This directory is the product: a production-grade retail platform built on the holiday-peak-lib framework. It runs end-to-end with real domain logic, real enterprise connectors, real SLOs, AGC weighted-canary routing, continuous agent evaluation, three-tier memory, and full observability — not a demo.
Canonical positioning: .github/instructions/repository-purpose.instructions.md. The framework lives in
lib/. The product lives here.
The product is composed of:
- 1 transactional microservice (
crud-service) — pure CRUD, ACID guarantees, event publishing - 26 agent services — bounded-context AI agents with dynamic planning, MCP tools, three-tier memory, SLM-first routing
- 1 Next.js frontend (
ui) — operator/admin/retail surfaces consumed by both agents and humans
Changes to apps are product changes: they require domain reasoning, eval impact analysis, SLO awareness, and operational runbook updates when behavior changes.
| App | Type | Purpose |
|---|---|---|
crm-campaign-intelligence |
Agent service | Generates campaign intelligence from CRM and funnel context. |
crm-profile-aggregation |
Agent service | Builds unified customer profiles from distributed CRM data. |
crm-segmentation-personalization |
Agent service | Segments customers and suggests personalization actions. |
crm-support-assistance |
Agent service | Produces support-assist context and recommended actions. |
crud-service |
Microservice | Owns transactional CRUD APIs and event publishing. |
ecommerce-cart-intelligence |
Agent service | Scores cart risk and recommends conversion improvements. |
ecommerce-catalog-search |
Agent service | Delivers catalog discovery and ACP-aligned search responses. |
ecommerce-checkout-support |
Agent service | Evaluates checkout blockers and completion guidance. |
ecommerce-order-status |
Agent service | Provides order and shipment status intelligence. |
ecommerce-product-detail-enrichment |
Agent service | Enriches product detail context for shopping experiences. |
inventory-alerts-triggers |
Agent service | Detects inventory alerts and trigger conditions. |
inventory-health-check |
Agent service | Assesses inventory health and anomaly signals. |
inventory-jit-replenishment |
Agent service | Recommends just-in-time replenishment actions. |
inventory-reservation-validation |
Agent service | Validates reservation requests against stock conditions. |
logistics-carrier-selection |
Agent service | Recommends carrier options and trade-offs. |
logistics-eta-computation |
Agent service | Computes ETA projections and delay risk signals. |
logistics-returns-support |
Agent service | Guides returns-related operational decisions. |
logistics-route-issue-detection |
Agent service | Detects route issues and recovery recommendations. |
product-management-acp-transformation |
Agent service | Transforms product data into ACP-aligned payloads. |
product-management-assortment-optimization |
Agent service | Ranks and optimizes assortment decisions. |
product-management-consistency-validation |
Agent service | Evaluates product consistency and completeness signals. |
product-management-normalization-classification |
Agent service | Normalizes and classifies product attributes. |
search-enrichment-agent |
Agent service | Enriches search-oriented product content asynchronously. |
truth-enrichment |
Agent service | Generates proposed truth-layer attribute enrichments. |
truth-export |
Agent service | Exports truth-layer attributes to downstream protocols/systems. |
truth-hitl |
Agent service | Supports human-in-the-loop review and approval queues. |
truth-ingestion |
Agent service | Ingests source product data into truth workflows. |
ui |
Frontend | Next.js portal for admin, operations, and retail workflows. |
Most Python app services are FastAPI services built with build_service_app, exposing synchronous interfaces and background Event Hub processing. Configuration is environment-driven, including Foundry model settings and optional Redis/Cosmos/Blob memory tiers.