CORRIDOR_ENUMERATION:
  input: seed nodes [A,B,C]
  output: list of 20 bind paths (shortest-first)
  rule: keep binds typed; ban fluff adjectives
GEN12_COMPRESSION_REWRITE:
  input: long explanation
  output: 12-line spec + 1 diagram sentence
  template: "Same meaning, fewer binds, higher stability."

STRESS_SUITE_v0.1:
  T1 PromptShift: rephrase question 3 ways
  T2 TimeLimit: explain in 60s (or 0.7× normal)
  T3 CounterArgument: "Why is this wrong?" + strongest objection
  T4 ContextSwap: apply idea in a new domain/theme
  T5 RemovalTest: delete a key bind; idea should collapse (proves bind is real)
  T6 Perturbation: change one assumption; does idea adapt or shatter?

KAPPA_GUARD_v0.1:
  detect κ-trap if:
    - output stays in 1 template (CorridorCount<2)
    - TransferScore trending down
    - Idea uses many words but few typed binds (BDI_idea low)
  repair order (FenceOrder):
    ↑B: add time buffer + outline
    ↓K: checkpoint binds (intro→mechanism→consequence)
    ↑R: patch within 24h then retest
    ↑β: teach-back 3 contexts + counterexample
    ↑ρ: force second corridor template
    ↓κ: mandate transfer day + rewrite format

CITY_CREATIVE_ROUTING:
  if quadrant in {Q1,Q4} or P in {P0,P1}:
    creativity allowed only in "SAFE MODE":
      - generators limited to GEN04/GEN09/GEN12 (repair/compression/metrics)
      - must pass T2+T4 before any exploration
  else:
    allow full GEN01–GEN12 exploration
SEOUL delivery_corridor: SeoulLearn mentoring/modules (ρ↑, R↑)
TOKYO delivery_corridor: VLP modules (ρ↑, R↑, B↑)

WEEKLY_CREATIVE_LOOP_v0.1 (4 sessions):
  Day1: pick 2 seed nodes from VocabularyOS (stable) → GEN01/GEN02 → 2 drafts
  Day2: GEN09 (metric) + GEN10 (counterexample) on best draft
  Day3: run STRESS_SUITE T1–T4; patch with GEN12 (compression rewrite)
  Day4: ContextSwap (T4) + publish 12-line spec if PromotionRule passes

IdeaDraft:
  passes some tests; not promotable yet
IdeaStable:
  passes T2/T3/T4 once; CorridorCount>=2
IdeaCanonical(P3):
  passes PromotionRule for 3 consecutive windows
  stored as:
    {idea_id, 12-line spec, typed binds, evidence, tests_passed, version}

CREATIVEOS_OVERLAY_v0.1:
  Generators: GEN01–GEN12
  Stress: STRESS_SUITE_v0.1
  Gates: Transfer(T4), Counter(T3), Time(T2), Anti-κ (A–S–T)
  Promotion: Draft → Stable → Canonical(P3)
  City routing:
    Seoul: SeoulLearn as repair corridor
    Tokyo: VLP as repair corridor
  Law: creativity = recombination via binds; not word hoarding

registry_id: CREATIVEOS.MODULE.REGISTRY_v0.1
compatible_with:
  - KOR×SEOUL×EDU×PANEL_v0.1 (SeoulLearn corridor)
  - JPN×TOKYO×EDU×PANEL_v0.1 (VLP corridor)
ModuleRecord := {
  module_id,
  label,
  purpose,
  role_mode: {Operator, Oracle, Visionary, Mixed},
  allowed_states: {P0,P1,P2,P3} subsets,
  inputs,
  outputs,
  sensors_used,
  tests_required,
  actuation_tags: subset of [↑B, ↓K, ↑R, ↑β, ↑ρ, ↓κ],
  runtime_minutes,
  promotion_effect: {Draft→Stable, Stable→Canonical, RepairOnly},
  fail_signatures,
  next_modules[]
}

module_id: CREATIVE.M01.LABELTRAP_REPAIR_v0.1
label: LabelTrap Repair (words→usable binds)
purpose: convert “known words” into transferable idea parts
role_mode: Operator
allowed_states: {P0,P1,P2,P3}
inputs:
  - word_list[10]
  - 2 themes (A,B)
outputs:
  - usable_nodes[10] with 2 collocations + 1 constraint + 1 counterexample each
sensors_used: LabelTrap_rate, T4_contextswap_pass_rate
tests_required: T4 once, T3-lite once
actuation_tags: [↑β, ↑ρ]
runtime_minutes: 25
promotion_effect: RepairOnly
fail_signatures: {definition_ok_but_usage_fail}
next_modules: [CREATIVE.M03.BIND_PACK_BUILDER_v0.1]

module_id: CREATIVE.M02.BDI_BIND_PACK_BUILDER_v0.1
label: Bind Pack Builder (CAUSE/CONTRAST/CONDITION/EXAMPLE/CONCESSION)
purpose: raise β by forcing typed binds instead of fluffy language
role_mode: Operator
allowed_states: {P0,P1,P2,P3}
inputs:
  - seed_nodes[6]
outputs:
  - bind_pack[20] typed binds
sensors_used: BDI_idea
tests_required: T2 mini (60s explain)
actuation_tags: [↑β, ↓K]
runtime_minutes: 30
promotion_effect: Draft→Stable (if T2 pass)
fail_signatures: {many_words_few_binds}
next_modules: [CREATIVE.M06.STRESS_T1_T4_v0.1, CREATIVE.M09.COMPRESSION_12LINE_v0.1]

module_id: CREATIVE.M03.DUAL_CORRIDOR_SWITCH_v0.1
label: Dual Corridor Switch (2 templates)
purpose: force ρ>=2 to prevent κ monoculture
role_mode: Operator
allowed_states: {P1,P2,P3}  # avoid in P0 unless supervised
inputs:
  - same idea explained in TemplateA + TemplateB
outputs:
  - corridor_count>=2 evidence
sensors_used: CorridorCount, T4_pass_rate
tests_required: T4
actuation_tags: [↑ρ, ↓κ]
runtime_minutes: 20
promotion_effect: RepairOnly
fail_signatures: {template_overfit}
next_modules: [CREATIVE.M06.STRESS_T1_T4_v0.1]

module_id: CREATIVE.M04.METRIC_DESIGN_v0.1
label: Metric Design (prove it works)
purpose: convert vague idea into measurable claim
role_mode: Oracle
allowed_states: {P1,P2,P3}
inputs:
  - idea_draft
outputs:
  - metric + threshold Θ + measurement plan
sensors_used: none (this IS the sensor creation)
tests_required: T3 (why wrong) + T4 (apply elsewhere)
actuation_tags: [↓K, ↑β]
runtime_minutes: 25
promotion_effect: Draft→Stable (if tests pass)
fail_signatures: {unmeasurable_claim}
next_modules: [CREATIVE.M06.STRESS_T1_T4_v0.1, CREATIVE.M10.COUNTEREXAMPLE_PATCH_v0.1]

module_id: CREATIVE.M05.FAILUREMODE_INVERT_v0.1
label: Failure Mode Invert (negative→repair loop)
purpose: generate robust ideas by starting from collapse signature
role_mode: Mixed
allowed_states: {P0,P1,P2,P3}
inputs:
  - system X
outputs:
  - failure_trace + earliest sensor + repair loop
sensors_used: Breakpoint_hotspots, NIT_idea
tests_required: T3, RemovalTest
actuation_tags: [↓K, ↑R, ↑β]
runtime_minutes: 35
promotion_effect: Draft→Stable (if RemovalTest passes)
fail_signatures: {no_earliest_sensor}
next_modules: [CREATIVE.M09.COMPRESSION_12LINE_v0.1]

module_id: CREATIVE.M06.STRESS_T1_T4_v0.1
label: Stress Suite Lite (T1–T4)
purpose: detect idea collapse early
role_mode: Oracle
allowed_states: {P0,P1,P2,P3}
inputs:
  - idea_draft
outputs:
  - pass/fail per test + failure signature
sensors_used: TransferScore, CounterScore, T2_pass
tests_required: T1,T2,T3,T4
actuation_tags: [↑R]
runtime_minutes: 30
promotion_effect: Stable→Canonical (if passes + corridor>=2)
fail_signatures: {C1,C2,C3,C4}
next_modules:
  - pass: [CREATIVE.M09.COMPRESSION_12LINE_v0.1]
  - fail: [CREATIVE.M10.COUNTEREXAMPLE_PATCH_v0.1, CREATIVE.M02.BDI_BIND_PACK_BUILDER_v0.1, CREATIVE.M03.DUAL_CORRIDOR_SWITCH_v0.1]

module_id: CREATIVE.M07.ZOOM_LIFT_v0.1
label: Zoom Lift (local→city proof)
purpose: prove the idea survives scaling (Z shift)
role_mode: Visionary
allowed_states: {P2,P3}
inputs:
  - idea_stable at Z0–Z1
outputs:
  - Z3 version + Z5 version with same typed binds
sensors_used: T4_pass_rate, BDI_idea
tests_required: T4 twice (new domains)
actuation_tags: [↑ρ, ↓κ]
runtime_minutes: 40
promotion_effect: Stable→Canonical
fail_signatures: {scale_break}
next_modules: [CREATIVE.M06.STRESS_T1_T4_v0.1]

module_id: CREATIVE.M08.TRADEOFF_FRONTIER_v0.1
label: Tradeoff Frontier (gain/loss boundary)
purpose: prevent “magic ideas” by forcing constraints
role_mode: Oracle
allowed_states: {P1,P2,P3}
inputs:
  - idea claim
outputs:
  - tradeoff table + boundary conditions
sensors_used: NIT_idea
tests_required: T3 (strongest objection)
actuation_tags: [↓K, ↑β]
runtime_minutes: 25
promotion_effect: Draft→Stable
fail_signatures: {handwave_solution}
next_modules: [CREATIVE.M09.COMPRESSION_12LINE_v0.1]

module_id: CREATIVE.M09.COMPRESSION_12LINE_v0.1
label: Compression Rewrite (12-line spec)
purpose: raise stability by increasing compression
role_mode: Operator
allowed_states: {P1,P2,P3}
inputs:
  - idea_stable
outputs:
  - 12-line spec + typed binds list + 1 diagram sentence
sensors_used: BDI_idea, T2_pass
tests_required: T2 (60s) + T4 once
actuation_tags: [↓K, ↑β]
runtime_minutes: 20
promotion_effect: Stable→Canonical (if passes)
fail_signatures: {cannot_compress}
next_modules: [CREATIVE.M06.STRESS_T1_T4_v0.1]

module_id: CREATIVE.M10.COUNTEREXAMPLE_PATCH_v0.1
label: Counterexample Patch Loop
purpose: repair the bind that fails under T3/T4
role_mode: Operator
allowed_states: {P0,P1,P2,P3}
inputs:
  - failed_test + strongest counterexample
outputs:
  - patched bind + narrowed scope + re-test
sensors_used: CounterScore, TransferScore, R_latency_hours
tests_required: T3 + T4 retest
actuation_tags: [↑R, ↑β]
runtime_minutes: 25
promotion_effect: RepairOnly
fail_signatures: {repeat_failure_same_bind}
next_modules: [CREATIVE.M06.STRESS_T1_T4_v0.1]

module_id: CREATIVE.M11.KAPPA_DEOVERFIT_WEEK_v0.1
label: κ De-Overfit Week (transfer day)
purpose: break monoculture patterns
role_mode: Operator
allowed_states: {P1,P2,P3}
inputs:
  - idea outputs from 1 template
outputs:
  - outputs in 3 formats (speech/email/report) + 2 new themes
sensors_used: CorridorCount, TransferScore, AST_k_trap_flag
tests_required: T4 twice
actuation_tags: [↓κ, ↑ρ]
runtime_minutes: 45
promotion_effect: RepairOnly
fail_signatures: {format_lock}
next_modules: [CREATIVE.M06.STRESS_T1_T4_v0.1]

module_id: CREATIVE.M12.SAFE_MODE_v0.1
label: Creative Safe Mode
purpose: allow creativity without collapse under high stress
role_mode: Mixed
allowed_states: {P0,P1}
inputs:
  - system failure or weak topic
outputs:
  - only: sensor + metric + compression + repair loop
  - disallow: open exploration generators
sensors_used: Ω_proxy, S_proxy, LabelTrap_rate, T2/T4
tests_required: T2 + T4 minimum
actuation_tags: [↑B, ↑R, ↑β, ↓K]
runtime_minutes: 30
promotion_effect: RepairOnly
fail_signatures: {exploration_causes_more_failures}
next_modules: [CREATIVE.M05.FAILUREMODE_INVERT_v0.1, CREATIVE.M04.METRIC_DESIGN_v0.1, CREATIVE.M09.COMPRESSION_12LINE_v0.1]

PLATFORM_BINDINGS_v0.1:
  SEOUL:
    delivery: SEOUL.EDU.PLATFORM.SEOULLEARN
    default_modules_for_Q1Q4: [M12, M01, M02, M10]
  TOKYO:
    delivery: TOKYO.EDU.PLATFORM.VLP
    default_modules_for_Q1Q4: [M12, M01, M02, M10]

CREATIVE_ROUTER_v0.1:
  if quadrant in {Q1,Q4} or P_state in {P0,P1}:
    dispatch: [M12] + first_two_needed_repairs
  else:
    dispatch: choose based on failure signature:
      C1 (fails T4): M03 or M11
      C2 (fails T3): M10 + M08
      C3 (fails T2): M09 + ↓K checkpointing
      C4 (template lock): M11
    then run M06 to gate promotion

CREATIVEOS_MODULE_REGISTRY_v0.1:
  Modules: M01..M12
  Tests: T1..T4 (+ RemovalTest optional)
  Promotion: Draft→Stable→Canonical via M06 + M09
  Anti-κ: M03 + M11
  SafeMode: M12 for P0/P1 or Q1/Q4
  Platform: SeoulLearn / VLP as delivery corridor

seed_nodes:
  A: "high-stakes gate raises coupling" (GATE_PRESSURE)
  B: "κ-trap = accuracy↑ while transfer↓" (KAPPA_TRAP)
  C: "T4 context swap measures transfer" (T4_SENSOR)
  D: "repair routing reduces collapse" (ROUTER_REPAIR)

GEN04_FAILUREMODE_INVERT:
  system X: "Exam-driven prep"
  failure_trace:
    gate_pressure → K↑ → κ↑ → template_overfit → transfer↓ → collapse under novelty
  earliest_sensor:
    T4_pass_rate trend down + AST_k_trap_flag
  repair_loop:
    enforce transfer-day + dual-corridor + T4 gating

bind_path:
  GATE_PRESSURE (CAUSE) KAPPA_TRAP
  KAPPA_TRAP (MECHANISM) TRANSFER_COLLAPSE
  T4_SENSOR (METRIC) TRANSFER_COLLAPSE
  ROUTER_REPAIR (MECHANISM) KAPPA_TRAP_REDUCTION
  TRANSFER_GATE (GOAL) SYSTEM_STABILITY

idea_id: IDEA.CITY.TRANSFER_FIRST_GATE_v0.1
spec_12:
1) High-stakes gates increase K and encourage single-template optimization (κ↑).
2) κ-trap is detected when Accuracy trends ↑ but Transfer (T4) trends ↓.
3) T4 ContextSwap is the city’s early-warning sensor for brittle learning.
4) If T4 declines for 2 consecutive windows, trigger “Transfer-First Gate.”
5) Transfer-First Gate = require 2 corridors (2 templates/methods) before advanced drilling.
6) Enforce a weekly Transfer Day (new theme + new format) to force ρ↑ and κ↓.
7) Route repairs through public corridor modules (SeoulLearn/VLP) for fast R↑.
8) Use checkpointing to decouple K (intro→mechanism→consequence).
9) Re-test T2 and T4 within 7 days after repair.
10) Promotion condition: T4>=0.70 and T3>=0.70 and T2>=0.70.
11) Output metric: κ-trap flags downtrend while accuracy holds.
12) Result: InnovationEvent allowed only when Ω>=1 during S>=Θ windows.

T1 promptshift: PASS (3 rephrasings still consistent)
T2 timelimit: PASS (60s coherent)
T3 counterargument: PASS ("Won’t it slow top students?" → allow when Ω margin exists)
T4 contextswap: PASS (applied to math/science not just writing)
RemovalTest: PASS (remove T4 gate → idea loses detection mechanism)

CorridorCount: 2 (policy + classroom implementation)
TransferScore(T4): 0.75
CounterScore(T3): 0.72
=> promote: Stable → Canonical(P3 candidate)

If district quadrant in {Q1,Q4}:
  dispatch modules: [M12 SAFE_MODE, M03 DUAL_CORRIDOR_SWITCH, M11 κ DEOVERFIT WEEK, M06 STRESS]
Else:
  dispatch modules: [M03, M11, M06, M09]

seed_nodes:
  A: "topic sentence often vague" (VAGUE_TS)
  B: "typed binds create coherence" (TYPED_BINDS)
  C: "time collapse occurs without compression" (TIME_COLLAPSE)
  D: "context swap requires mechanism not adjectives" (TRANSFER_NEED)

GEN09_METRIC_DESIGN:
  goal: improve paragraph quality under time
  metric: BDI_score_raw + T2_pass_rate + T4_pass_rate
  threshold Θ: BDI_score_raw >= target AND T2>=0.70 AND T4>=0.70

bind_path:
  VAGUE_TS (CAUSE) LOW_BDI
  TYPED_BINDS (MECHANISM) HIGH_BDI
  HIGH_BDI (MECHANISM) TIME_RESILIENCE
  MECHANISM_SPINE (MECHANISM) TRANSFER_SUCCESS

idea_id: IDEA.STUDENT.MECHANISM_SPINE_PARAGRAPH_v0.1
spec_12:
1) Write topic sentence as a mechanism claim, not a feeling/adjective.
2) Use 3 typed binds only: CAUSE, EXAMPLE, CONSEQUENCE.
3) Sentence 2: CAUSE (because/therefore) — state mechanism.
4) Sentence 3: EXAMPLE — concrete proof.
5) Sentence 4: CONSEQUENCE — what changes because of the mechanism.
6) Ban filler adjectives unless tied to a typed bind.
7) Under time limit, stop after consequence and submit.
8) Repair loop: rewrite weakest sentence with stronger bind.
9) Transfer drill: apply same mechanism spine to a new theme (T4).
10) Counter drill: write “why this is wrong” in 1 sentence (T3-lite).
11) Track BDI (valid binds/100w) weekly.
12) Promotion: stable when T2 and T4 both pass ≥0.70.

T1 promptshift: PASS (same spine works across prompts)
T2 timelimit: PASS (fast because only 4 sentences)
T3 counterargument: PASS (one-sentence “why wrong” added)
T4 contextswap: PASS (theme change still works)

M02 BDI_BIND_PACK_BUILDER → M09 COMPRESSION_12LINE → M06 STRESS
If LabelTrap detected: prepend M01 LABELTRAP_REPAIR

seed_nodes:
  A: "education = regeneration pipeline" (REGEN_PIPELINE)
  B: "school = time compression escalator" (ESCALATOR)
  C: "router repairs prevent collapse" (REPAIR_ROUTER)
  D: "corridors = redundancy" (RHO_REDUNDANCY)

GEN05_ANALOGY_BRIDGE:
  map from: network reliability engineering
  to: education system design
  claim: the key is routing + redundancy + repair latency, not just throughput

bind_path:
  REGEN_PIPELINE (MECHANISM) CIV_STABILITY
  ESCALATOR (CONDITION) THROUGHPUT_ONLY
  THROUGHPUT_ONLY (FAILUREMODE) KAPPA_TRAP
  REPAIR_ROUTER (MECHANISM) STABILITY_UNDER_LOAD
  RHO_REDUNDANCY (MECHANISM) TRANSFER_SURVIVAL

idea_id: IDEA.CROSSDOMAIN.EDU_AS_ROUTED_REPAIR_v0.1
spec_12:
1) Education is a repair-routing network that regenerates capability over time.
2) School is one delivery lane (a time compression escalator), not the system itself.
3) When one lane dominates, κ rises: students overfit to one template and transfer fails.
4) A stable education system requires redundancy (ρ): multiple corridors to the same skill.
5) It requires buffers (B): time/sleep/slack so learning survives stress windows.
6) It requires fast repair (R): failures patched within 24h, not after exams.
7) The city’s early sensor is transfer (T4), not word-count or homework volume.
8) Bind strength (β) is measured via BDI: typed binds that hold under time.
9) When stress S rises near gates Θ, require Ω>=1 via routing and redundancy.
10) Therefore: “more tuition” or “more drilling” is not a solution if κ rises.
11) Correct actuation order is FenceOrder: ↑B → ↓K → ↑R → ↑β → ↑ρ → ↓κ.
12) Result: stable innovation (learning gains) without collapse under novelty.

T1 promptshift: PASS
T2 timelimit: PASS
T3 counterargument: PASS (schools still matter as a lane; claim is “not equal”)
T4 contextswap: PASS (applies to healthcare training, workplace training)

CorridorCount: 3 (education, network reliability, city routing)
TransferScore: 0.78
CounterScore: 0.74
=> Canonical(P3 candidate)

CREATIVEOS_WORKED_EXAMPLES_v0.1:
  Example1: IDEA.CITY.TRANSFER_FIRST_GATE_v0.1 (policy)
  Example2: IDEA.STUDENT.MECHANISM_SPINE_PARAGRAPH_v0.1 (skill protocol)
  Example3: IDEA.CROSSDOMAIN.EDU_AS_ROUTED_REPAIR_v0.1 (analogy corridor)
  Each includes: seed_nodes → bind_path → spec_12 → T-tests → promotion

system_id: CREATIVEOS×CITY_DASHBOARD_v0.1
compatible_with:
  - KOR×SEOUL×EDU×PANEL_v0.1
  - JPN×TOKYO×EDU×PANEL_v0.1
granularity:
  - Z4 (SEOUL.GU.<KEY> / TOKYO.WARD.<KEY> / TOKYO.ISLAND.MUNI.<KEY>)
cadence: weekly (YYYY-WW)
roles:
  Operator (submit ideas + tests)
  Oracle (gate + promote)
  Visionary (new corridor design when Ω margin exists)

CreativePanelFields:
  district_id
  cohort_id
  time_window
  sample_n_ideas
  # Flow
  drafts_count
  stress_runs_count          # number of M06 executions
  stable_count               # passed T2+T3+T4 once
  canonical_count            # promoted this week
  cumulative_canonical
  # Survival
  T2_pass_rate               # time-limit survival
  T3_pass_rate               # counterargument survival
  T4_pass_rate               # transfer survival
  removaltest_pass_rate      # optional (bind integrity)
  # Structure
  avg_BDI_idea               # binds per 100w × coherence
  avg_corridor_count         # mean distinct templates per idea
  novelty_score_mean         # rough domain-distance × bind-path length
  # Risk
  kappa_trap_rate            # % ideas with (A↑ & Transfer↓) or template lock
  collapse_signature_rate    # C1–C4 occurrences / drafts
  avg_repair_latency_hours   # median hours draft→patched
  # Integration with EducationOS
  linked_quadrant            # Q1–Q4 from Education panel
  linked_P_state             # P0–P3
  S_proxy
  Omega_proxy
  # Router
  modules_dispatched         # list of Creative module IDs
  delivery_corridor          # SeoulLearn / VLP / Mixed

CSI =
  0.30*T2_pass_rate +
  0.30*T4_pass_rate +
  0.20*T3_pass_rate +
  0.20*clip01(avg_BDI_idea / BDI_TARGET_BY_GRADE)

### 2.2 κ Risk Index (KRI)
```text
KRI =
  0.50*kappa_trap_rate +
  0.30*(1 - avg_corridor_count/2 clipped 0..1) +
  0.20*collapse_signature_rate

### 2.3 Innovation Activation Score (IAS)
```text
IAS =
  canonical_count / max(1, drafts_count)
  × clip01(Omega_proxy)
  × clip01(S_proxy / Theta_default)

---
## 3) Quadrant Overlay (Creative × Education)
```text
CreativeQuadrant:
  if linked_quadrant in {Q1,Q4}:
    mode = SAFE_MODE_ONLY      # allow only M12/M05/M09 repair/compression
  else:
    mode = FULL_GENERATORS     # M01–M12 allowed
CreativeState:
  C0: CSI<0.60 and KRI>0.50         # brittle + monoculture
  C1: CSI<0.75 but improving        # developing
  C2: CSI>=0.75 and KRI<=0.40       # stable creative base
  C3: C2 + IAS high 2 windows       # innovation active

Columns:
  district_id | CSI | KRI | IAS | canonical_count | avg_repair_latency_hours | linked_P_state
Sort:
  1) CSI desc
  2) KRI asc

Filter:
  KRI > 0.50 OR T4_pass_rate < 0.65
Output:
  district_id, collapse_signature_rate, kappa_trap_rate,
  recommended_modules (from router)

drafts_count
  → stress_runs_count
    → stable_count
      → canonical_count
ConversionRates:
  draft_to_stable = stable_count / drafts_count
  stable_to_canonical = canonical_count / stable_count

avg_corridor_count by district
Heatmap:
  <1.5 = red (template lock)
  1.5–2.0 = amber
  >=2.0 = green

CREATIVE_CITY_ROUTER_v0.1:
if KRI > 0.50:
  dispatch: [M11 KAPPA_DEOVERFIT_WEEK, M03 DUAL_CORRIDOR_SWITCH]
if CSI < 0.60:
  dispatch: [M12 SAFE_MODE, M02 BDI_BIND_PACK_BUILDER, M10 COUNTEREXAMPLE_PATCH]
if T4_pass_rate < 0.65:
  dispatch: [M03, M06 STRESS_T1_T4]
if avg_repair_latency_hours > 24:
  dispatch: [M10 COUNTEREXAMPLE_PATCH, enforce 24h sprint]
if IAS high and CSI>=0.75:
  allow: M07 ZOOM_LIFT (city-scale publication)

SEOUL → SEOUL.EDU.PLATFORM.SEOULLEARN
TOKYO → TOKYO.EDU.PLATFORM.VLP

CITY_COMPARE_v0.1:
  metrics:
    mean_CSI
    mean_KRI
    mean_IAS
    total_canonical
    median_repair_latency
  normalized_by:
    drafts_per_1000_students (optional)
Output:
  rank_by_CSI
  risk_by_KRI
  innovation_by_IAS

CREATIVEOS_CITY_DASHBOARD_v0.1:
  Inputs: idea drafts + M06 stress results + Education panel linkage
  Metrics: CSI, KRI, IAS
  States: C0–C3
  Router: module dispatch based on CSI/KRI/T4/repair_latency
  Views:
    - Leaderboard
    - Early Warning
    - Promotion Funnel
    - Corridor Diversity Map
  Platforms:
    Seoul → SeoulLearn
    Tokyo → VLP

system_id: CIVOS×CREATIVE_INDEX×CITY_ENVELOPE_v0.1
places_supported:
  - KOR×SEOUL
  - JPN×TOKYO
zoom: Z5 (city) → Z6 (nation comparability)
cadence: weekly (YYYY-WW) with quarterly rollups
inputs:
  - CreativeOS City Dashboard metrics (CSI, KRI, IAS, repair latency, promotions)
  - EducationOS panel linkage (Ω_proxy, S_proxy, quadrant, P_state)
outputs:
  - CityInnovationStabilityEnvelope (CISE)
  - CityCreativeRiskState (CR0–CR3)
  - recommended city actuation (FenceOrder + Creative modules)

City Creativity = regenerative capability in “idea space”
Collapse in idea space happens when:
  idea-loss / brittleness rate  >  idea-repair / regeneration rate
So we measure:
  - Creative regeneration throughput (new stable ideas)
  - Brittleness drift (κ-trap + transfer failure)
  - Repair speed (latency)
  - Stress windows (S relative to Θ) using Education panel coupling

for each district i:
  CSI_i, KRI_i, IAS_i,
  canonical_count_i,
  drafts_count_i,
  avg_repair_latency_i,
  T4_pass_rate_i,
  linked_Omega_i,
  linked_S_i,
  linked_Theta_i

w_i = drafts_count_i / sum(drafts_count_all)   # activity-weighted
# (You can later use student population weights.)

S_city =
  weighted_mean_i( linked_S_i / linked_Theta_i )    # >1 means stress windows

CSI_city = weighted_mean_i(CSI_i)
KRI_city = weighted_mean_i(KRI_i)
T4_city  = weighted_mean_i(T4_pass_rate_i)
Rlat_city = weighted_median_i(avg_repair_latency_i)
R_eff_city = 1 / (1 + Rlat_city/24)
Omega_city_creative =
  (0.45*CSI_city + 0.25*T4_city + 0.30*R_eff_city)
  /
  (0.60 + 0.60*KRI_city)

canonical_city = sum_i(canonical_count_i)
drafts_city    = sum_i(drafts_count_i)
PromotionRate_city = canonical_city / max(1, drafts_city)
G_city =
  PromotionRate_city * clip01(CSI_city) * clip01(T4_city)

D_city =
  0.55*KRI_city +
  0.25*(1 - T4_city) +
  0.20*clip01(1 - CSI_city)

R_creative = D_city / max(eps, G_city)
eps = 0.001

CISE := {
  place_id,
  time_window,
  S_city,
  Omega_city_creative,
  CSI_city,
  KRI_city,
  T4_city,
  Rlat_city,
  G_city,
  D_city,
  R_creative,
  state: {CR0,CR1,CR2,CR3},
  recommended_actuation: [...]
}

CR3 (Innovation Active):
  S_city>=1 AND Omega_city_creative>=1 AND R_creative<1 AND CSI_city>=0.75 AND KRI_city<=0.40
CR2 (Stable Base):
  Omega_city_creative>=1 AND CSI_city>=0.75 AND KRI_city<=0.45
CR1 (Developing / Fragile):
  Omega_city_creative in [0.85,1.0) OR CSI_city in [0.60,0.75)
CR0 (Creative Attrition Risk):
  Omega_city_creative<0.85 OR R_creative>=1 OR KRI_city>0.55

CITY_ACTUATION_POLICY_v0.1:
if state==CR0:
  enforce: Creative Safe Mode city-wide during S_city>=1 windows
  dispatch modules:
    [M12 SAFE_MODE, M02 BDI_BIND_PACK, M10 COUNTEREXAMPLE_PATCH, M11 κ_DEOVERFIT]
  and FenceOrder emphasis:
    ↑B (buffers) + ↑R (repair latency) + ↓κ (monoculture)
if state==CR1:
  dispatch:
    [M03 DUAL_CORRIDOR_SWITCH, M06 STRESS_T1_T4, M09 COMPRESSION_12LINE]
  target:
    CSI_city↑, T4_city↑, Rlat_city↓
if state==CR2:
  allow:
    full generators
  add:
    M07 ZOOM_LIFT for a few districts to publish city-scale canonical ideas
if state==CR3:
  publish:
    Canonical(P3) bundles weekly
  maintain:
    κ guard (M11 monthly)

CITY_COMPARE_RECORD := {
  place_id,
  window,
  CISE.state,
  S_city,
  Omega_city_creative,
  R_creative,
  CSI_city,
  KRI_city,
  canonical_city,
  median_repair_latency
}

CIVOS_CREATIVE_CITY_ENVELOPE_v0.1:
  Aggregate from Z4 Creative Dashboard + Education Ω/S linkage
  Compute:
    S_city = mean(linked_S/Θ)
    Ω_city_creative = f(CSI,T4,R_eff,KRI)
    G_city (regeneration) vs D_city (brittleness drift)
    R_creative = D_city / G_city
  Output: CISE + CR0–CR3 state + actuation policy
  Purpose: measure city innovation as regenerative capability under load