arch-evolution

Plan architecture migrations, technical debt repayment, and governance processes with phased roadmaps.

7.1k|1.1k|Updated Mar 5, 2026
One-click install
npx skills add https://github.com/openJiuwen-ai/jiuwenswarm --skill arch-evolution
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: arch-evolution
Source: https://github.com/openJiuwen-ai/jiuwenswarm/tree/main/jiuwenswarm/resources/agent/workspace/plugins/agent_templates/system-architect/skills/arch-evolution
Command: npx skills add https://github.com/openJiuwen-ai/jiuwenswarm --skill arch-evolution

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Modernizing legacy systems without breaking production is hard: teams lack a safe migration strategy, technical debt goes unquantified, and architecture decisions drift away from team structure. This Skill provides a structured methodology for evolving architecture safely from current state to target state.

Core Features & Use Cases

  • Strangler Fig Migration Planning: Step-by-step legacy replacement with facade gateways, parallel running, and staged traffic shifting (1% to 100%) with rollback plans.
  • Technical Debt Management: D1-D4 severity rating matrix, repayment prioritization rules, and debt kanban practices reserving 15-20% iteration capacity.
  • Architecture Governance: ARB review board workflows, ADR lifecycle management, technology radar quadrants, and Team Topologies alignment with Conway's Law.
  • Database Schema Evolution: Index strategy, rollback-safe migrations, large-table phased changes, multi-tenancy patterns, and sharding decisions.
  • Use Case: A team needs to split a monolith into services. Use this Skill to assess the current state, define the target architecture, build a four-phase migration roadmap with rollback criteria, and set up an ARB process to govern ongoing decisions.

Quick Start

Ask the agent to create a phased migration roadmap from our current monolith to microservices using the strangler fig pattern, including technical debt ratings and rollback plans for each phase.

Frequently Asked Questions about arch-evolution

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

FAQPage Schema
How do I migrate a monolith to microservices without downtime?

Use the strangler fig pattern: place an API gateway in front of the legacy system, migrate one function at a time, run old and new in parallel to verify consistency, then shift traffic gradually from 1% to 100%. Always keep a one-click rollback path through the gateway.

How to prioritize technical debt repayment across teams?

Rate debt on a D1-D4 matrix by impact scope and urgency: D1 items blocking core flows get fixed immediately, D2 items enter the next two iterations, D3 goes into quarterly planning, and D4 is fixed opportunistically. Reserve 15-20% of each iteration's capacity for debt repayment.

What is the safest way to change schema on large database tables?

For tables over 10 million rows, use a phased approach: add a nullable column, dual-write from the application, backfill historical data in batches, switch reads to the new field, then drop the old column. Every migration must have a corresponding rollback.

Which multi-tenancy database pattern should I choose?

Choose based on tenant count and isolation needs: database-per-tenant for under 50 tenants needing strong isolation, schema-per-tenant for 50-5000 tenants, and shared schema with tenant_id plus row-level security for large-scale SaaS. Each index in shared schema must be prefixed with tenant_id.

When should an Architecture Review Board reject a proposal?

A proposal should be sent back when it lacks alternatives analysis, quality attribute impact, risk mitigation, dependency mapping, or a rollback plan. The ARB records decisions as ADRs and tracks execution through milestone checkpoints.