@wildo-ai/framework-knowledge 1.1.12 → 1.1.14

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (101) hide show
  1. package/dist/claude-marketplace/.claude-plugin/marketplace.json +1 -1
  2. package/dist/claude-marketplace/plugins/wildo/.claude-plugin/plugin.json +1 -1
  3. package/dist/claude-marketplace/plugins/wildo/references/composites-and-the-loop.md +12 -4
  4. package/dist/claude-marketplace/plugins/wildo/skills/build/SKILL.md +4 -0
  5. package/dist/esm/framework-knowledge.definitions.d.ts +1 -1
  6. package/dist/esm/framework-knowledge.definitions.d.ts.map +1 -1
  7. package/dist/esm/framework-knowledge.definitions.js +10 -0
  8. package/dist/esm/framework-knowledge.definitions.js.map +1 -1
  9. package/dist/esm/playbooks/bottom_up.foundation/playbook.definition.d.ts +2 -1
  10. package/dist/esm/playbooks/bottom_up.foundation/playbook.definition.d.ts.map +1 -1
  11. package/dist/esm/playbooks/bottom_up.foundation/playbook.definition.js +8 -1
  12. package/dist/esm/playbooks/bottom_up.foundation/playbook.definition.js.map +1 -1
  13. package/dist/knowledge/assurance-authority-bases/wildo-safe-baseline-v1/external-framework-mapping.json +1 -1
  14. package/dist/knowledge/manifest.json +7 -7
  15. package/dist/knowledge/playbooks/bottom_up.foundation/method.md +6 -0
  16. package/dist/knowledge/playbooks/bottom_up.foundation/playbook.json +6 -3
  17. package/dist/knowledge/playbooks/bottom_up.maturity-reading/method.md +9 -0
  18. package/dist/knowledge/playbooks/bottom_up.maturity-reading/playbook.json +1 -1
  19. package/dist/knowledge/playbooks/refinement.compliance-document-candidate/method.md +11 -2
  20. package/dist/knowledge/playbooks/refinement.compliance-document-candidate/playbook.json +1 -1
  21. package/dist/knowledge/playbooks/top_down.requirements-definition/method.md +14 -0
  22. package/dist/knowledge/playbooks/top_down.requirements-definition/playbook.json +1 -1
  23. package/dist/knowledge/references/application-maturity.config.md +43 -0
  24. package/dist/knowledge/references/engine.client-bundle.allow-list.md +6 -0
  25. package/dist/knowledge/references/example.backend.addon-wiring.module.md +5 -1
  26. package/dist/knowledge/references/example.backend.custom-operation.declared-refusal-code.md +6 -0
  27. package/dist/knowledge/references/example.config.navigation.full.md +14 -0
  28. package/dist/knowledge/references/example.infra.environment-config.local.md +2 -2
  29. package/dist/knowledge/references/example.resource.config-wiring.retention-policy-operations.md +21 -0
  30. package/dist/knowledge/references/example.resource.status-lifecycle.quote.md +1 -1
  31. package/dist/knowledge/references/example.resource.ui-behavior.operation-icon.md +8 -2
  32. package/dist/knowledge/references/platform.config.environment.deployment.md +60 -4
  33. package/dist/knowledge/references/platform.config.infrastructure-host.md +1 -0
  34. package/dist/knowledge/references/platform.config.platform-key-management.md +5 -2
  35. package/dist/knowledge/references/requirements.engine-substance.md +38 -0
  36. package/dist/knowledge/references/requirements.functional-requirement.config.md +11 -0
  37. package/dist/knowledge/references/requirements.non-functional-requirement.config.md +11 -0
  38. package/dist/knowledge/references/saas.approval-seals.admin-endpoints.md +20 -0
  39. package/dist/knowledge/references/saas.backend.custom-operation.hook-transaction-posture.md +12 -1
  40. package/dist/knowledge/references/saas.backend.flows-actors.function-tool.md +72 -0
  41. package/dist/knowledge/references/saas.backend.system-access-service.md +30 -1
  42. package/dist/knowledge/references/saas.frontend.app-shell.shell-motion-contract.md +30 -0
  43. package/dist/knowledge/references/saas.frontend.chart-dashboard-view.md +3 -3
  44. package/dist/knowledge/references/saas.frontend.chart-dashboard.md +38 -35
  45. package/dist/knowledge/references/saas.models.chart-dashboard-tile-sizes.md +40 -0
  46. package/dist/knowledge/references/saas.models.flows-actors.agent-content-guardrails-environment.md +52 -0
  47. package/dist/knowledge/references/saas.models.frontend-provider-public-config-contract.md +6 -0
  48. package/dist/knowledge/references/saas.models.organization-dashboard-layout.md +61 -0
  49. package/dist/knowledge/references/saas.models.privacy.data-residency.md +188 -31
  50. package/dist/knowledge/references/saas.models.providers.compliance-choice.md +9 -0
  51. package/dist/knowledge/references/saas.models.resource-config.operation-condition-parameters.md +8 -1
  52. package/dist/knowledge/references/saas.specifications.processor-register.md +10 -0
  53. package/dist/knowledge/registries/assurance-authority-bases.json +1 -1
  54. package/dist/knowledge/registries/framework-topics.json +41 -1
  55. package/dist/knowledge/registries/playbooks.json +1 -1
  56. package/dist/knowledge/registries/source-tags.json +1147 -1077
  57. package/dist/knowledge/skills/backend-addon-wiring/SKILL.md +5 -1
  58. package/dist/knowledge/skills/backend-custom-operation/references/declared-refusal-codes.md +28 -1
  59. package/dist/knowledge/skills/backend-custom-operation/references/framework-hook-transaction-posture-api.md +32 -1
  60. package/dist/knowledge/skills/compliance-document-authoring/references/document-composition-privacy-notice.md +11 -2
  61. package/dist/knowledge/skills/compliance-document-authoring/references/framework-processor-register-authoring.md +16 -0
  62. package/dist/knowledge/skills/compliance-document-authoring/references/framework-provider-compliance-choice-api.md +197 -31
  63. package/dist/knowledge/skills/config-app-shell/references/framework-shell-layout-api.md +30 -0
  64. package/dist/knowledge/skills/config-navigation/SKILL.md +14 -0
  65. package/dist/knowledge/skills/config-saas/references/compliance-declarations.md +19 -0
  66. package/dist/knowledge/skills/config-saas/references/framework-key-management-api.md +58 -1
  67. package/dist/knowledge/skills/electronic-signature-deliverables/references/what-makes-a-request-signed.md +20 -3
  68. package/dist/knowledge/skills/flows-actors-autonomous-agents/SKILL.md +12 -2
  69. package/dist/knowledge/skills/flows-actors-chat-assistant/SKILL.md +16 -4
  70. package/dist/knowledge/skills/flows-actors-chat-assistant/references/framework-agent-content-guardrails-api.md +144 -6
  71. package/dist/knowledge/skills/flows-actors-chat-assistant/references/framework-flows-actors-tool-api.md +72 -0
  72. package/dist/knowledge/skills/flows-actors-declared-flow/SKILL.md +12 -2
  73. package/dist/knowledge/skills/flows-actors-llm-agent-invocation/SKILL.md +12 -2
  74. package/dist/knowledge/skills/frontend-component-catalog/SKILL.md +1 -1
  75. package/dist/knowledge/skills/frontend-component-catalog/references/framework-component-catalog.md +3 -2
  76. package/dist/knowledge/skills/frontend-component-catalog/references/framework-component-usage.md +13 -0
  77. package/dist/knowledge/skills/infra-environment-config/SKILL.md +36 -8
  78. package/dist/knowledge/skills/infra-environment-config/references/framework-agent-content-guardrails-environment-api.md +67 -0
  79. package/dist/knowledge/skills/infra-environment-config/references/framework-environment-deployment-api.md +60 -4
  80. package/dist/knowledge/skills/infra-environment-config/references/framework-infrastructure-host-api.md +1 -0
  81. package/dist/knowledge/skills/infra-environment-config/references/framework-platform-key-management-api.md +5 -2
  82. package/dist/knowledge/skills/layout-2d-adoption/references/framework-shell-motion-api.md +30 -0
  83. package/dist/knowledge/skills/provider-backend-package/SKILL.md +17 -3
  84. package/dist/knowledge/skills/provider-frontend-file-picker/references/framework-frontend-public-config-api.md +6 -0
  85. package/dist/knowledge/skills/provider-frontend-module/references/package-frontend-facets.md +4 -1
  86. package/dist/knowledge/skills/resource-config-wiring/SKILL.md +6 -2
  87. package/dist/knowledge/skills/resource-config-wiring/references/framework-operation-audit-api.md +4 -2
  88. package/dist/knowledge/skills/resource-config-wiring/references/framework-retention-policy-operations.md +21 -0
  89. package/dist/knowledge/skills/resource-config-wiring/references/framework-system-access-reads-api.md +30 -1
  90. package/dist/knowledge/skills/resource-status-lifecycle/SKILL.md +3 -2
  91. package/dist/knowledge/skills/resource-status-lifecycle/references/framework-operation-condition-parameters-api.md +8 -1
  92. package/dist/knowledge/skills/resource-status-lifecycle/references/quote-lifecycle-worked-example.md +3 -3
  93. package/dist/knowledge/skills/resource-ui-behavior/SKILL.md +8 -2
  94. package/dist/knowledge/skills/views-chart-authoring/SKILL.md +5 -1
  95. package/dist/knowledge/skills/views-composite-authoring/SKILL.md +1 -1
  96. package/dist/knowledge/skills/views-composite-authoring/references/framework-chart-dashboard-api.md +166 -39
  97. package/dist/knowledge/skills/website-bundle-discipline/SKILL.md +6 -0
  98. package/dist/knowledge/skills/website-bundle-discipline/references/website-bundle-rules.md +6 -0
  99. package/dist/knowledge/topic-index.json +292 -153
  100. package/dist/tsconfig.build.tsbuildinfo +1 -1
  101. package/package.json +4 -4
@@ -1,11 +1,11 @@
1
1
  {
2
2
  "artifactVersion": "3",
3
- "frameworkVersion": "1.1.12",
4
- "generatorVersion": "1.1.12",
5
- "generatedAt": "2026-11-12T14:36:04.596Z",
3
+ "frameworkVersion": "1.1.14",
4
+ "generatorVersion": "1.1.14",
5
+ "generatedAt": "2026-11-16T14:58:04.129Z",
6
6
  "packageName": "@wildo-ai/framework-knowledge",
7
- "topicCount": 1066,
8
- "referenceFileCount": 923,
7
+ "topicCount": 1070,
8
+ "referenceFileCount": 927,
9
9
  "payload": {
10
10
  "topicIndexPath": "topic-index.json",
11
11
  "registriesDirectory": "registries",
@@ -21,13 +21,13 @@
21
21
  "registryName": "source-tags",
22
22
  "filePath": "registries/source-tags.json",
23
23
  "description": "Source-derived tag registry for extracted framework references.",
24
- "entryCount": 923
24
+ "entryCount": 927
25
25
  },
26
26
  {
27
27
  "registryName": "framework-topics",
28
28
  "filePath": "registries/framework-topics.json",
29
29
  "description": "Topic registry derived from extracted framework-knowledge tags.",
30
- "entryCount": 1066
30
+ "entryCount": 1070
31
31
  },
32
32
  {
33
33
  "registryName": "playbooks",
@@ -334,6 +334,12 @@ slugs (`eu, ch`) or `none`, and the foundation executor writes it, deterministic
334
334
  deployment model to choose, an authored region outside or no compliance profile is reported
335
335
  on the run result with the regions that would satisfy the residency, and
336
336
  `wildo context launch-blockers` lists it until the choice is made.
337
+ - **The at-rest promise is the operator's, never asked and never written.** When the
338
+ configuration already declares `dataResidency.storedAtRestTerritories`, the executor keeps it
339
+ when it writes the territories, and places a provider none of whose regions processes inside
340
+ on a region that keeps everything at rest inside it. When it declares none, an unplaced
341
+ provider is reported with the regions an at-rest list over the same places would admit, so
342
+ the operator can decide whether to make that weaker promise.
337
343
  - **`none` writes nothing**, and an authored residency is never removed by it: removing a
338
344
  residency withdraws a commitment, which a person does on purpose. A deferred answer writes
339
345
  nothing and the run says the residency is still open.
@@ -46,7 +46,7 @@
46
46
  "outputs": [],
47
47
  "stepOutputs": [],
48
48
  "method": {
49
- "guidance": "# Foundation — Domain Plan → Configured, Bootable App Skeleton (v1 scaffold)\n\n### Target Application Framing\n\nThis knowledge is for configuring behavior in the target application.\nWildo is the framework surface being used, not the product being built.\n\nRead and apply these patterns from the application-developer point of view:\n\n- how the application should configure the relevant app-level file\n- how the application should use framework-provided factories, enums, and helpers\n- how the application should avoid mixing frontend, backend, and infrastructure concerns\n\n\nThis method composes existing tooling: `wildo init` creates the SaaS skeleton and\nattempts knowledge delivery, `wildo assets sync` refreshes that knowledge, and\n`wildo setup --non-interactive` followed by `wildo local init` configures and verifies\nthe local environment. It is guidance consumed by the application-creation lane,\nnot an automatically executed CLI script. The registered foundation executor composes\nmodule and resource shells and tenancy configuration from the domain plan; it does\nnot execute this full init/assets/setup sequence. Fulfil those prerequisites and\npostconditions separately and record actual command outcomes. Tool availability\nalone does not prove that a foundation run succeeded.\n\n## Where you are in the journey\n\nThis is the **foundation** playbook — the point in the **bottom-up** phase where\nthe *plan* becomes a *bootable app*. It runs **after** the `bottom_up.domain-plan`\nplaybook produced (and the creator approved) the **domain plan**: the module\ndecomposition, the per-module resource inventory, and the inter-resource\nrelationship intents. Its preconditions — `phase_is bottom_up` and\n`artifact_exists domain-plan` and `artifact_expected vision-brief` — mean the journey\nhas entered the build-shaping phase and the load-bearing artifact is present: the\n**domain plan**, whose modules are registered and whose tenancy decides two of the three\nderived capabilities. The vision is SEQUENCING here rather than readiness (#1434): nothing\nin this playbook reads its content, and it is enforced harder upstream — the market\nsnapshot and the requirements both gate on it, and the domain plan gates on those. The\nroute-bound `transition_committed` fact that authorized and applied crossing from top-down into\nbottom-up was the domain-plan playbook's duty (it is the first bottom-up playbook);\nyou do not re-cross that boundary — the domain plan's existence already sits behind\nit.\n\nYou are not re-opening the domain plan here — you are turning an *approved* plan\ninto the *configured application skeleton the build loop then fills in*.\n\n## What this playbook does\n\nYou are the **foundation engineer**. This playbook reads the approved domain plan\nand the vision brief, then configures and brings up the application skeleton so the\nbuild loop has a real workspace to author into. First inspect the destination: create\na fresh skeleton with `wildo init` in an empty directory, or preserve the existing\napplication. Then perform the configuration-authoring and bring-up steps below.\nThe method describes their order; it does not itself execute the commands.\n\n1. **The engine capabilities are DERIVED for you**, from the approved domain plan's tenancy and the\n business model's revenue model. You do not choose them — see \"Engine capabilities\" below.\n2. **Declare the domain-plan's modules** at the SaaS root (they are registered across\n the runtimes after the skeleton is present, so it boots with them).\n3. **Select the providers** the enabled capabilities need.\n4. **Record the compliance declarations you were given** — the operating entity and\n the jurisdictions, from the elicitation answers, into `./wildo.saas.config.ts`.\n See `config-saas` and its `references/compliance-declarations.md`.\n5. **Verify and bring up the configured project** — check the existing skeleton\n and knowledge delivery, run `wildo setup --non-interactive`, then `wildo local init`.\n\nYou emit **no artifact**. The deliverable is the app itself: `./wildo.saas.config.ts`,\nthe materialized skeleton, and a verified local initialization. Record *why* each\ncapability, module and provider was chosen in the config's own comments, where the\nnext reader of that file will find it.\n\nThis is a **foundation pass**. You enable capabilities, declare modules as shells,\nselect providers, and bring the skeleton up. Deterministic composition may create\nresource shells from the domain plan. Business-specific fields, processing, UI behavior,\nviews and navigation are the downstream build work; a generated shell does not\ncomplete that work.\n\n## Read state first — the artifacts you read, the files you write\n\nThis playbook is **idempotent** and, on a re-run, a *refinement*. Before configuring\nanything, inspect the target directory. For a new application, run `wildo init`\nin an empty destination first, then refine its generated configuration from the\napproved plan. Do not write the planned config into an empty destination and then\nexpect initial scaffolding to merge around it. For an existing application, keep\nits source and use the documented repair/composition paths; never scaffold over it.\nThen:\n\n1. Read the current journey state and scope from what this run is given — the loaded\n specifications and the workspace sections of your brief — and do not re-derive it. Confirm the domain plan is present and\n approved, and check whether a `./wildo.saas.config.ts` already exists from an\n earlier run (refine it, do not re-author — see below).\n2. Read the **domain plan** (`domain-plan` family — the modules to register + their\n resource inventory). The executor also reads the landed **business model** if there is\n one, for the billing capability alone; it may legitimately be absent, per the top-down\n phase-gate's optional posture.\n\n **Nothing else is read.** This list used to continue with the market snapshot (\"its ICPs sharpen\n capability selection\"), the requirements and the roadmap (\"which modules are MVP-critical, so you\n enable what the first release needs\"). None of the three reaches the executor, and none ever did\n (#1434). They are no longer declared as inputs either, so the declaration and the prose now say\n the same thing.\n\n// Operator affordances (the executor itself is handed the domain-plan VALUE directly):\n// `wildo context journey` — the eligibility board, runs in flight\n// `wildo context info specification-artifacts` — per-family landed state and rendered content\n\nThe executor **reads** `domain-plan` (handed to it, relationship drafts included) and the landed\n`business-model`; it **writes** `./wildo.saas.config.ts` and the materialized skeleton — app files,\ncommitted to git, which is where the record of what was decided lives.\n\nThis line used to name six families (#1434). Four of them reached nothing.\n\n## Step 1 — Derive and enable the engine capabilities\n\nEngine capabilities are the framework surfaces the app turns on (organizations, billing,\nnotifications, audit trails, and so on). **THE EXECUTOR DERIVES THEM. Nothing here chooses one.**\n\nExactly three are derivable today, and each from one declared value:\n\n| capability | derived from |\n|---|---|\n| `ORGANIZATIONS` | the domain plan's `tenancy` is `organizations` |\n| `EMAIL_TRANSACTIONAL` | the same — tenant invitations send mail |\n| `BILLING` | the business model's revenue model charges users |\n\n**The application's line can exclude a capability, and then nothing derives it.** An internal tool has no billing at\nall: on that line the `billing-provider` question is not asked, the executor never enables `BILLING` whatever else\nimplies it, and a configuration that enables it anyway is reported on the run result as contradicting the line\n(`engineCapabilitiesContradictingLine`), never switched off for you.\n\n**Compliance constraints do NOT select a capability, and this method said for months that they were\n\"the primary driver\".** They are not, and the reason is recorded where the derivation lives\n(`FOUNDATION_CAPABILITY_BASIS`, which is TOTAL over the capability vocabulary and names why each\nremaining one is underivable). The clearest case is the one this method used to point at: where an\naudit trail must LIVE — a ledger, a SIEM — is an operator decision about retention and independence\nrather than a product fact, and leaving that capability out removes no control objective, because\nassurance reads the application's own audit store.\n\nA capability nothing derives is not enabled by guessing. It is enabled deliberately, later, by\nwhoever decides it — which is why the basis map records a reason per capability instead of a\nquestion.\n\nHistorical note, kept because it is the defect rather than a detail (#1434): this section described a\nderivation the runtime cannot perform — four steps of \"walk the constraints\", \"treat these as\nsharpeners\", \"every enabled capability must trace to a compliance constraint\" — and a reader believed\ncompliance constraints shaped the foundation. Nothing walked anything. The market snapshot, the\nrequirements and the roadmap were declared inputs of this playbook for the same reason and reached\nthe same nobody; they are no longer declared.\n\nThe `config-saas` skill carries the vocabulary for the `engineCapabilities` block of\n`./wildo.saas.config.ts`. The executor writes that block; read the skill when you need to understand\nwhat an enabled capability means, not to decide one.\n\n## Step 1b — Apply the capability composition's optional packages\n\nWhen a `capability-composition` has landed (`bottom_up.capability-composition`), the executor\napplies every scenario it marks NEEDED after the domain scenarios below — the marketing website\n(`add-website`) and the documentation portal (`add-technical-doc`) today — with the application's\nown host facts: its slug and display name from `wildo.saas.config.ts`, the public domain from its\ninfrastructure environments (a remote one first, `localhost` when none declares one), and its port\nblock (website = app port + 1, docs = app port + 2, the convention the dogfood applications use).\nA package whose workspace already exists is left alone and reported as already composed. The run\nrecord names **every scenario applied and every scenario declined with the declaration each came\nfrom** (`compositionScenarios`), so a reader sees why the application has a website — or why not —\nwithout re-deriving it.\n\n## Step 1c — Lay down the line's access posture (internal tool)\n\nWho may hold an account is, for an internal tool, decided by what the application IS: the company's own people,\nadmitted by the company. So the executor, after declaring the organization type that owns the application, writes the\nline's posture through the same function the `wildo/internal-tool` blueprint applies at `wildo init --line`\n(`applyLineAccessPosture`), and reports it on the run result as `accessPosture`:\n\n| what | written | decided by |\n|---|---|---|\n| `userTypes.<each user type of the owner>.auth.registration.mode` in `./backend-api/src/saas-config.backend.ts` | `RegistrationMode.INVITATION_ONLY` | the line (not asked) |\n| organization creation | nothing to write: no door creates an organization of the owner type, and the configuration refuses a `creationPolicy` or `selfServiceCreation` on it | the line, through the owner it declares |\n| `[AuthMethod.ORG_OPENID_CONNECT]` and `[AuthMethod.ORG_SAML]` on those user types, and `CoreFeature.SINGLE_SIGN_ON` in the default feature profile | `true` / granted | the `corporate-identity` answer: `identity_provider` or `identity_provider_and_directory` |\n| `scimProvisioning` on the owner organization type, and `CoreFeature.DIRECTORY_PROVISIONING` | `{ enabled: true }` / granted | the `corporate-identity` answer: `identity_provider_and_directory` |\n\nOnly an internal tool is asked `corporate-identity`; every other line, and an application with no line yet, is not.\nWhen it is unanswered, the domain plan's owner type joining as `DIRECTORY_PROVISIONED` counts as\n`identity_provider_and_directory`; when the answer and the plan disagree, the answer is applied and the disagreement is\nlisted in `accessPosture.owed`. `none` writes nothing, and nothing already open is switched off. What the run cannot\nwrite (no owner type, user types it cannot read, a file shape it cannot address, a user type no owner lists) is listed\nthere too, with its remedy; nothing is guessed.\n\n`frontendServices.<app>.usersManagement.<user type>.register` stays as declared. It gates self-service sign-up at the\nfrontend service (password sign-up and social sign-up alike), which invitation-only registration already refuses at\nthe API, and the backend logs the overlap at startup; accepting an invitation does not read it.\n\nWhat this does NOT set up: the identity-provider connection, the verified domain and the SCIM token. Those carry\nsecrets or proofs and are records the company creates in the running application, through the product. The SHAPE of the\ncompany's identity set-up (its units, its directory's department mapping, its single sign-on posture) is a committed\ndeclaration, authored next.\n\n**The organization identity declaration.** Once the owner type is found and the corporate identity is known, the executor\nwrites `./backend-api/src/organization-identity.ts` (`APPLICATION_ORGANIZATION_IDENTITY`) and passes it as\n`organizationIdentity` in `./backend-api/src/bootstrap-config.ts`, before the host's overrides\n(`composeOwnerOrganizationIdentityDeclaration`). The backend applies it at every boot to the owner organization. It is\nreported on the run result as `organizationIdentity`:\n\n| part | written | decided by |\n|---|---|---|\n| `singleSignOn` | `requireSSO: true`, `requireEnterpriseSSOOnly: true`, `allowPasswordFallback: true`, no break-glass account | the `corporate-identity` answer: `identity_provider` or `identity_provider_and_directory`; `none` declares no posture |\n| `units` | empty | nothing typed names the company's units: OWED, never invented |\n| `directory` | not declared | each binding names a declared unit, so it is OWED with the units (`identity_provider_and_directory`) |\n| `name` | not declared | nothing typed names the company: OWED, never invented. Until it is added, every person reads the platform's placeholder \"Default Organization\"; the backend writes a declared name over that placeholder only: once the organization is named, a changed `name` is reported as drift (with the rename that adopts it), never written over the stored name |\n\nPasswords stay allowed until the company names an emergency administrator: refusing them with no break-glass account\nwould lock the company out the day its identity provider is unreachable. The owed list says to name one in\n`singleSignOn.breakGlassAccounts` and then set `allowPasswordFallback: false`. The file is rewritten only while it is\nexactly what a derivation wrote; once someone edits it (units added, an account named), it is kept and the run says so.\nAn unanswered `corporate-identity` writes nothing and is owed.\n\n**The owner type takes the domain plan's names, through a declared rename.** The platform seeds the owner\norganization for the first administrator, typed from the configuration, when the application is initialized. The\ncreation playbooks run on a started application, so by the time this step runs it usually has been, and an\norganization's type and a user's types are fixed at creation. So where the domain plan names the owner type or its\npeople differently (`group`, `employee` against the scaffold's `workspace`, `member`), the executor RENAMES them before\nlaying the posture down (`renameOwnerToDomainPlanNames`): the new keys carry `renamedFrom` naming the old ones, every\nreference follows (the first administrator's user types, the frontends' user management, relationship kinds), and the\nbackend re-types the seeded organization and every account still holding the old key at its next start, recording each\nat CRITICAL. The rename is all or nothing: one it cannot make (a spread key, a plan name already declared as another\ntype, several user types on either side) writes nothing and is listed in `accessPosture.owed`, for the build's\norganization-model task to finish. The build must never declare a second type with `ownsApplication`.\n\n## Step 1d — Report where records live outside the application\n\nWhen the domain plan's `organizationModel.recordLocations` puts a resource's records in a system of record (an ERP, a\nCRM), the run result lists each system under `systemsOfRecord`: its kind, the resources it holds, and the credentials its\ninstance will owe, read off the facet the `@wildo-ai/systems-of-record` library declares for that kind and named as the\nbackend reads them (`WILDO_SOR_<INSTANCE>__<SLOT>`). The instance is named by its kind (`systemsOfRecord.odoo`). Where\nthe facet offers several credential kinds, the result names them: which one the company uses is its choice. Report\nthose credentials as owed, like a provider's.\n\nThe executor declares neither the instance nor the library. The build's binding task declares the instance (vendor,\ndialect, partition, personal-data origin) against the facet; once it does, `wildo config sync` (and `wildo local dev`)\ndeclares the library on the backend and links it in one step, which is the only rule that writes both halves.\n\n## Step 2 — Declare the domain-plan's modules (register across runtimes after materialization)\n\nThe skeleton supplies the root workspaces and registry structure; declaring a\nmodule in configuration alone does not implement its package or contributions.\n\n1. **Declare each approved module at the SaaS root.** Use `config-saas` for the\n module's application configuration and preserve existing module identifiers.\n2. **Materialize and register module shells.** Use the delivered module-composition\n scenario and `module-registration` guidance to create the corresponding shared,\n backend and frontend contributions. Inspect the resulting registry exports;\n `config sync` is not a substitute for authored module implementation. At foundation\n depth the shells must load; business-specific resource behavior remains the build\n loop's work.\n3. **Declare every domain-plan module — and no more.** A module the domain plan named\n that is missing from the config is a GAP; a module in the config the domain plan did\n not name is an unjustified-module CONCERN. Record the declared modules (each traced\n to its domain-plan module) as a comment beside it in the config.\n\n**Do not author the resources here.** The domain plan's per-module resource inventory\ntells you *what* resources each module will own, but authoring their schemas, fields,\nand config is the **build loop's** job. If you catch yourself writing a resource\nschema, stop: that is out of this playbook's depth.\n\n## Step 3 — Select the providers\n\nEnabled capabilities that reach an external service need a **provider** (billing → a\npayment provider; notifications/email → an email provider; file storage → an object\nstore). For each such capability, select and wire the provider scope +\n`providerConfigurations` in `./wildo.saas.config.ts` (the `config-saas` skill covers\nthe provider surfaces). A value the vendor issues per deployment — an OAuth client ID, a\nfrontend SDK DSN — is authored in each environment's\n`./infrastructure/<env>/wildo.infra.<env>.config.ts > providerConfigurations` instead (the\n`infra-environment-config` skill):\n\n- **Every provider matches a declared capability.** A provider wired for no enabled\n capability is a CONCERN; an enabled capability that needs a provider with none\n selected is a GAP.\n- **Do not seed secrets here.** Provider *selection* (which provider, which scope) is\n foundation; the per-environment secret material is seeded by the deterministic\n bring-up / environment step, not authored inline. Record the selected providers\n (each matched to its capability) as a comment beside it in the config.\n\n**The choice itself is the creator's, and the interview carries it.** Billing and SMS are\nthe two capabilities whose every provider needs an account, so the executor cannot pick one:\nit reads the `billing-provider` and `sms-provider` answers from the dossier, declares the chosen\nprovider in the backend scope, selects it as primary and enables the capability. The provider's\nrequired credentials are recorded on the run result as OWED — the application boots, and the\nprovider answers once an operator sets them. Answering `none` — the interview requires an answer to\nevery question, so \"decide later\" is an option — leaves the capability owed exactly as before. The\noptions are the engine providers serving the capability plus `none`; the interview validator refuses\nanything else before the run starts. Email is the exception the executor decides itself: it has a\ncredential-free development substitute (`email-preview`) and needs no choice.\n\n## Step 4 — The compliance declarations are written by the executor (only what was told)\n\nNo agent writes these: the foundation executor writes them deterministically from the\ninterview, after the capabilities and the composition. Each answer follows a grammar the\ninterview holds it to at every door, so what reaches the executor is always readable.\n\n- **`operating-entity`** is labelled parts: `name: …`, `address: …`, `privacy contact: …`,\n one per line or separated by `;`. Each stated part the configuration lacks is written to\n `operatorComplianceIdentity.legalEntityName`, `.registeredAddress` or `.privacyContact`.\n Any subset is a valid answer: every part is independently declarable, and a missing one is\n reported downstream as outstanding (`wildo context launch-blockers`), never filled in.\n- **`operating-jurisdictions`** is `<code>: <why>` entries (`fr: established here; eu:\n customers resident here`). Each jurisdiction the configuration lacks is appended to\n `operatingJurisdictions.jurisdictions`, with its basis. A place without a basis is refused at\n the door: \"eu\" alone does not say which obligation applies.\n- **An authored value is never replaced.** A part or a jurisdiction the configuration already\n declares differently is kept, and the run result names it in `operatorDeclarations`. Nothing\n is removed, and nothing the interview does not ask (lawful bases, roles, hosting regions) is\n touched: those are legal or operational determinations nobody typed here.\n- **Nothing is invented.** The executor writes only text the creator gave. It never derives a\n company name, an address, a privacy contact or a lawful basis.\n- **Unanswered is an advisory, not a refusal.** A deferred or unanswered question writes\n nothing, and the run result says so. An answer stored before the grammar existed is reported\n as unreadable, with nothing written, until it is answered again.\n\nThis step CANNOT fail the run: a configuration the executor cannot read or edit is reported on\nthe run result with nothing written, and the application boots.\n\n### The data residency is written by the executor\n\nThe `data-residency-territories` answer is not prose: the interview stores it as jurisdiction\nslugs (`eu, ch`) or `none`, and the foundation executor writes it, deterministically.\n\n- **Every data-serving provider is placed first.** A provider the residency leaves exactly one\n region for (only offered vendor-hosted, one region whose places all lie inside) gets that\n region in `providerComplianceChoices`. An authored choice is never overwritten.\n- **The residency is written only when nothing is left unplaced**, into\n `operatorComplianceIdentity.dataResidency.territories`. `wildo config sync` refuses a\n declared residency that leaves a provider unplaced, so writing it earlier would stop the\n application starting.\n- **Anything left is named, never guessed.** A provider with several regions inside, none, a\n deployment model to choose, an authored region outside or no compliance profile is reported\n on the run result with the regions that would satisfy the residency, and\n `wildo context launch-blockers` lists it until the choice is made.\n- **`none` writes nothing**, and an authored residency is never removed by it: removing a\n residency withdraws a commitment, which a person does on purpose. A deferred answer writes\n nothing and the run says the residency is still open.\n\n(A `foundation-manifest` artifact was declared here until 2026-08-25 and is now\nretired with the app-creation verification/digest aspect. It was a bookkeeping record\nof decisions the config already carries — deleting it costs no fact about the app.)\n\n## Step 5 — Verify materialization and bring up the environment\n\nRun the existing commands from the intended application root with their documented\nprerequisites. Keep application structure, delivered knowledge and running services\nas separately verified results.\n\n1. **Verify the skeleton.** A fresh project starts with `wildo init` in an empty\n destination before the configuration-authoring steps above. Check `frontend/`,\n `shared-lib/`, `backend-api/`, `specifications/` and the generated configuration.\n On a repeat run, preserve existing source; use init's repair path or the relevant\n composition scenario instead of regenerating the project as a new application.\n\n2. **Verify application knowledge delivery.** Init attempts delivery from the\n installed packages and reports delivered or skipped assets. If dependency\n installation was skipped or a package is unavailable, install the required\n packages, then refresh the supported skills/rules and framework assets:\n\n ```bash\n wildo assets sync\n ```\n\n <!-- fwk-provenance-waiver: the command below is named in order to be REFUSED — `wildo fwk` needs a framework workspace and errors inside a generated application, which is exactly why an agent that might reach for it has to recognise the name; translating it away would leave the prohibition with nothing to prohibit. -->\n Read the result and inspect the delivered application guidance. Do not assume\n arbitrary agents/hooks are installed, and do not substitute a framework-developer\n delivery flow for this one — this command is how an application receives them.\n\n3. **Run setup** — configure the app from the authored `./wildo.saas.config.ts` and\n the `./infrastructure/` env configs (the `infra-environment-config` skill owns the\n typed infra env files; the `local-environment-bootstrap` skill owns the `wildo`\n lifecycle). Headless setup is registered and fail-closed on the administrator seed\n environment variables; it mints the environment generation without placeholders.\n\n ```bash\n wildo setup --non-interactive\n ```\n\n4. **Initialize and verify the project** — run the registered convergence journey.\n It regenerates artifacts, starts and readiness-gates backing services, reconciles\n platform Mongo state and credentials, starts native platform services, registers the\n app once, emits all runtime projections, and executes a signed managed-config read.\n\n ```bash\n wildo local init\n ```\n\n5. **Know what a PostgreSQL store owes — and that you do not write its baseline here.** Read the\n store selection you authored in the infra env config:\n\n ```bash\n rg -n \"applicationDatabase\" infrastructure/*/wildo.infra.*.config.ts\n ```\n\n Absent, or MongoDB: there is nothing to do — a collection exists on first write, there is no\n plan, and no migration is ever owed.\n\n `engine: DatabaseEngine.POSTGRESQL`: nothing creates a table implicitly, so until the INITIAL\n migration exists the backend refuses to boot on every table it declares. Do NOT accept a baseline\n with `wildo db diff --update-lockfile`: that writes the lockfile and NO migration, so it records a\n plan whose tables were never created and the boot gate still refuses.\n\n The build loop owns that migration. For an application that declares PostgreSQL and has no\n `backend-api/migrations/schema-plan.lock.json`, the dispatch host derives a `database-migrations`\n step; it runs `wildo db generate`, which, with no lockfile, renders the whole model as the initial\n migration and writes the baseline in the same invocation. The two are one change and are committed\n together. It needs a built `backend-api` (the plan is computed by importing the compiled backend\n registry), so report the store you found and leave the migration to that step.\n\nReport the status of each bring-up step to the creator from its real exit and\npostcondition results. A skipped knowledge delivery or unsuccessful initialization\nis not completion, even when source files exist.\n\n## Refine, don't re-author\n\n- **Keep what still holds.** On a re-run, do not disable a capability, drop a module,\n or re-select a provider that is already correct. Only enable a newly-justified\n capability, register a newly-added domain-plan module, or correct a provider where\n the definition set genuinely changed.\n- **Reuse existing module ids and provider scopes.** A module already registered keeps\n its id; a provider already selected keeps its scope — never mint a second one for the\n same module/capability.\n- **Follow the definition set.** If the domain plan added a module since the last run,\n reflect it (register the module) — do not re-derive the whole configuration from\n scratch. The delta is visible in `git diff` on the config, which is the record. A new\n compliance constraint does NOT enable a capability here; see Step 1 for why.\n\n## Quality gate — `foundation-configured-coherently` (declarative at scaffold depth)\n\nThe foundation is checked against the `foundation-configured-coherently` **reviewer\nchecklist** (see the end of this method): five invariants (capability derivation, module registration, provider coverage,\nsequence order, and foundation-depth boundary discipline), each marked PASS / CONCERN /\nGAP.\n\nThis checklist is **declarative**: no runtime executes it (rubric-only judge gates were retired at the flexible-intelligence hard cut — judgment runs only through a compiled Evaluation Treatment, and this playbook has none yet), so it ships as the **creator's review checklist** — nothing runs it automatically. The **trust floor** for accepting the foundation today is the\ncreator's `approve_result` review of the produced config and the bring-up result. There is\nintentionally **no deterministic gate**: config validation is a `wildo` CLI concern\n(`config sync` / validate), not a validator in the journey-artifact schemaRef\nvocabulary. Existing command permission and effect boundaries apply independently\nof this reviewer checklist.\n\n## Scaffold-depth boundary\n\nKeep this playbook at foundation depth. Explicitly **out of scope here**:\n\n- **Implementing replacement CLI tooling.** Use the existing init, asset delivery,\n setup and local initialization owners. If a command fails, retain its diagnostic\n evidence and resolve the prerequisite or concrete defect rather than inventing\n another materializer inside the application.\n- **Resource / field / schema authoring.** The config declares modules as shells and\n enables capabilities. The foundation executor can materialize resource shells\n and scope them to the chosen tenancy. Filling those shells with business-specific\n fields, rules, UI behavior, views and navigation belongs to the later build work.\n- **Re-introducing a foundation artifact.** The `foundation-manifest` family was\n retired on 2026-08-25 with the app-creation verification/digest aspect. Do **NOT**\n reshape `_core` to bring it back: the config file and its git history already carry\n every decision it recorded.\n- **Deep capability / provider modelling.** A defended capability-selection model\n (per-compliance-regime obligation → capability mapping, provider cost/feature\n matrices, multi-provider fallback) is a later depth pass. You derive capabilities\n from the definition set with the v1 heuristic above — not a solved obligation graph.\n\nIf you find yourself perfecting any of the above, stop and leave the slice at scaffold\ndepth — deeper capability modelling belongs to a dedicated later pass,\nholistically.\n\n## What you emit\n\n- The authored **`./wildo.saas.config.ts`** (an app file) — the enabled engine\n capabilities, the registered modules, and the selected providers — plus the\n registered module shells across the runtimes and the\n brought-up skeleton.\n- The **bring-up result** reported to the creator: the status of each step, taken from\n its real exit/postcondition rather than from intent. This is a report, not a stored\n artifact — nothing reads it back.\n\nThe creator reviews the result (approval posture `approve_result`) before it is\naccepted as the journey's foundation, which the bottom-up build loop then fills in\nwith authored resources, operations, UX, and integrations.\n\n## References\n\n- `.claude/skills/config-saas/SKILL.md` — the `./wildo.saas.config.ts` authoring\n surface: `engineCapabilities`, `modules`, provider scopes, and `providerConfigurations`.\n- `.claude/skills/module-registration/SKILL.md` — register each domain-plan module\n across the shared / frontend / backend registries so the skeleton boots with it\n (distinct from `config-saas`, which only declares module existence at the root).\n- `.claude/skills/infra-environment-config/SKILL.md` — the typed `./infrastructure/`\n environment configs (backing services, public domain, service auth) the bring-up needs.\n- `.claude/skills/local-environment-bootstrap/SKILL.md` — the `wildo` lifecycle\n (`setup`, `local init`, `local reinit`, `local dev`) for deterministic initialization,\n repair, clean-room recovery, and runtime start.\n\n## Reviewer checklist — `foundation-configured-coherently`\n\nFormerly authored as an `llm_judge` quality gate; the generic rubric-only gate kind was retired at the flexible-intelligence hard cut (judgment now runs only through a compiled Evaluation Treatment). The checklist itself is creator-facing review guidance and is maintained here as guidance against the current tooling:\n\nJudge whether the authored foundation (the ./wildo.saas.config.ts composition + the materialized skeleton) is a coherent realisation of the approved domain plan and the top-down definition set. Assess five invariants and mark each PASS / CONCERN / GAP with a one-line evidence note. (1) CAPABILITY DERIVATION — every enabled engine capability traces to a compliance/vision, business-model, market, or domain-plan need; unjustified capabilities are a CONCERN and missing required capabilities are a GAP. (2) MODULE REGISTRATION — every approved domain-plan module is declared and registered across runtimes; missing modules are a GAP and extra modules are a CONCERN. (3) PROVIDER COVERAGE — every provider serves a declared capability and every provider-backed capability has a provider. (4) SEQUENCE ORDER AND PROOF — materialize a fresh skeleton in an empty destination (or preserve an existing project) → refine config and module registrations → verify knowledge delivery → wildo setup --non-interactive → wildo local init. Treat setup intent, container health, or an initialization marker as insufficient: local init must reach its signed managed-config-read postcondition. (5) BOUNDARY DISCIPLINE — stay at foundation depth: deterministic resource shells are allowed, but business-specific fields, processing, UI behavior, views and navigation belong to later build work. Use the existing init/assets owners; do not replace them. Headless setup and local init run once their prerequisites are satisfied. This reviewer checklist is declarative; creator approval is the trust floor.\n",
49
+ "guidance": "# Foundation — Domain Plan → Configured, Bootable App Skeleton (v1 scaffold)\n\n### Target Application Framing\n\nThis knowledge is for configuring behavior in the target application.\nWildo is the framework surface being used, not the product being built.\n\nRead and apply these patterns from the application-developer point of view:\n\n- how the application should configure the relevant app-level file\n- how the application should use framework-provided factories, enums, and helpers\n- how the application should avoid mixing frontend, backend, and infrastructure concerns\n\n\nThis method composes existing tooling: `wildo init` creates the SaaS skeleton and\nattempts knowledge delivery, `wildo assets sync` refreshes that knowledge, and\n`wildo setup --non-interactive` followed by `wildo local init` configures and verifies\nthe local environment. It is guidance consumed by the application-creation lane,\nnot an automatically executed CLI script. The registered foundation executor composes\nmodule and resource shells and tenancy configuration from the domain plan; it does\nnot execute this full init/assets/setup sequence. Fulfil those prerequisites and\npostconditions separately and record actual command outcomes. Tool availability\nalone does not prove that a foundation run succeeded.\n\n## Where you are in the journey\n\nThis is the **foundation** playbook — the point in the **bottom-up** phase where\nthe *plan* becomes a *bootable app*. It runs **after** the `bottom_up.domain-plan`\nplaybook produced (and the creator approved) the **domain plan**: the module\ndecomposition, the per-module resource inventory, and the inter-resource\nrelationship intents. Its preconditions — `phase_is bottom_up` and\n`artifact_exists domain-plan` and `artifact_expected vision-brief` — mean the journey\nhas entered the build-shaping phase and the load-bearing artifact is present: the\n**domain plan**, whose modules are registered and whose tenancy decides two of the three\nderived capabilities. The vision is SEQUENCING here rather than readiness (#1434): nothing\nin this playbook reads its content, and it is enforced harder upstream — the market\nsnapshot and the requirements both gate on it, and the domain plan gates on those. The\nroute-bound `transition_committed` fact that authorized and applied crossing from top-down into\nbottom-up was the domain-plan playbook's duty (it is the first bottom-up playbook);\nyou do not re-cross that boundary — the domain plan's existence already sits behind\nit.\n\nYou are not re-opening the domain plan here — you are turning an *approved* plan\ninto the *configured application skeleton the build loop then fills in*.\n\n## What this playbook does\n\nYou are the **foundation engineer**. This playbook reads the approved domain plan\nand the vision brief, then configures and brings up the application skeleton so the\nbuild loop has a real workspace to author into. First inspect the destination: create\na fresh skeleton with `wildo init` in an empty directory, or preserve the existing\napplication. Then perform the configuration-authoring and bring-up steps below.\nThe method describes their order; it does not itself execute the commands.\n\n1. **The engine capabilities are DERIVED for you**, from the approved domain plan's tenancy and the\n business model's revenue model. You do not choose them — see \"Engine capabilities\" below.\n2. **Declare the domain-plan's modules** at the SaaS root (they are registered across\n the runtimes after the skeleton is present, so it boots with them).\n3. **Select the providers** the enabled capabilities need.\n4. **Record the compliance declarations you were given** — the operating entity and\n the jurisdictions, from the elicitation answers, into `./wildo.saas.config.ts`.\n See `config-saas` and its `references/compliance-declarations.md`.\n5. **Verify and bring up the configured project** — check the existing skeleton\n and knowledge delivery, run `wildo setup --non-interactive`, then `wildo local init`.\n\nYou emit **no artifact**. The deliverable is the app itself: `./wildo.saas.config.ts`,\nthe materialized skeleton, and a verified local initialization. Record *why* each\ncapability, module and provider was chosen in the config's own comments, where the\nnext reader of that file will find it.\n\nThis is a **foundation pass**. You enable capabilities, declare modules as shells,\nselect providers, and bring the skeleton up. Deterministic composition may create\nresource shells from the domain plan. Business-specific fields, processing, UI behavior,\nviews and navigation are the downstream build work; a generated shell does not\ncomplete that work.\n\n## Read state first — the artifacts you read, the files you write\n\nThis playbook is **idempotent** and, on a re-run, a *refinement*. Before configuring\nanything, inspect the target directory. For a new application, run `wildo init`\nin an empty destination first, then refine its generated configuration from the\napproved plan. Do not write the planned config into an empty destination and then\nexpect initial scaffolding to merge around it. For an existing application, keep\nits source and use the documented repair/composition paths; never scaffold over it.\nThen:\n\n1. Read the current journey state and scope from what this run is given — the loaded\n specifications and the workspace sections of your brief — and do not re-derive it. Confirm the domain plan is present and\n approved, and check whether a `./wildo.saas.config.ts` already exists from an\n earlier run (refine it, do not re-author — see below).\n2. Read the **domain plan** (`domain-plan` family — the modules to register + their\n resource inventory). The executor also reads the landed **business model** if there is\n one, for the billing capability alone; it may legitimately be absent, per the top-down\n phase-gate's optional posture.\n\n **Nothing else is read.** This list used to continue with the market snapshot (\"its ICPs sharpen\n capability selection\"), the requirements and the roadmap (\"which modules are MVP-critical, so you\n enable what the first release needs\"). None of the three reaches the executor, and none ever did\n (#1434). They are no longer declared as inputs either, so the declaration and the prose now say\n the same thing.\n\n// Operator affordances (the executor itself is handed the domain-plan VALUE directly):\n// `wildo context journey` — the eligibility board, runs in flight\n// `wildo context info specification-artifacts` — per-family landed state and rendered content\n\nThe executor **reads** `domain-plan` (handed to it, relationship drafts included) and the landed\n`business-model`; it **writes** `./wildo.saas.config.ts` and the materialized skeleton — app files,\ncommitted to git, which is where the record of what was decided lives.\n\nThis line used to name six families (#1434). Four of them reached nothing.\n\n## Step 1 — Derive and enable the engine capabilities\n\nEngine capabilities are the framework surfaces the app turns on (organizations, billing,\nnotifications, audit trails, and so on). **THE EXECUTOR DERIVES THEM. Nothing here chooses one.**\n\nExactly three are derivable today, and each from one declared value:\n\n| capability | derived from |\n|---|---|\n| `ORGANIZATIONS` | the domain plan's `tenancy` is `organizations` |\n| `EMAIL_TRANSACTIONAL` | the same — tenant invitations send mail |\n| `BILLING` | the business model's revenue model charges users |\n\n**The application's line can exclude a capability, and then nothing derives it.** An internal tool has no billing at\nall: on that line the `billing-provider` question is not asked, the executor never enables `BILLING` whatever else\nimplies it, and a configuration that enables it anyway is reported on the run result as contradicting the line\n(`engineCapabilitiesContradictingLine`), never switched off for you.\n\n**Compliance constraints do NOT select a capability, and this method said for months that they were\n\"the primary driver\".** They are not, and the reason is recorded where the derivation lives\n(`FOUNDATION_CAPABILITY_BASIS`, which is TOTAL over the capability vocabulary and names why each\nremaining one is underivable). The clearest case is the one this method used to point at: where an\naudit trail must LIVE — a ledger, a SIEM — is an operator decision about retention and independence\nrather than a product fact, and leaving that capability out removes no control objective, because\nassurance reads the application's own audit store.\n\nA capability nothing derives is not enabled by guessing. It is enabled deliberately, later, by\nwhoever decides it — which is why the basis map records a reason per capability instead of a\nquestion.\n\nHistorical note, kept because it is the defect rather than a detail (#1434): this section described a\nderivation the runtime cannot perform — four steps of \"walk the constraints\", \"treat these as\nsharpeners\", \"every enabled capability must trace to a compliance constraint\" — and a reader believed\ncompliance constraints shaped the foundation. Nothing walked anything. The market snapshot, the\nrequirements and the roadmap were declared inputs of this playbook for the same reason and reached\nthe same nobody; they are no longer declared.\n\nThe `config-saas` skill carries the vocabulary for the `engineCapabilities` block of\n`./wildo.saas.config.ts`. The executor writes that block; read the skill when you need to understand\nwhat an enabled capability means, not to decide one.\n\n## Step 1b — Apply the capability composition's optional packages\n\nWhen a `capability-composition` has landed (`bottom_up.capability-composition`), the executor\napplies every scenario it marks NEEDED after the domain scenarios below — the marketing website\n(`add-website`) and the documentation portal (`add-technical-doc`) today — with the application's\nown host facts: its slug and display name from `wildo.saas.config.ts`, the public domain from its\ninfrastructure environments (a remote one first, `localhost` when none declares one), and its port\nblock (website = app port + 1, docs = app port + 2, the convention the dogfood applications use).\nA package whose workspace already exists is left alone and reported as already composed. The run\nrecord names **every scenario applied and every scenario declined with the declaration each came\nfrom** (`compositionScenarios`), so a reader sees why the application has a website — or why not —\nwithout re-deriving it.\n\n## Step 1c — Lay down the line's access posture (internal tool)\n\nWho may hold an account is, for an internal tool, decided by what the application IS: the company's own people,\nadmitted by the company. So the executor, after declaring the organization type that owns the application, writes the\nline's posture through the same function the `wildo/internal-tool` blueprint applies at `wildo init --line`\n(`applyLineAccessPosture`), and reports it on the run result as `accessPosture`:\n\n| what | written | decided by |\n|---|---|---|\n| `userTypes.<each user type of the owner>.auth.registration.mode` in `./backend-api/src/saas-config.backend.ts` | `RegistrationMode.INVITATION_ONLY` | the line (not asked) |\n| organization creation | nothing to write: no door creates an organization of the owner type, and the configuration refuses a `creationPolicy` or `selfServiceCreation` on it | the line, through the owner it declares |\n| `[AuthMethod.ORG_OPENID_CONNECT]` and `[AuthMethod.ORG_SAML]` on those user types, and `CoreFeature.SINGLE_SIGN_ON` in the default feature profile | `true` / granted | the `corporate-identity` answer: `identity_provider` or `identity_provider_and_directory` |\n| `scimProvisioning` on the owner organization type, and `CoreFeature.DIRECTORY_PROVISIONING` | `{ enabled: true }` / granted | the `corporate-identity` answer: `identity_provider_and_directory` |\n\nOnly an internal tool is asked `corporate-identity`; every other line, and an application with no line yet, is not.\nWhen it is unanswered, the domain plan's owner type joining as `DIRECTORY_PROVISIONED` counts as\n`identity_provider_and_directory`; when the answer and the plan disagree, the answer is applied and the disagreement is\nlisted in `accessPosture.owed`. `none` writes nothing, and nothing already open is switched off. What the run cannot\nwrite (no owner type, user types it cannot read, a file shape it cannot address, a user type no owner lists) is listed\nthere too, with its remedy; nothing is guessed.\n\n`frontendServices.<app>.usersManagement.<user type>.register` stays as declared. It gates self-service sign-up at the\nfrontend service (password sign-up and social sign-up alike), which invitation-only registration already refuses at\nthe API, and the backend logs the overlap at startup; accepting an invitation does not read it.\n\nWhat this does NOT set up: the identity-provider connection, the verified domain and the SCIM token. Those carry\nsecrets or proofs and are records the company creates in the running application, through the product. The SHAPE of the\ncompany's identity set-up (its units, its directory's department mapping, its single sign-on posture) is a committed\ndeclaration, authored next.\n\n**The organization identity declaration.** Once the owner type is found and the corporate identity is known, the executor\nwrites `./backend-api/src/organization-identity.ts` (`APPLICATION_ORGANIZATION_IDENTITY`) and passes it as\n`organizationIdentity` in `./backend-api/src/bootstrap-config.ts`, before the host's overrides\n(`composeOwnerOrganizationIdentityDeclaration`). The backend applies it at every boot to the owner organization. It is\nreported on the run result as `organizationIdentity`:\n\n| part | written | decided by |\n|---|---|---|\n| `singleSignOn` | `requireSSO: true`, `requireEnterpriseSSOOnly: true`, `allowPasswordFallback: true`, no break-glass account | the `corporate-identity` answer: `identity_provider` or `identity_provider_and_directory`; `none` declares no posture |\n| `units` | empty | nothing typed names the company's units: OWED, never invented |\n| `directory` | not declared | each binding names a declared unit, so it is OWED with the units (`identity_provider_and_directory`) |\n| `name` | not declared | nothing typed names the company: OWED, never invented. Until it is added, every person reads the platform's placeholder \"Default Organization\"; the backend writes a declared name over that placeholder only: once the organization is named, a changed `name` is reported as drift (with the rename that adopts it), never written over the stored name |\n\nPasswords stay allowed until the company names an emergency administrator: refusing them with no break-glass account\nwould lock the company out the day its identity provider is unreachable. The owed list says to name one in\n`singleSignOn.breakGlassAccounts` and then set `allowPasswordFallback: false`. The file is rewritten only while it is\nexactly what a derivation wrote; once someone edits it (units added, an account named), it is kept and the run says so.\nAn unanswered `corporate-identity` writes nothing and is owed.\n\n**The owner type takes the domain plan's names, through a declared rename.** The platform seeds the owner\norganization for the first administrator, typed from the configuration, when the application is initialized. The\ncreation playbooks run on a started application, so by the time this step runs it usually has been, and an\norganization's type and a user's types are fixed at creation. So where the domain plan names the owner type or its\npeople differently (`group`, `employee` against the scaffold's `workspace`, `member`), the executor RENAMES them before\nlaying the posture down (`renameOwnerToDomainPlanNames`): the new keys carry `renamedFrom` naming the old ones, every\nreference follows (the first administrator's user types, the frontends' user management, relationship kinds), and the\nbackend re-types the seeded organization and every account still holding the old key at its next start, recording each\nat CRITICAL. The rename is all or nothing: one it cannot make (a spread key, a plan name already declared as another\ntype, several user types on either side) writes nothing and is listed in `accessPosture.owed`, for the build's\norganization-model task to finish. The build must never declare a second type with `ownsApplication`.\n\n## Step 1d — Report where records live outside the application\n\nWhen the domain plan's `organizationModel.recordLocations` puts a resource's records in a system of record (an ERP, a\nCRM), the run result lists each system under `systemsOfRecord`: its kind, the resources it holds, and the credentials its\ninstance will owe, read off the facet the `@wildo-ai/systems-of-record` library declares for that kind and named as the\nbackend reads them (`WILDO_SOR_<INSTANCE>__<SLOT>`). The instance is named by its kind (`systemsOfRecord.odoo`). Where\nthe facet offers several credential kinds, the result names them: which one the company uses is its choice. Report\nthose credentials as owed, like a provider's.\n\nThe executor declares neither the instance nor the library. The build's binding task declares the instance (vendor,\ndialect, partition, personal-data origin) against the facet; once it does, `wildo config sync` (and `wildo local dev`)\ndeclares the library on the backend and links it in one step, which is the only rule that writes both halves.\n\n## Step 2 — Declare the domain-plan's modules (register across runtimes after materialization)\n\nThe skeleton supplies the root workspaces and registry structure; declaring a\nmodule in configuration alone does not implement its package or contributions.\n\n1. **Declare each approved module at the SaaS root.** Use `config-saas` for the\n module's application configuration and preserve existing module identifiers.\n2. **Materialize and register module shells.** Use the delivered module-composition\n scenario and `module-registration` guidance to create the corresponding shared,\n backend and frontend contributions. Inspect the resulting registry exports;\n `config sync` is not a substitute for authored module implementation. At foundation\n depth the shells must load; business-specific resource behavior remains the build\n loop's work.\n3. **Declare every domain-plan module — and no more.** A module the domain plan named\n that is missing from the config is a GAP; a module in the config the domain plan did\n not name is an unjustified-module CONCERN. Record the declared modules (each traced\n to its domain-plan module) as a comment beside it in the config.\n\n**Do not author the resources here.** The domain plan's per-module resource inventory\ntells you *what* resources each module will own, but authoring their schemas, fields,\nand config is the **build loop's** job. If you catch yourself writing a resource\nschema, stop: that is out of this playbook's depth.\n\n## Step 3 — Select the providers\n\nEnabled capabilities that reach an external service need a **provider** (billing → a\npayment provider; notifications/email → an email provider; file storage → an object\nstore). For each such capability, select and wire the provider scope +\n`providerConfigurations` in `./wildo.saas.config.ts` (the `config-saas` skill covers\nthe provider surfaces). A value the vendor issues per deployment — an OAuth client ID, a\nfrontend SDK DSN — is authored in each environment's\n`./infrastructure/<env>/wildo.infra.<env>.config.ts > providerConfigurations` instead (the\n`infra-environment-config` skill):\n\n- **Every provider matches a declared capability.** A provider wired for no enabled\n capability is a CONCERN; an enabled capability that needs a provider with none\n selected is a GAP.\n- **Do not seed secrets here.** Provider *selection* (which provider, which scope) is\n foundation; the per-environment secret material is seeded by the deterministic\n bring-up / environment step, not authored inline. Record the selected providers\n (each matched to its capability) as a comment beside it in the config.\n\n**The choice itself is the creator's, and the interview carries it.** Billing and SMS are\nthe two capabilities whose every provider needs an account, so the executor cannot pick one:\nit reads the `billing-provider` and `sms-provider` answers from the dossier, declares the chosen\nprovider in the backend scope, selects it as primary and enables the capability. The provider's\nrequired credentials are recorded on the run result as OWED — the application boots, and the\nprovider answers once an operator sets them. Answering `none` — the interview requires an answer to\nevery question, so \"decide later\" is an option — leaves the capability owed exactly as before. The\noptions are the engine providers serving the capability plus `none`; the interview validator refuses\nanything else before the run starts. Email is the exception the executor decides itself: it has a\ncredential-free development substitute (`email-preview`) and needs no choice.\n\n## Step 4 — The compliance declarations are written by the executor (only what was told)\n\nNo agent writes these: the foundation executor writes them deterministically from the\ninterview, after the capabilities and the composition. Each answer follows a grammar the\ninterview holds it to at every door, so what reaches the executor is always readable.\n\n- **`operating-entity`** is labelled parts: `name: …`, `address: …`, `privacy contact: …`,\n one per line or separated by `;`. Each stated part the configuration lacks is written to\n `operatorComplianceIdentity.legalEntityName`, `.registeredAddress` or `.privacyContact`.\n Any subset is a valid answer: every part is independently declarable, and a missing one is\n reported downstream as outstanding (`wildo context launch-blockers`), never filled in.\n- **`operating-jurisdictions`** is `<code>: <why>` entries (`fr: established here; eu:\n customers resident here`). Each jurisdiction the configuration lacks is appended to\n `operatingJurisdictions.jurisdictions`, with its basis. A place without a basis is refused at\n the door: \"eu\" alone does not say which obligation applies.\n- **An authored value is never replaced.** A part or a jurisdiction the configuration already\n declares differently is kept, and the run result names it in `operatorDeclarations`. Nothing\n is removed, and nothing the interview does not ask (lawful bases, roles, hosting regions) is\n touched: those are legal or operational determinations nobody typed here.\n- **Nothing is invented.** The executor writes only text the creator gave. It never derives a\n company name, an address, a privacy contact or a lawful basis.\n- **Unanswered is an advisory, not a refusal.** A deferred or unanswered question writes\n nothing, and the run result says so. An answer stored before the grammar existed is reported\n as unreadable, with nothing written, until it is answered again.\n\nThis step CANNOT fail the run: a configuration the executor cannot read or edit is reported on\nthe run result with nothing written, and the application boots.\n\n### The data residency is written by the executor\n\nThe `data-residency-territories` answer is not prose: the interview stores it as jurisdiction\nslugs (`eu, ch`) or `none`, and the foundation executor writes it, deterministically.\n\n- **Every data-serving provider is placed first.** A provider the residency leaves exactly one\n region for (only offered vendor-hosted, one region whose places all lie inside) gets that\n region in `providerComplianceChoices`. An authored choice is never overwritten.\n- **The residency is written only when nothing is left unplaced**, into\n `operatorComplianceIdentity.dataResidency.territories`. `wildo config sync` refuses a\n declared residency that leaves a provider unplaced, so writing it earlier would stop the\n application starting.\n- **Anything left is named, never guessed.** A provider with several regions inside, none, a\n deployment model to choose, an authored region outside or no compliance profile is reported\n on the run result with the regions that would satisfy the residency, and\n `wildo context launch-blockers` lists it until the choice is made.\n- **The at-rest promise is the operator's, never asked and never written.** When the\n configuration already declares `dataResidency.storedAtRestTerritories`, the executor keeps it\n when it writes the territories, and places a provider none of whose regions processes inside\n on a region that keeps everything at rest inside it. When it declares none, an unplaced\n provider is reported with the regions an at-rest list over the same places would admit, so\n the operator can decide whether to make that weaker promise.\n- **`none` writes nothing**, and an authored residency is never removed by it: removing a\n residency withdraws a commitment, which a person does on purpose. A deferred answer writes\n nothing and the run says the residency is still open.\n\n(A `foundation-manifest` artifact was declared here until 2026-08-25 and is now\nretired with the app-creation verification/digest aspect. It was a bookkeeping record\nof decisions the config already carries — deleting it costs no fact about the app.)\n\n## Step 5 — Verify materialization and bring up the environment\n\nRun the existing commands from the intended application root with their documented\nprerequisites. Keep application structure, delivered knowledge and running services\nas separately verified results.\n\n1. **Verify the skeleton.** A fresh project starts with `wildo init` in an empty\n destination before the configuration-authoring steps above. Check `frontend/`,\n `shared-lib/`, `backend-api/`, `specifications/` and the generated configuration.\n On a repeat run, preserve existing source; use init's repair path or the relevant\n composition scenario instead of regenerating the project as a new application.\n\n2. **Verify application knowledge delivery.** Init attempts delivery from the\n installed packages and reports delivered or skipped assets. If dependency\n installation was skipped or a package is unavailable, install the required\n packages, then refresh the supported skills/rules and framework assets:\n\n ```bash\n wildo assets sync\n ```\n\n <!-- fwk-provenance-waiver: the command below is named in order to be REFUSED — `wildo fwk` needs a framework workspace and errors inside a generated application, which is exactly why an agent that might reach for it has to recognise the name; translating it away would leave the prohibition with nothing to prohibit. -->\n Read the result and inspect the delivered application guidance. Do not assume\n arbitrary agents/hooks are installed, and do not substitute a framework-developer\n delivery flow for this one — this command is how an application receives them.\n\n3. **Run setup** — configure the app from the authored `./wildo.saas.config.ts` and\n the `./infrastructure/` env configs (the `infra-environment-config` skill owns the\n typed infra env files; the `local-environment-bootstrap` skill owns the `wildo`\n lifecycle). Headless setup is registered and fail-closed on the administrator seed\n environment variables; it mints the environment generation without placeholders.\n\n ```bash\n wildo setup --non-interactive\n ```\n\n4. **Initialize and verify the project** — run the registered convergence journey.\n It regenerates artifacts, starts and readiness-gates backing services, reconciles\n platform Mongo state and credentials, starts native platform services, registers the\n app once, emits all runtime projections, and executes a signed managed-config read.\n\n ```bash\n wildo local init\n ```\n\n5. **Know what a PostgreSQL store owes — and that you do not write its baseline here.** Read the\n store selection you authored in the infra env config:\n\n ```bash\n rg -n \"applicationDatabase\" infrastructure/*/wildo.infra.*.config.ts\n ```\n\n Absent, or MongoDB: there is nothing to do — a collection exists on first write, there is no\n plan, and no migration is ever owed.\n\n `engine: DatabaseEngine.POSTGRESQL`: nothing creates a table implicitly, so until the INITIAL\n migration exists the backend refuses to boot on every table it declares. Do NOT accept a baseline\n with `wildo db diff --update-lockfile`: that writes the lockfile and NO migration, so it records a\n plan whose tables were never created and the boot gate still refuses.\n\n The build loop owns that migration. For an application that declares PostgreSQL and has no\n `backend-api/migrations/schema-plan.lock.json`, the dispatch host derives a `database-migrations`\n step; it runs `wildo db generate`, which, with no lockfile, renders the whole model as the initial\n migration and writes the baseline in the same invocation. The two are one change and are committed\n together. It needs a built `backend-api` (the plan is computed by importing the compiled backend\n registry), so report the store you found and leave the migration to that step.\n\nReport the status of each bring-up step to the creator from its real exit and\npostcondition results. A skipped knowledge delivery or unsuccessful initialization\nis not completion, even when source files exist.\n\n## Refine, don't re-author\n\n- **Keep what still holds.** On a re-run, do not disable a capability, drop a module,\n or re-select a provider that is already correct. Only enable a newly-justified\n capability, register a newly-added domain-plan module, or correct a provider where\n the definition set genuinely changed.\n- **Reuse existing module ids and provider scopes.** A module already registered keeps\n its id; a provider already selected keeps its scope — never mint a second one for the\n same module/capability.\n- **Follow the definition set.** If the domain plan added a module since the last run,\n reflect it (register the module) — do not re-derive the whole configuration from\n scratch. The delta is visible in `git diff` on the config, which is the record. A new\n compliance constraint does NOT enable a capability here; see Step 1 for why.\n\n## Quality gate — `foundation-configured-coherently` (declarative at scaffold depth)\n\nThe foundation is checked against the `foundation-configured-coherently` **reviewer\nchecklist** (see the end of this method): five invariants (capability derivation, module registration, provider coverage,\nsequence order, and foundation-depth boundary discipline), each marked PASS / CONCERN /\nGAP.\n\nThis checklist is **declarative**: no runtime executes it (rubric-only judge gates were retired at the flexible-intelligence hard cut — judgment runs only through a compiled Evaluation Treatment, and this playbook has none yet), so it ships as the **creator's review checklist** — nothing runs it automatically. The **trust floor** for accepting the foundation today is the\ncreator's `approve_result` review of the produced config and the bring-up result. There is\nintentionally **no deterministic gate**: config validation is a `wildo` CLI concern\n(`config sync` / validate), not a validator in the journey-artifact schemaRef\nvocabulary. Existing command permission and effect boundaries apply independently\nof this reviewer checklist.\n\n## Scaffold-depth boundary\n\nKeep this playbook at foundation depth. Explicitly **out of scope here**:\n\n- **Implementing replacement CLI tooling.** Use the existing init, asset delivery,\n setup and local initialization owners. If a command fails, retain its diagnostic\n evidence and resolve the prerequisite or concrete defect rather than inventing\n another materializer inside the application.\n- **Resource / field / schema authoring.** The config declares modules as shells and\n enables capabilities. The foundation executor can materialize resource shells\n and scope them to the chosen tenancy. Filling those shells with business-specific\n fields, rules, UI behavior, views and navigation belongs to the later build work.\n- **Re-introducing a foundation artifact.** The `foundation-manifest` family was\n retired on 2026-08-25 with the app-creation verification/digest aspect. Do **NOT**\n reshape `_core` to bring it back: the config file and its git history already carry\n every decision it recorded.\n- **Deep capability / provider modelling.** A defended capability-selection model\n (per-compliance-regime obligation → capability mapping, provider cost/feature\n matrices, multi-provider fallback) is a later depth pass. You derive capabilities\n from the definition set with the v1 heuristic above — not a solved obligation graph.\n\nIf you find yourself perfecting any of the above, stop and leave the slice at scaffold\ndepth — deeper capability modelling belongs to a dedicated later pass,\nholistically.\n\n## What you emit\n\n- The authored **`./wildo.saas.config.ts`** (an app file) — the enabled engine\n capabilities, the registered modules, and the selected providers — plus the\n registered module shells across the runtimes and the\n brought-up skeleton.\n- The **bring-up result** reported to the creator: the status of each step, taken from\n its real exit/postcondition rather than from intent. This is a report, not a stored\n artifact — nothing reads it back.\n\nThe creator reviews the result (approval posture `approve_result`) before it is\naccepted as the journey's foundation, which the bottom-up build loop then fills in\nwith authored resources, operations, UX, and integrations.\n\n## References\n\n- `.claude/skills/config-saas/SKILL.md` — the `./wildo.saas.config.ts` authoring\n surface: `engineCapabilities`, `modules`, provider scopes, and `providerConfigurations`.\n- `.claude/skills/module-registration/SKILL.md` — register each domain-plan module\n across the shared / frontend / backend registries so the skeleton boots with it\n (distinct from `config-saas`, which only declares module existence at the root).\n- `.claude/skills/infra-environment-config/SKILL.md` — the typed `./infrastructure/`\n environment configs (backing services, public domain, service auth) the bring-up needs.\n- `.claude/skills/local-environment-bootstrap/SKILL.md` — the `wildo` lifecycle\n (`setup`, `local init`, `local reinit`, `local dev`) for deterministic initialization,\n repair, clean-room recovery, and runtime start.\n\n## Reviewer checklist — `foundation-configured-coherently`\n\nFormerly authored as an `llm_judge` quality gate; the generic rubric-only gate kind was retired at the flexible-intelligence hard cut (judgment now runs only through a compiled Evaluation Treatment). The checklist itself is creator-facing review guidance and is maintained here as guidance against the current tooling:\n\nJudge whether the authored foundation (the ./wildo.saas.config.ts composition + the materialized skeleton) is a coherent realisation of the approved domain plan and the top-down definition set. Assess five invariants and mark each PASS / CONCERN / GAP with a one-line evidence note. (1) CAPABILITY DERIVATION — every enabled engine capability traces to a compliance/vision, business-model, market, or domain-plan need; unjustified capabilities are a CONCERN and missing required capabilities are a GAP. (2) MODULE REGISTRATION — every approved domain-plan module is declared and registered across runtimes; missing modules are a GAP and extra modules are a CONCERN. (3) PROVIDER COVERAGE — every provider serves a declared capability and every provider-backed capability has a provider. (4) SEQUENCE ORDER AND PROOF — materialize a fresh skeleton in an empty destination (or preserve an existing project) → refine config and module registrations → verify knowledge delivery → wildo setup --non-interactive → wildo local init. Treat setup intent, container health, or an initialization marker as insufficient: local init must reach its signed managed-config-read postcondition. (5) BOUNDARY DISCIPLINE — stay at foundation depth: deterministic resource shells are allowed, but business-specific fields, processing, UI behavior, views and navigation belong to later build work. Use the existing init/assets owners; do not replace them. Headless setup and local init run once their prerequisites are satisfied. This reviewer checklist is declarative; creator approval is the trust floor.\n",
50
50
  "skillRefs": [
51
51
  "config-saas",
52
52
  "module-registration",
@@ -73,12 +73,15 @@
73
73
  },
74
74
  {
75
75
  "ref": "operating-entity",
76
- "question": "Which legal entity operates this service? Give the parts you know, each on its own line (or separated by `;`) and labelled: `name: <registered company name>`, `address: <registered address>`, `privacy contact: <the email, URL or postal address a person uses to reach you about privacy>`. Answer only the parts you actually know today — anything you leave out is recorded as outstanding and nothing is blocked by it. Do NOT supply a placeholder or a plausible-looking value: a produced privacy notice quotes these as fact, and the address is what a person would use to serve a legal request on a real company.",
76
+ "question": "Which legal entity operates this service? Give the parts you know, each on its own line (or separated by `;`) and labelled: `name: <registered company name>`, `address: <registered address>`, `privacy contact: <the email, URL or postal address a person uses to reach you about privacy>`. Answer only the parts you actually know today — anything you leave out is recorded as outstanding and nothing is blocked by it. If you know none of them yet (no company exists), defer the question rather than answer it. Do NOT supply a placeholder or a plausible-looking value: a produced privacy notice quotes these as fact, and the address is what a person would use to serve a legal request on a real company.",
77
77
  "kind": "free_text",
78
78
  "answerFormat": "operator_identity_parts",
79
79
  "depth": "essential",
80
80
  "whenUnknown": "ask",
81
- "regulatedDomain": true
81
+ "regulatedDomain": true,
82
+ "nonAnswers": [
83
+ "deferred"
84
+ ]
82
85
  },
83
86
  {
84
87
  "ref": "operating-jurisdictions",
@@ -47,6 +47,15 @@ the task that reported it, and the follow-up task the build host derived to answ
47
47
  when its follow-up lands with nothing left to report, or when a landed later run states the
48
48
  requirement met.
49
49
 
50
+ A requirement the application's CONFIGURATION does not meet also reads at most `BUILT`. A requirement
51
+ declares what it cannot be met without (`engineCapabilities`, `languages`), and the reading compares
52
+ that with the capabilities `wildo.saas.config.ts` enables and the languages it serves. Each one that
53
+ is off is a configuration gap on the row, naming where to set it. A requirement that declares a need
54
+ while the configuration cannot be read is not called proven either.
55
+
56
+ The unattended drive calls the creation's goal reached only when no requirement carries an open item:
57
+ nothing addresses it, no build work for it landed, a reported gap, or a configuration gap.
58
+
50
59
  The ladder is ordered, and the stage is the highest one whose threshold the population meets. The
51
60
  default threshold is **every requirement at that level** — one number, and not a guess: a share
52
61
  below 1 would be an invented tolerance, and this family measures rather than flatters.
@@ -50,7 +50,7 @@
50
50
  ],
51
51
  "stepOutputs": [],
52
52
  "method": {
53
- "guidance": "# Maturity Reading — Where This Application Stands\n\n### Target Application Framing\n\nThis knowledge is for configuring behavior in the target application.\nWildo is the framework surface being used, not the product being built.\n\nRead and apply these patterns from the application-developer point of view:\n\n- how the application should configure the relevant app-level file\n- how the application should use framework-provided factories, enums, and helpers\n- how the application should avoid mixing frontend, backend, and infrastructure concerns\n\n\n## What this step does\n\nThe build lane ends on four typechecks and a migration-drift probe, and its own method says it \"does\nnot boot the application, run its test suite\". Nothing related what was BUILT to what the\napplication's requirements ASKED FOR.\n\nThis step records that relation as the `application-maturity` family: **one row per requirement the\napplication declares**, the evidence that exists for it, and the stage that follows.\n\n## There is no \"done\", and that is the design\n\nThe highest stage does not mean finished. It means that today, every requirement has the strongest\nevidence this framework can produce for it — and the next requirement somebody writes lowers the\nreading again, with no code change anywhere. That is what the measurement is FOR.\n\nSo nothing in the lane may gate on this family, and no other step may read a stage as permission.\n\n## It is derived, never judged\n\nNothing here is an opinion, and no model produces it. Each row's level is reached by a measurement:\n\n| level | reached when | measured from |\n|---|---|---|\n| `UNADDRESSED` | nothing the application decided points at the requirement | the decided-element graph |\n| `DECIDED` | a decided element links to it — a feature draft, an experience surface, a worker, a build task | inbound links on the same graph |\n| `BUILT` | a build task addressing it has LANDED | the landing probe, over the working tree |\n| `VERIFIED` | that landed task's declared product was PROVEN | the same probe, minus what it could not verify |\n| `OBSERVED` | something watched the running application do it | runtime observations |\n\nA requirement a coding run still REPORTS unmet reads at most `BUILT`, whatever the probe found. The\nrun that built the work said what is missing, so its file is not proof. The row lists each open gap,\nthe task that reported it, and the follow-up task the build host derived to answer it. A gap closes\nwhen its follow-up lands with nothing left to report, or when a landed later run states the\nrequirement met.\n\nThe ladder is ordered, and the stage is the highest one whose threshold the population meets. The\ndefault threshold is **every requirement at that level** — one number, and not a guess: a share\nbelow 1 would be an invented tolerance, and this family measures rather than flatters.\n\n## The denominator is the requirements family, and nothing else\n\nEvery row comes from `requirements`, functional and non-functional, in its own order. Not the\nrequirements that happen to be referenced somewhere, not a count of build tasks, not a share of\nfiles. A requirement nothing addresses is a ROW at `UNADDRESSED` — visible, counted, and lowering\nthe reading.\n\nThat is the difference between a measurement and a file count: a denominator you choose yourself is\nnot a measurement at all.\n\n## A family that addresses requirements joins on its own\n\n\"Which elements address this requirement\" is read from the decided-element graph's inbound links, so\nevery family that declares an edge to `requirements.functional` or `requirements.non-functional`\nalready contributes — and one that declares such an edge tomorrow contributes without this method,\nor the derivation, knowing it exists.\n\n## What the reading says that a number cannot\n\nFour findings, and each names the application rather than the artifact:\n\n- a requirement **nothing addresses** — the reading's whole point;\n- a requirement whose build work **landed and could not be proven**, so its evidence stops at\n `BUILT`. A task that declares no product cannot be probed, and that is a known, deliberate state\n rather than a failure;\n- a requirement a coding run still **reports partly unmet**, naming what is missing and the\n follow-up that answers it. It stops at `BUILT` until that gap closes;\n- **no runtime observation exists at all**, which today is true of every application. The finding\n names what would fill it — an engine-behaviour e2e scenario that names the requirement it\n exercises — rather than leaving the top of the ladder quietly empty.\n\n## What you emit\n\n- The **maturity reading** at `specifications/src/application-maturity/index.ts`. It is landed by\n the runner, not written by hand.\n\n## Verification\n\n- The row count equals the number of requirements the application declares.\n- Every `requirementRef` names a requirement in the `requirements` family.\n- `rg -c \"UNADDRESSED\" specifications/src/application-maturity/index.ts` — how many requirements\n nothing has picked up yet.\n",
53
+ "guidance": "# Maturity Reading — Where This Application Stands\n\n### Target Application Framing\n\nThis knowledge is for configuring behavior in the target application.\nWildo is the framework surface being used, not the product being built.\n\nRead and apply these patterns from the application-developer point of view:\n\n- how the application should configure the relevant app-level file\n- how the application should use framework-provided factories, enums, and helpers\n- how the application should avoid mixing frontend, backend, and infrastructure concerns\n\n\n## What this step does\n\nThe build lane ends on four typechecks and a migration-drift probe, and its own method says it \"does\nnot boot the application, run its test suite\". Nothing related what was BUILT to what the\napplication's requirements ASKED FOR.\n\nThis step records that relation as the `application-maturity` family: **one row per requirement the\napplication declares**, the evidence that exists for it, and the stage that follows.\n\n## There is no \"done\", and that is the design\n\nThe highest stage does not mean finished. It means that today, every requirement has the strongest\nevidence this framework can produce for it — and the next requirement somebody writes lowers the\nreading again, with no code change anywhere. That is what the measurement is FOR.\n\nSo nothing in the lane may gate on this family, and no other step may read a stage as permission.\n\n## It is derived, never judged\n\nNothing here is an opinion, and no model produces it. Each row's level is reached by a measurement:\n\n| level | reached when | measured from |\n|---|---|---|\n| `UNADDRESSED` | nothing the application decided points at the requirement | the decided-element graph |\n| `DECIDED` | a decided element links to it — a feature draft, an experience surface, a worker, a build task | inbound links on the same graph |\n| `BUILT` | a build task addressing it has LANDED | the landing probe, over the working tree |\n| `VERIFIED` | that landed task's declared product was PROVEN | the same probe, minus what it could not verify |\n| `OBSERVED` | something watched the running application do it | runtime observations |\n\nA requirement a coding run still REPORTS unmet reads at most `BUILT`, whatever the probe found. The\nrun that built the work said what is missing, so its file is not proof. The row lists each open gap,\nthe task that reported it, and the follow-up task the build host derived to answer it. A gap closes\nwhen its follow-up lands with nothing left to report, or when a landed later run states the\nrequirement met.\n\nA requirement the application's CONFIGURATION does not meet also reads at most `BUILT`. A requirement\ndeclares what it cannot be met without (`engineCapabilities`, `languages`), and the reading compares\nthat with the capabilities `wildo.saas.config.ts` enables and the languages it serves. Each one that\nis off is a configuration gap on the row, naming where to set it. A requirement that declares a need\nwhile the configuration cannot be read is not called proven either.\n\nThe unattended drive calls the creation's goal reached only when no requirement carries an open item:\nnothing addresses it, no build work for it landed, a reported gap, or a configuration gap.\n\nThe ladder is ordered, and the stage is the highest one whose threshold the population meets. The\ndefault threshold is **every requirement at that level** — one number, and not a guess: a share\nbelow 1 would be an invented tolerance, and this family measures rather than flatters.\n\n## The denominator is the requirements family, and nothing else\n\nEvery row comes from `requirements`, functional and non-functional, in its own order. Not the\nrequirements that happen to be referenced somewhere, not a count of build tasks, not a share of\nfiles. A requirement nothing addresses is a ROW at `UNADDRESSED` — visible, counted, and lowering\nthe reading.\n\nThat is the difference between a measurement and a file count: a denominator you choose yourself is\nnot a measurement at all.\n\n## A family that addresses requirements joins on its own\n\n\"Which elements address this requirement\" is read from the decided-element graph's inbound links, so\nevery family that declares an edge to `requirements.functional` or `requirements.non-functional`\nalready contributes — and one that declares such an edge tomorrow contributes without this method,\nor the derivation, knowing it exists.\n\n## What the reading says that a number cannot\n\nFour findings, and each names the application rather than the artifact:\n\n- a requirement **nothing addresses** — the reading's whole point;\n- a requirement whose build work **landed and could not be proven**, so its evidence stops at\n `BUILT`. A task that declares no product cannot be probed, and that is a known, deliberate state\n rather than a failure;\n- a requirement a coding run still **reports partly unmet**, naming what is missing and the\n follow-up that answers it. It stops at `BUILT` until that gap closes;\n- **no runtime observation exists at all**, which today is true of every application. The finding\n names what would fill it — an engine-behaviour e2e scenario that names the requirement it\n exercises — rather than leaving the top of the ladder quietly empty.\n\n## What you emit\n\n- The **maturity reading** at `specifications/src/application-maturity/index.ts`. It is landed by\n the runner, not written by hand.\n\n## Verification\n\n- The row count equals the number of requirements the application declares.\n- Every `requirementRef` names a requirement in the `requirements` family.\n- `rg -c \"UNADDRESSED\" specifications/src/application-maturity/index.ts` — how many requirements\n nothing has picked up yet.\n",
54
54
  "skillRefs": [],
55
55
  "skillBodyRefs": [],
56
56
  "derivedFromAnswers": [],
@@ -775,8 +775,17 @@ region, the places that region processes in, and a `status` (`inside`, `outside`
775
775
  `no-established-profile`). State the territories from this value, never from prose. `null` means no
776
776
  residency is declared: say the service makes no residency commitment rather than naming a region.
777
777
  `providers: null` means the providers' profiles were not read, so state the commitment without claiming
778
- every vendor keeps it. A configuration with any status other than `inside` is refused by `wildo config
779
- sync`, so a landed application normally shows only `inside`.
778
+ every vendor keeps it. A configuration with any status other than `inside` or `inside-at-rest` is
779
+ refused by `wildo config sync`, so a landed application normally shows only those two.
780
+
781
+ `storedAtRestTerritories` is a second, weaker promise the operator may opt into: a vendor whose region
782
+ keeps the data in those places and processes it elsewhere. Such a vendor has status `inside-at-rest` and
783
+ carries `storageTerritories`. Name it and say exactly that: the data is stored in those places, and it is
784
+ processed elsewhere. Never write that it processes inside the residency, and never fold it into a sentence
785
+ about the vendors that do. `storedAtRestTerritories: null` means the operator made no such promise.
786
+ When the entry carries a `storageCondition`, the storage holds only on that account arrangement (an add-on, a
787
+ plan, a subuser, data created after residency was set up): state it beside the places, in the vendor's terms,
788
+ so the reader sees what the promise rests on.
780
789
 
781
790
  **Completeness** — do not round any state up.
782
791