# Ecosystem Verification Matrix > **Status:** Hardened baseline > **Owner:** `learning_ai_common_plat` > **Purpose:** Define how ecosystem work is verified before it is considered complete. --- ## 1. Principle A cross-product feature is not done when one repo compiles. It is done when: 1. the shared contract is updated 2. the producing repo emits the expected behavior 3. the consuming repo handles it correctly 4. the tracker reflects the implementing commit(s) --- ## 2. Verification Layers ### Layer 1: Doc verification - contract doc updated - checklist updated - tracker updated - open decisions called out if unresolved ### Layer 2: Contract verification - schema/type examples compile or validate - sample payloads are present - event names and field names are stable ### Layer 3: Producer verification - source repo emits or stores the expected contract shape - unit/integration tests added where appropriate ### Layer 4: Consumer verification - downstream repo accepts and uses the contract correctly - no translation hacks or undocumented fallback behavior ### Layer 5: End-to-end flow verification - the actual user journey works across repo boundaries --- ## 3. Phase-1 Golden Flow Checks ### Transcript -> Note -> Memory - [ ] transcript artifact created - [ ] note artifact created or linked - [ ] memory artifact created or proposed - [ ] canonical events emitted - [ ] provenance preserved across all steps - [ ] tracker updated with commit links ### Plan -> Routine -> Habit - [ ] plan artifact exists - [ ] routine generated from plan - [ ] habit generated from plan - [ ] canonical events emitted - [ ] user-visible flow validated - [ ] tracker updated with commit links ### Cowork -> Trail -> Notes -> Memory - [ ] agent run logged - [ ] output artifact generated - [ ] ActionTrail can replay or inspect the run - [ ] downstream note or memory artifacts linked - [ ] approval trail preserved where required - [ ] tracker updated with commit links --- ## 4. Required Completion Evidence For every major implementation item, attach: 1. commit hash(es) 2. affected repos 3. verification commands or tests 4. residual risks 5. tracker update --- ## 5. Repo-Level Verification Owners - `learning_ai_common_plat` - shared schema validation in `packages/events` - platform-service route tests - cowork-service route tests - admin-web proxy and dashboard route tests - `learning_voice_ai_agent` - transcript producer tests and artifact persistence checks - `learning_ai_notes` - transcript and trail import route/library tests - `learning_multimodal_memory_agents` - memory import route/library tests - `learning_ai_flowmonk` - runtime route tests and Phase 2 artifact flow tests - `learning_ai_clock` - routine import tests for Phase 2 - `learning_ai_efforise` - habit import tests for Phase 2 - `learning_ai_trails` - trail artifact import and causation tests for Phase 3 - `oss/learning_ai_claw-cowork` - native audit/runtime emission verification where local Rust tooling is available --- ## 6. Repo-Level Evidence Matrix | Repo | Primary owner | Required evidence | Baseline commands | | ----------------------------------- | ----------------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | `learning_ai_common_plat` | Codex | contract changes, tracker updates, route tests, shared type validation | `pnpm --filter @bytelyst/events build`, `pnpm --filter @lysnrai/platform-service exec vitest run ...`, `pnpm --filter @lysnrai/cowork-service exec vitest run ...` | | `learning_voice_ai_agent` | Codex | transcript producer verification, artifact persistence checks | `python3 -m pytest tests/test_ecosystem_phase1.py -q` | | `learning_ai_notes` | Windsurf or Codex | import route verification, schema conformance, downstream lineage | `pnpm --filter @notelett/backend run test` or focused Vitest route coverage | | `learning_multimodal_memory_agents` | Cursor or Codex | memory import verification, runtime path validation, focused web tests | focused Vitest command for the ecosystem import path | | `learning_ai_flowmonk` | Codex | backend runtime route tests, web build, live runtime browser verification when runtime UI changes | `pnpm --filter @flowmonk/backend exec vitest run src/modules/agent-runtime/routes.test.ts`, `npm run build`, `npx playwright test ...` | | `learning_ai_clock` | Codex | routine-import verification against the shared artifact contract | focused route or library tests for routine import | | `learning_ai_efforise` | Codex | habit-import verification against the shared artifact contract | focused route or library tests for habit import | | `learning_ai_trails` | Codex | trail artifact causation verification and import-path checks | `node --experimental-strip-types --test backend/src/lib/ecosystem-phase3.test.ts` | | `oss/learning_ai_claw-cowork` | Codex | native Rust verification for audit, session/task projections, and checkpoint surfaces | `cargo test --workspace` | Rules: 1. Every tracker row that claims completion should cite at least one repo-specific verification command. 2. Cross-repo flows should include both producer and consumer evidence. 3. Live browser or runtime walkthrough evidence is required when user-facing behavior is a first-class requirement, not just a convenience. --- ## 7. Implementation Checklist - [x] finalize phase-1 golden flow checks - [x] define repo-level verification owners - [x] define minimum evidence for “done” - [x] define repo-level evidence ownership and baseline command expectations - [ ] add links from tracker rows to verification evidence when converting local hashes to fully linked audit rows Commits: - `41fa2cd` drafted the initial version - `065c025` tracked verification hardening as remaining work