team-arch-opt

Coordinate architecture refactoring through analysis, prioritized planning, execution, and validation.

1|Updated May 12, 2026
One-click install
npx skills add https://github.com/tcytan/tecton --skill team-arch-opt
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: team-arch-opt
Source: https://github.com/tcytan/tecton/tree/main/.claude/skills/team-arch-opt
Command: npx skills add https://github.com/tcytan/tecton --skill team-arch-opt

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps teams systematically improve an existing codebase’s architecture by coordinating analysis, refactoring design, implementation, and verification—without skipping safety checks.

Core Features & Use Cases

  • Unified multi-agent pipeline: Analyzer → Designer → Refactorer → Validator + Reviewer with structured outputs at each stage.
  • Parallel refactoring strategies: Supports single, fan-out (branch-by-refactor), independent (multi-target), and auto mode that chooses an approach based on refactoring count.
  • Review-fix safety loop: Automatically creates targeted fix tasks when validation or review reports issues.
  • Evidence-based deliverables: Produces baseline metrics, ranked architecture issues, a prioritized refactoring plan, validation results, and a review report.

Quick Start

Run a coordinated architecture optimization session by giving the task description “Refactor the gateway runtime modules to reduce coupling and improve layering in the tecton repository.”

Frequently Asked Questions about team-arch-opt

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

FAQPage Schema
How do I coordinate multi-agent architecture optimization to refactor a codebase?

Multi-agent architecture optimization is coordinated through a pipeline of Analyzer, Designer, Refactorer, Validator, and Reviewer phases. This pipeline systematically analyzes code structure, designs prioritized refactoring steps, executes changes, and validates outcomes with structured outputs at each stage.

What is a fan-out branch refactor and when should I use it?

A fan-out branch refactor is a parallel refactoring strategy where independent changes are executed across separate branches. You should use this approach, or let auto mode select it, when you have multiple refactoring targets to improve coupling, layering, duplication, or dead code simultaneously.

How does the review-fix safety loop handle validation failures during code refactoring?

The review-fix safety loop handles validation failures by automatically creating targeted fix tasks when validation or review reports issues. This enforces safety checks during architecture optimization and escalates fix cycles until the refactoring outcomes meet the required validation verdicts.

Do I need a YAML entrypoint to use this architecture optimization skill?

Yes, you need a YAML-defined skill entrypoint to use this architecture optimization skill. The YAML entrypoint is required to orchestrate role-based phases with dependency-aware task spawning for analyzing code structure and executing coordinated multi-agent refactoring plans.

What evidence-based deliverables are produced after reducing code coupling and layering?

Evidence-based deliverables produced after reducing code coupling and layering include baseline metrics, ranked architecture issues, a prioritized refactoring plan, validation results, and a review report. These structured outputs document the entire architecture optimization pipeline from analysis to verification.