projection-patterns

Build projections and read models from event streams for CQRS queries.

Updated Apr 4, 2026
One-click install
npx skills add https://github.com/emilneuraz-ai/neuraz-web --skill projection-patterns-emilneuraz-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: projection-patterns
Source: https://github.com/emilneuraz-ai/neuraz-web/tree/main/.agents/skills/.agents/skills/projection-patterns
Command: npx skills add https://github.com/emilneuraz-ai/neuraz-web --skill projection-patterns-emilneuraz-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Convert event streams into projections and read models to support CQRS queries and fast, scalable reads.

Core Features & Use Cases

  • Projection architectures and types (Live, Catchup, Persistent, Inline) for real-time and historical views.
  • Templates for multi-table projections, aggregated views, and search-oriented read models.
  • Safe, idempotent replay and checkpointing to enable rebuilds and resilience.

Quick Start

Import the projection templates and wire your event store to start building read models from streams.

Frequently Asked Questions about projection-patterns

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build read models from event streams for CQRS queries?

You build read models by applying projections to event streams, transforming events into queryable formats. This enables fast CQRS reads by separating write event streams from optimized, scalable read-side data structures.

What are the different types of projection patterns in event-sourced architectures?

Projection patterns in event sourcing include Live, Catchup, Persistent, and Inline types. These variations handle real-time and historical views differently, allowing you to balance latency and rebuild capabilities for read models.

How do I ensure transactional integrity when projecting events to multiple tables?

Transactional integrity for multi-table projections is maintained using templates that coordinate writes across tables. This ensures read models remain consistent and accurate during complex event stream processing and state rebuilding.

Can I safely replay event streams to rebuild projections without data corruption?

Yes, you can safely replay event streams to rebuild projections using idempotent replay and checkpointing mechanisms. This allows resilient recovery and rebuilding of read models without duplicating data or corrupting state.

When do I need catchup projections versus live projections for read models?

You need catchup projections when reconstructing historical state from an event stream, and live projections for processing only new events in real-time. Combining both allows catchup to build baseline state before switching to live updates.