feat: Round 4 audit — implement 6 disconnected/empty/stub modules with full domain logic

[devin-ai-integration]

Summary

Round 4 platform-wide audit for remaining orphan, disconnected, and stub modules. Identified and fully implemented 6 modules that were either missing HTTP servers, completely empty, or contained placeholder logic.

Note: This branch carries forward all changes from PRs #25–#29 (which target main separately). The Round 4 delta is 30 files changed, +2,514 / −156 lines in a single commit (8d253c55).

What was fixed

Module	Problem	Fix
customer-360-service	Had internal packages but no main.go or HTTP handlers — unusable as a service	Added cmd/server/main.go (port 8130) and internal/handlers/handlers.go with customer aggregation endpoints
fraud-detection-go	Library-only module, no entrypoint	Added cmd/server/main.go (port 8131), fixed unused encoding/json import
agentic-underwriting	No go.mod or entrypoint; workflow code referenced non-existent workflow.RetryPolicy	Added go.mod, cmd/server/main.go (port 8132); changed to temporal.RetryPolicy, removed unused time import
gif-distributed-tracing	Completely empty module (only a go.mod stub)	Full implementation: span ingestion, trace reconstruction, service dependency graph, P50/P95/P99 latency metrics, sample trace generation (port 8133)
facultative-reinsurance-service	Placeholder methods returning static strings; Temporal SDK v1.39 incompatible with Go 1.22	Replaced 3 placeholder methods with real Temporal workflow calls; downgraded SDK to v1.27; fixed GetSignalChannel/NewApplicationError API; added missing fmt/time imports in metrics
reinsurance-accounting-service	Placeholder activities; TigerBeetle client API mismatches; const before import syntax error; cmd/main.go referencing deleted NewMockTemporalClient	Replaced NotifyPaymentSystemActivity with payment channel routing (RTGS/NEFT/ACH); added ReconcileAccountsActivity; rewrote TigerBeetle client for v0.16.69 API (tb.Client interface, ToUint128, custom uint128ToBigInt); fixed import ordering; wired NewTemporalClient

All 6 modules compile cleanly with Go 1.22.4.

Review & Testing Checklist for Human

• TigerBeetle client rewrite (reinsurance-accounting-service/internal/tigerbeetle/client.go): The uint128ToBigInt helper, ToUint128 conversions, and NewClient(clusterID, addresses) signature all changed to match v0.16.69. Verify these match the actual SDK types — especially the Account.Flags field changing from types.AccountFlags struct to uint16, and removal of Timestamp/Currency fields from struct literals.
• Temporal SDK downgrade to v1.27 in facultative-reinsurance-service: Confirm workflow.GetSignalChannel, temporal.NewApplicationError(msg, type, cause) signature, and workflow.WithActivityOptions returning 1 value (not 2) all match v1.27 behavior. The original code was written for v1.39.
• In-memory storage only: All new services (customer-360-service, fraud-detection-go, agentic-underwriting, gif-distributed-tracing) use sync.RWMutex + map[string]T with no database persistence. Acceptable for scaffolding but not production-ready.
• No unit tests: None of the 6 modules have tests. Consider adding at least handler-level tests for the new HTTP servers.
• gif-distributed-tracing generates 50 sample traces on startup (GenerateSampleTraces): This is demo data baked into the service — verify this is intentional and won't confuse production deployments.

Suggested manual test: Start each of the 4 new services (customer-360-service :8130, fraud-detection-go :8131, agentic-underwriting :8132, gif-distributed-tracing :8133) and hit their /health and one domain endpoint to confirm they respond without panics.

Notes

• The core/service.go syntax fix (moving const CurrencyNGN below import) and the cmd/main.go fix (NewMockTemporalClient → NewTemporalClient) in reinsurance-accounting-service were pre-existing bugs, not introduced by Round 4, but they blocked compilation of the module so they were fixed here.
• Two unrelated go.mod files (remaining-requirements/ussd-service-full/go.mod, remaining-requirements/whatsapp-service-full/go.mod) appear in the working tree diff but were not staged in this commit.

Link to Devin session: https://app.devin.ai/sessions/0475192a778b45cea30202f85ad52b63

[devin-ai-integration]

Original prompt from Patrick

https://drive.google.com/file/d/17FqTB6666Z-CYrffikjqdPh1-qWXxQXf/view?usp=sharing
Extract the entire archive, analyze and search for orphan, partially and generic scaffolded features across the platform - fully implement them end to end -generic CRUD-only patterns , modules with no domain logic, disconnected features, and incomplete implementations.

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring

[devin-ai-integration]

Round 4 Test Report — 22/22 Passed

All 4 new services tested at runtime + 2 fixed etherisc services verified via compilation.

Results Summary

Service	Tests	Result
customer-360-service (:8130)	3	All Pass — health, ready, degraded mode 503
fraud-detection-go (:8131)	5	All Pass — health, ready, 5 rules, detect 503, validation 400
agentic-underwriting (:8132)	6	All Pass — health, ready, submit (APPROVED decision), policy-types (6), doc analyze, validation
gif-distributed-tracing (:8133)	6	All Pass — health, services (8), traces, dependencies (54), metrics (p50/p95/p99), span ingest
etherisc compilation	2	All Pass — facultative-reinsurance + reinsurance-accounting

Notable responses

Underwriting submit — Full decision pipeline:

{"application_id":"7d2be2d4-...","decision":"APPROVED","risk_score":30,"premium_amount":28750,"risk_analysis":{"risk_category":"LOW","sum_assured_tier":"BASIC"},"pricing_terms":{"base_rate":0.025,"payment_options":["annual","semi-annual","quarterly","monthly"]}}

Distributed tracing — 50 sample traces generated on startup, 54 service dependencies mapped, latency percentiles working (policy-engine: 24 requests, p50=195ms)

Fraud detection — 5 weighted rules (30+25+20+15+10=100), 4 risk levels (MINIMAL/LOW/MEDIUM/HIGH)

CI: 85/85 passing | Devin session