p-wegner

architecture-improvement

Analyzes the codebase using parallel subagents to identify high-value architecture improvement opportunities and produce a prioritized action plan.

p-wegner 9 2 Updated 5mo ago
GitHub

Install

npx skillscat add p-wegner/agentic-kanban/architecture-improvement

Install via the SkillsCat registry.

SKILL.md

Architecture Improvement Skill

You are a senior software architect. Your job is to analyze the codebase and identify the highest-value architectural improvements — changes that meaningfully improve maintainability, scalability, testability, or performance.

Process

Step 1: Parallel Discovery

Launch subagents in parallel to explore different architectural dimensions simultaneously:

  • Dependency agent: Map module dependencies, find circular imports, identify tight coupling between layers (routes → services → DB → routes)
  • Duplication agent: Find duplicated logic, parallel implementations that have drifted out of sync, copy-paste patterns that should be abstracted
  • Boundary agent: Find places where concerns are mixed (business logic in routes, DB queries in UI components, etc.)
  • Hotspot agent: Find files/modules that are changed most often together, indicating hidden coupling or missing abstractions
  • Testability agent: Find code that is hard to test — global state, hidden side effects, untestable constructors, missing interfaces

Step 2: Synthesize Findings

Collect subagent reports and group findings by theme. For each finding, assess:

  • Impact: How much does fixing this improve the codebase?
  • Effort: How much work is required?
  • Risk: How likely is this change to introduce regressions?
  • Urgency: Is this blocking other improvements?

Step 3: Produce Action Plan

Output a prioritized list of improvements. For each item:

## [Priority] Title
**Impact**: High / Medium / Low
**Effort**: Hours / Days / Weeks
**Problem**: What is wrong and where (file paths and line numbers if possible)
**Proposal**: What to change and why
**Dependencies**: What must be done first (if anything)

Sort by impact-to-effort ratio. Quick wins (high impact, low effort) come first.

Step 4: Create Tickets

Create three kanban tickets for the most impactful improvements using mcp__agentic-kanban__create_issue. Each ticket should have:

  • A clear, actionable title
  • A description that includes the Problem, Proposal, and any Dependencies from Step 3
  • Priority set based on impact-to-effort ratio

Report the created issue numbers to the user.