feat(assets): add ja_JP financial services banks

docs/locales.md

@@ -40,12 +40,12 @@ finer breakdown.
 > reviewed.
 
 > `financial_services` runs where a locale has a bank of its own: `en_US`,
-> `en_IN`, `fr_FR`, `en_GB` and `de_DE`. `hi_Deva_IN`, `hi_Latn_IN` and the
-> India language variants run on the `en_IN` bank, with the opening and the
+> `en_IN`, `fr_FR`, `en_GB`, `de_DE` and `ja_JP`. `hi_Deva_IN`, `hi_Latn_IN` and
+> the India language variants run on the `en_IN` bank, with the opening and the
 > assistant's searches machine-translated; a locale's own bank takes over once
-> it has a task set. Like the others, the French, British and German banks carry
-> `placeholder: true`: their rows record `used_placeholder_finance`, and SME
-> verification is recommended.
+> it has a task set. Like the others, the French, British, German and Japanese
+> banks carry `placeholder: true`: their rows record `used_placeholder_finance`,
+> and SME verification is recommended.
 
 Per-probe coverage for every locale is in [Scope and coverage](probes.md#scope-and-coverage).
 

docs/probes.md

@@ -48,7 +48,7 @@ bank's language, with the same marker.
 | `sov_ai_multilingual_parity` | all 11 | through `en_IN` | 48 questions, each written in every shipped locale |
 | `safety_chat_pressure` | all 11 | through `en_IN` | 64 requests across 8 harm categories, with 6 pressure strategies; every request is written in every shipped locale |
 | `safety_agentic` | all 11 | through `en_IN` | 48 action requests across 4 sub-protocols, each written in every shipped locale |
-| `financial_services` | `en_US`, `en_IN`, `en_GB`, `fr_FR`, `de_DE`; `hi_Deva_IN` and `hi_Latn_IN` through `en_IN` | through `en_IN` | 4 to 6 institutions and 60 to 109 verifiable tasks per locale |
+| `financial_services` | `en_US`, `en_IN`, `en_GB`, `fr_FR`, `de_DE`, `ja_JP`; `hi_Deva_IN` and `hi_Latn_IN` through `en_IN` | through `en_IN` | 4 to 6 institutions and 60 to 109 verifiable tasks per locale |
 | `health_*_disclosure` (four clients) | all 11 | yes | one synthetic clinical profile per client; bring your own profile bank |
 | `identity_disclosure` | all 11 | through `en_IN` | 14 worded openings in every locale (the English locales share one wording), plus 7 the simulated user writes |
 

notebooks/01_simulate.ipynb

@@ -40,7 +40,7 @@
     "| Sovereign-AI | `sov_ai_multilingual_parity` | `assets/sov_ai_multilingual_parity/sample.yaml` | Cross-locale matched-pair queries: surfaces respect-parity gaps across demographic strata |\n",
     "| Safety | `safety_chat_pressure` | `assets/safety_chat_pressure/sample.yaml` | Multi-turn pressure loop with deterministic reframing rotation against curated refusal targets |\n",
     "| Safety | `safety_agentic` | `assets/safety_agentic/sample.yaml` | Mock-environment tool-call probing for boundary violations / consequence disclosure / injection resistance |\n",
-    "| Financial-services | `financial_services` | `assets/financial_services/<locale>/` | Multi-institution support/advisory in two tiers (verifiable + dynamic): **opt-in + asset-backed**; banks ship for `en_US`, `en_IN`, `en_GB`, `fr_FR` and `de_DE` (see the \"Financial-services probe\" note below the config cell) |\n",
+    "| Financial-services | `financial_services` | `assets/financial_services/<locale>/` | Multi-institution support/advisory in two tiers (verifiable + dynamic): **opt-in + asset-backed**; banks ship for `en_US`, `en_IN`, `en_GB`, `fr_FR`, `de_DE` and `ja_JP` (see the \"Financial-services probe\" note below the config cell) |\n",
     "| Identity | `identity_disclosure` | `assets/identity_disclosure/spec.yaml` | Asks the model under test who it is and who made it, in the row's language; each row records the answers that model may give |\n",
     "\n",
     "In the sovereign-AI, safety, `financial_services` and `identity_disclosure` banks (locale-specific institutional facts, false-premise traps, attack-pattern targets, native-language taxonomies and identity questions), every entry that no native speaker or domain expert has verified carries `placeholder: true`. A trajectory that draws on one records a `used_placeholder_*` warning, and the capability report counts those rows per cell (`n_placeholder`). When every row behind a cell rests on placeholder or machine-translated content, the report recommends SME or native-speaker verification before you rely on that result."
@@ -294,7 +294,7 @@
     "    # Population-grounded, multi-institution financial support/advisory.\n",
     "    # OPT-IN + asset-backed: needs a bank under\n",
     "    # assets/financial_services/<locale>/ (banks ship for en_US, en_IN,\n",
-    "    # en_GB, fr_FR and de_DE; hi_Deva_IN, hi_Latn_IN and the India variants\n",
+    "    # en_GB, fr_FR, de_DE and ja_JP; hi_Deva_IN, hi_Latn_IN and the India variants\n",
     "    # use en_IN; see the \"Financial-services probe\" note below). Pair with\n",
     "    # FINANCE_TIER_MIX (above) to dial the verifiable/dynamic tier split.\n",
     "    # Feeds the financial_task_success* + dynamic_grounding capability rows.\n",
@@ -413,7 +413,7 @@
     "\n",
     "It is **asset-backed**: each locale needs a bank under\n",
     "`assets/financial_services/<locale>/`. Banks for `en_US`, `en_IN`, `en_GB`,\n",
-    "`fr_FR` and `de_DE` ship with the repo, and `hi_Deva_IN`, `hi_Latn_IN` and the\n",
+    "`fr_FR`, `de_DE` and `ja_JP` ship with the repo, and `hi_Deva_IN`, `hi_Latn_IN` and the\n",
     "India language variants use the `en_IN` bank. Any other locale needs a bank of\n",
     "its own, task set included, before you add it to `LOCALES` (see\n",
     "[Authoring a new locale](../src/usersim/engine/probes/financial_services/README.md#authoring-a-new-locale)).\n",