Exit Criteria Signoff — V1 Launch Gate

Gate owner: Release Captain + every domain owner named in §10 below. Required for §33.4. Each launch consumes one signed instance at docs/releases/v1/verification/exit-criteria/YYYY-MM-DD.md.

V1 is not feature-complete until every exit criterion below has a PASS or PASS-WITH-CONDITIONS signoff with evidence committed to the repo. Lacking that, GA cannot proceed regardless of other gates.

Release scope (2026-08-05). This gate is scoped to what the release under signoff actually ships. For V1.0 that is the four rooms — Tara, Nyx, Arete, Nisaba — on web + PWA, charging through the crypto rail only. Veritas and Metis are deferred to V1.2 and the native apps and fiat rails to V1.1 (see V1/BRAND.md).

Read "every domain" below as "every domain this release ships". A criterion demanding coverage of a deferred room would either block a launch on a room nobody can open, or be signed off against evidence that proves nothing — both worse than an honest, narrower gate. The deferred rooms return to this gate in the release that opens them.

1. Customer web feels like one coherent product across every room this release ships#

Evidence:

• Customer web app: apps/oshun/web/ (Next.js).
• Customer mobile app: apps/oshun/mobile/ (React Native + native shells).
• Shared design system: apps/oshun/web/src/design-system/lilith + cross-platform tokens distributed to mobile via the shared style package.
• Cross-domain shell tests:
  • apps/oshun/web/e2e/critical-web-journey-coverage.spec.ts (matrix of the shipped domains with keyboard / a11y / visual / resilience coverage references).
  • apps/oshun/web/e2e/cross-domain-flows.spec.ts.
  • apps/oshun/mobile/e2e/flows/shell/deep-links-six-domains.yaml — mobile evidence, and therefore V1.1 signoff rather than V1.0.
• Design-language audits: apps/oshun/web/e2e/domain-readability.spec.ts, apps/oshun/web/e2e/home-domain-narrative-order.spec.ts.

Verification owner: Product Lead + Design Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

2. Tara is visibly the product center#

Evidence:

• Home composition: apps/oshun/web/src/components/lilith/CustomerToday.tsx is documented as "Lead column: today's Tara sit" — Tara owns the lead column on the broadsheet front page.
• Home route: apps/oshun/web/src/app/page.tsx renders CustomerToday.
• Continuity tests:
  • apps/oshun/web/e2e/home-tara-centering-continuity.spec.ts.
  • apps/oshun/web/e2e/home-tara-practice.spec.ts (where present) and home-domain-narrative-order.spec.ts.

Verification owner: Tara PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

3. Nisaba and Metis are fully integrated, not just routed#

Evidence:

• Cross-domain spec: apps/oshun/web/e2e/nisaba-metis-study-prompts.spec.ts.
• Arete bridge: apps/oshun/web/e2e/arete-metis-study-prompts.spec.ts.
• BFF cross-domain recommendations: apps/oshun/bff/src/__tests__/cross-domain-recommendations-route.test.ts.
• Search ranking cross-domain integration: apps/oshun/bff/src/search/ranking.ts (Nisaba + Metis signals).
• Continue-route surface: apps/oshun/bff/src/__tests__/continue-route.test.ts.
• Library and activity-route integration: apps/oshun/bff/src/__tests__/library-route.test.ts, activity-route.test.ts.

Verification owner: Nisaba PM + Metis PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

4. Sophia grounding is visible everywhere evidence matters#

Evidence:

• Sophia core: apps/sophia/, libs/sophia/.
• Sophia store integration into Studio: apps/yemaya/studio-web/src/stores/sophiaStore.ts.
• Grounding visibility in Veritas: apps/oshun/web/e2e/veritas-full-journey.spec.ts (source drill + counterclaim).
• Assistant grounding-state surface: apps/oshun/web/e2e/assistant-grounding-state.spec.ts.
• Metis evidence emission: libs/metis/models/src/assessment-evidence/assessment-evidence-pack.ts and .test.ts.
• Iris evals for grounded-answer rate.

Verification owner: Sophia PM + Iris PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

5. Cross-domain continuity is useful, credible, and tested#

Evidence:

• Integration test harness: testing/integration/examples/cross-domain.integration.test.ts.
• Cross-domain flows: apps/oshun/web/e2e/cross-domain-flows.spec.ts, apps/oshun/web/e2e/activity-cross-domain-continuity.spec.ts, apps/oshun/web/e2e/home-mixed-domain-recommendations.spec.ts.
• Continuity card surfaces: apps/oshun/web/e2e/continuity-cards.spec.ts.
• BFF continuity routes: apps/oshun/bff/src/__tests__/continue-route.test.ts.

Verification owner: Shell PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

6. Admin web + admin mobile operate every launch-critical workflow#

Evidence:

• Admin web: apps/oshun/admin/.
• Admin mobile: apps/oshun/admin-mobile/.
• Admin E2E coverage (catalog in apps/oshun/admin/e2e/, 40+ specs):
  • Queue triage: unified-inbox-flows.spec.ts, inbox-bulk-action-flows.spec.ts.
  • Review/approve: review-decision-flows.spec.ts, high-risk-approval-flows.spec.ts.
  • Publish gate: policy-approval-promotion-rollback-flows.spec.ts.
  • Audit trail: audit-log-explorer-flows.spec.ts.
  • Persona promotion: persona-approval-flows.spec.ts, persona-governance-editor.spec.ts, persona-release-flows.spec.ts.
  • Model promotion: model-governance-flows.spec.ts, model-promotion-flows.spec.ts.
  • Privacy DSAR: privacy-dsar-end-to-end.spec.ts, privacy-deletion-workflow-flows.spec.ts.
  • Incidents: incident-orchestration-flows.spec.ts, incident-impact-comms-flows.spec.ts, incident-runbook-postmortem-flows.spec.ts.

Verification owner: Operations Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

7. Oshun Studio supports authoring / editorial / asset / taxonomy / versioning / localization / collaboration / customer-curation with grounded source binding, persona/tone policy, and release-gate enforcement#

Evidence:

• Authoring + editorial: apps/oshun/web/e2e/studio-authoring-workspace-smoke.spec.ts.
• Asset preview: apps/oshun/web/e2e/studio-asset-preview-pipeline-smoke.spec.ts.
• Taxonomy: libs/oshun/analytics/src/admin-taxonomy.ts, customer-taxonomy.ts, assistant-taxonomy.ts.
• Versioning: apps/oshun/web/e2e/studio-hathor-entity-version-management-smoke.spec.ts (representative per-domain version manager; coverage matrix in apps/oshun/web/e2e/support/critical-web-journeys.ts).
• Localization: route surfaced at apps/oshun/web/src/app/studio/internationalization-localization/page.tsx; glossary editor at libs/oshun/studio-authoring/src/localization-workflow/glossary-editor.ts; workflow runbook docs catalogued under docs/releases/v1/runbooks/ (Studio i18n/L10n suite).
• Collaboration: apps/oshun/web/e2e/studio-realtime-collaboration-substrate-smoke.spec.ts.
• Source binding: apps/yemaya/studio-web/src/stores/sophiaStore.ts + apps/yemaya/studio-web/e2e/studio-web.smoke.spec.ts.
• Persona/tone policy: libs/oshun/persona-policy-lilith/ (tone bands + cross-product harness), libs/oshun/persona-registry/ (calibration rubric), apps/oshun/admin/e2e/persona-governance-editor.spec.ts.
• Release-gate enforcement: libs/oshun/generation-control-isis/src/release-gate-model.ts + release-gate-model.test.ts, libs/isis/3d-quality-gates/src/delivery/release-gates.ts, apps/oshun/web/e2e/studio-launch-readiness-governance-smoke.spec.ts.

Verification owner: Studio PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

8. Agentic AI Studio operational with governance gates enforced and tenant-scoped budgets honored#

Evidence:

• Agentic store: apps/oshun/bff/src/admin/admin-agentic-operations-store.ts.
• Agentic panel UI: apps/oshun/admin/src/components/AgenticOperationsPanel.tsx (run table with data-run-id, data-run-severity, data-run-approval, cost ledger, kill switches).
• Tenant-scoped budgets: libs/oshun/tenant-console/src/tenant-model/agentic-config.ts.
• Approval gates: apps/oshun/admin/e2e/high-risk-approval-flows.spec.ts.
• Agentic ops flows: apps/oshun/admin/e2e/agentic-operations-flows.spec.ts, copilot-lifecycle-flows.spec.ts, copilot-audit-invariants.spec.ts.
• Library: libs/oshun/agentic-studio test suite.

Verification owner: Agentic AI PM + Operations Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

9. Tenant Console operational with SSO/SCIM/OneRoster/role-customization/audit/bulk-ops/API-keys/webhooks/lifecycle-comms/help-authoring/status-banner — strictly isolated from operator Oshun Admin#

Evidence:

• Tenant-admin app (customer-facing): apps/oshun/tenant-admin/.
• Identity:
  • SSO: libs/oshun/tenant-console/src/identity/sso.ts + sso.spec.ts (added in §33.3 remediation).
  • SCIM + OneRoster diff: libs/oshun/tenant-console/src/identity/scim.ts + scim.spec.ts (added in §33.3 remediation).
  • Federation: libs/oshun/tenant-console/src/identity/federation-and-auth-policy.ts.
• Roles: libs/oshun/tenant-console/src/roles/.
• Audit: libs/oshun/tenant-console/src/audit-explorer/.
• Bulk ops + OneRoster export: libs/oshun/tenant-console/src/bulk-ops/bulk-export.ts + bulk-export.test.ts.
• API keys + webhooks: libs/oshun/tenant-console/src/integrations/api-keys.ts, webhooks.ts, webhook-simulator.ts.
• Lifecycle comms: libs/oshun/tenant-console/src/notifications/.
• Help authoring: libs/oshun/tenant-console/src/help-center/.
• Status banner: libs/oshun/tenant-console/src/status-page/ + status-page-extras.test.ts.
• Tenant isolation: tests/security/tenant-isolation/bff-tenant-isolation.security.test.ts.

Strict isolation from operator Admin — separate app surfaces: apps/oshun/admin/ (operators) and apps/oshun/tenant-admin/ (customers). Tenant-isolation tests enforce the data-plane separation.

Verification owner: Tenant Console PM + Security Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

10. Iris / Psyche / Lilith / Sophia / Isis / Metis ownership boundaries true in implementation, not just in docs#

Evidence:

• Domain matrix: DOMAINS/ with architecture.md, features.md, specifications.md per domain.
• Domain ownership matrix: DOMAINS/README.md (assignments).
• Apps + libs partitioned by domain — verified by nx project graph and the tsconfig.base.json path mappings (each domain owns its own scope @iris/*, @psyche/*, @lilith/*, @sophia/*, @isis/*, @metis/*).
• Cross-domain integration is via published contracts only: libs/contracts/ + per-domain contracts.

Domain owners (responsible for confirming boundaries on signoff):

• Iris — Iris PM
• Psyche — Psyche PM
• Lilith — Lilith PM
• Sophia — Sophia PM
• Isis — Isis PM
• Metis — Metis PM

Decision (per-domain): PASS / PASS-WITH-CONDITIONS / HOLD.

11. External model sources cannot bypass review / rights / safety / provenance / Isis release gates#

Evidence:

• Isis release gates:
  • libs/oshun/generation-control-isis/src/release-gate-model.ts + tests.
  • libs/isis/ai-providers/src/release-gates/ + tests.
  • libs/isis/3d-quality-gates/src/delivery/release-gates.ts + tests.
• Provenance tracking: libs/sophia/ingestion/src/provenance/ + libs/isis/ai-providers/src/provenance/ + libs/isis/outputs/src/provenance/
  • per-asset provenance audit (sampling 100 items per domain).
• Watermark verification: docs/runbooks/watermark-verification-failure.md drill; verifier code under libs/aphrodite/stream-watermarking/, libs/aja/content-watermarking/, and libs/yemaya/living-scenes-runtime/src/provenance/audio-watermark/.
• Safety gate: every external-provider call routes through libs/iris/privacy/safety/ (privacy-class enforcement) and libs/iris/agents/computer-use/safety/ (agent execution safety) before publication.
• Rights gate: libs/isis/ai-providers/src/release-gates/ + Sophia source-credibility scoring (libs/veritas/core/src/source.ts).

Verification owner: Isis PM + Iris PM + Security Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

12. Metis high-stakes outputs emit claim-level evidence and pass grounding / pedagogy / safety / rights / academic-integrity / standards / drift gates#

Evidence:

• Claim-level evidence emission: libs/metis/models/src/assessment-evidence/assessment-evidence-pack.ts + .test.ts.
• Evidence requirements: libs/metis/learning/src/mastery/evidence-requirements.ts.
• Grounding eval: docs/launch/quality-review.md §1.8 (grounded-rate ≥ 0.95).
• Pedagogy eval: docs/launch/quality-review.md §1.8 (rubric pass rate ≥ 0.90).
• Safety eval: same §1.8 (crisis-flow 1.00).
• Rights eval: same §1.8 (violation rate ≤ 0.001).
• Academic integrity eval: same §1.8 (violation rate ≤ 0.001).
• Standards eval: same §1.8 (alignment ≥ 0.95).
• Drift eval: same §1.8 (KL ≤ 0.05).
• E2E: apps/metis/web/e2e/critical-paths.spec.ts, assessments.spec.ts, apps/metis/admin/e2e/admin-critical-flows.spec.ts.

Verification owner: Metis PM + Iris PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

13. Public website + app store listings + deep links + public status page ready for launch#

Evidence:

• Public surfaces: apps/oshun/web/public/ (favicon, icons, manifest, robots, screenshots, service worker).
• Deep-link associations: apps/oshun/web/src/lib/deep-link-associations.ts + apps/oshun/web/src/__tests__/deep-link-association-files.test.ts.
• Canonical deep links: apps/oshun/web/e2e/canonical-deep-links.spec.ts, canonical-deep-links-metis.spec.ts, apps/oshun/mobile/e2e/flows/shell/deep-links-six-domains.yaml.
• Lilith public routes: apps/oshun/web/e2e/lilith-public-routes.spec.ts.
• PWA install + update + offline: apps/oshun/web/e2e/pwa-install-update-offline.spec.ts.
• Status page: libs/oshun/tenant-console/src/status-page/ + tests.

Verification owner: Marketing Lead + Mobile PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

14. Recommendation candidate generation, signal ingestion, offline evaluation, observability operational#

Evidence:

• BFF recommendation routes: apps/oshun/bff/src/recommendations/.
• Cross-domain recommendation telemetry: apps/oshun/bff/src/__tests__/cross-domain-recommendation-telemetry.test.ts.
• Search ranking and candidate generation: apps/oshun/bff/src/search/ranking.ts.
• Experimentation: libs/oshun/search-discovery/src/experiments/.
• Observability: libs/oshun/analytics/src/dashboards-*.ts, alerts-manifest.ts, tracing-manifest.ts.

Verification owner: Search/Discovery PM + SRE Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

15. Aja embodied-instruction services required by Metis are ready or explicitly deferred with approved scope boundaries#

Evidence:

• Aja services: apps/aja/svc-motion-ai/, apps/aja/svc-motion-pipeline/, apps/aja/svc-reference-video/.
• Aja libraries: libs/aja/ (40+ sub-libraries including motion-formats, motion-pipeline-sdk, content-moderation, content-security, privacy-protection, motion-validation, etc.).
• Aja CLI + docs: apps/aja/cli/, apps/aja/docs/.
• Scope boundaries: documented in DOMAINS/aja/ (architecture + features).

Verification owner: Aja PM + Metis PM.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

16. All critical workflows have risk-tiered automated verification#

Evidence:

• Automation matrix: docs/testing/automation-matrix.md.
• Critical-journey mapping: docs/launch/critical-journeys.md.
• Risk-tier evidence per workflow:
  • High-risk (publishing, model promotion, persona release, kill-switch, DSAR, tenant isolation): require Playwright E2E + library unit + integration tests + chaos drill.
  • Medium-risk (review queues, billing, content edits, cross-domain continuity): require Playwright E2E + library unit tests.
  • Low-risk (cosmetic, UX polish): require Vitest unit + visual regression.
• Verification: libs/yemaya/production-verification/, testing/integration/contract-verification.test.ts.

Verification owner: QA Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

17. Accessibility, performance, security, resilience, observability, documentation, training, runbooks, canaries, rollback triggers, post-deploy monitors all satisfy launch gates#

Evidence:

Verification owner: SRE Lead + QA Lead + Security Lead.

Decision: PASS / PASS-WITH-CONDITIONS / HOLD.

18. Signoff capture#

# Exit-criteria signoff
date: YYYY-MM-DD
release tag: vX.Y.Z

| Criterion | Owner | Decision | Conditions | TS |
| --- | --- | --- | --- | --- |
| 1. Web/mobile coherence | Product Lead + Design Lead | PASS / HOLD | ... | ts |
| 2. Tara product center | Tara PM | PASS / HOLD | ... | ts |
| 3. Nisaba+Metis integrated | Nisaba PM + Metis PM | PASS / HOLD | ... | ts |
| 4. Sophia grounding visible | Sophia PM + Iris PM | PASS / HOLD | ... | ts |
| 5. Cross-domain continuity | Shell PM | PASS / HOLD | ... | ts |
| 6. Admin web+mobile critical workflows | Operations Lead | PASS / HOLD | ... | ts |
| 7. Oshun Studio | Studio PM | PASS / HOLD | ... | ts |
| 8. Agentic AI Studio | Agentic AI PM + Ops Lead | PASS / HOLD | ... | ts |
| 9. Tenant Console (isolated) | Tenant Console PM + Security Lead | PASS / HOLD | ... | ts |
| 10. Domain ownership boundaries | Per-domain PM × 6 | PASS / HOLD | ... | ts |
| 11. External-model gates not bypassable | Isis PM + Iris PM + Security Lead | PASS / HOLD | ... | ts |
| 12. Metis high-stakes evidence | Metis PM + Iris PM | PASS / HOLD | ... | ts |
| 13. Public/app-store/deep-links/status | Marketing Lead + Mobile PM | PASS / HOLD | ... | ts |
| 14. Recommendations operational | Search PM + SRE Lead | PASS / HOLD | ... | ts |
| 15. Aja services ready/deferred | Aja PM + Metis PM | PASS / HOLD | ... | ts |
| 16. Risk-tiered automated verification | QA Lead | PASS / HOLD | ... | ts |
| 17. A11y/perf/sec/resilience/obs/etc. | SRE Lead + QA Lead + Security Lead | PASS / HOLD | ... | ts |

aggregate decision: GO / NO-GO

A single HOLD on any criterion blocks GA until the criterion converts to PASS or PASS-WITH-CONDITIONS with linked conditions tracked.

19. Re-opening conditions#

Any of these auto-reopens the signoff:

• A criterion's referenced evidence file is deleted, broken, or regresses in a CI run between signoff and GA.
• A criterion's PM rescinds their PASS.
• A new beta-blocking issue is opened that maps to any criterion.

20. Cross-references#

• Go/no-go: docs/launch/go-no-go.md.
• Critical journeys: docs/launch/critical-journeys.md.
• Quality review: docs/launch/quality-review.md.
• Safety review: docs/launch/safety-review.md.
• Dogfood drill: docs/launch/dogfood-drill.md.
• Private beta: docs/launch/private-beta.md.
• Beta-blocking remediation: docs/launch/beta-blocking-remediation.md.
• Rollout plan: docs/launch/rollout-plan.md.
• Rollback plan: docs/launch/rollback-plan.md.
• Post-deploy monitoring: docs/launch/post-deploy-monitoring.md.
• Canary analysis: docs/launch/canary-analysis.md.