Agent-Skills.md

Agent Skills: Architecture Audit

Protocol for auditing structural debt, logic leakage, and fragmentation across Web, Mobile, and Backend. (triggers: package.json, pubspec.yaml, go.mod, pom.xml, nest-cli.json, architecture audit, code review, tech debt, logic leakage, refactor)

UncategorizedID: hoangnguyen0403/agent-skills-standard/common-architecture-audit

Repository

hoangnguyen0403/agent-skills-standard
HoangNguyen0403License: Apache-2.0
362103

Install this agent skill to your local

pnpm dlx add-skill https://github.com/HoangNguyen0403/agent-skills-standard/tree/HEAD/skills/common/common-architecture-audit

Skill Files

Browse the full folder contents for common-architecture-audit.

Download Skill

Loading file tree…

skills/common/common-architecture-audit/SKILL.md

Skill Metadata

Name
common-architecture-audit
Description
Audit structural debt, logic leakage, and monolithic components across Web, Mobile, and Backend codebases. Use when reviewing architecture, assessing tech debt, detecting logic in wrong layers, or identifying God classes.

Architecture Audit

Priority: P1 (STANDARD)

1. Discover Structural Duplication

Identify split sources of truth by searching for redundant directory patterns.

  • Compare Service.ts vs ServiceNew.ts vs ServiceV2.ts.
  • Check for /v1, /v2 or "Refactor" folders.

See implementation examples for detection scripts.

2. Detect Logic Leakage (by Ecosystem)

Find business logic trapped in wrong layer.

  • Web (React/Next.js/Vue): grep -rE "useEffect|useState|useMemo" components --include="*.tsx" | wc -l — if components/ hook count > 20x hooks/ folder, architecture monolithic.
  • Mobile (Flutter/React Native): grep -rE "http\.|dio\.|socket\." lib/widgets --include="*.dart" | wc -l — I/O or state mutation > 5 lines in build() High Debt.
  • Backend (NestJS/Go/Spring): grep -rE "Repository\.|Query\.|db\." src/controllers --include="*.ts" | wc -l — Controllers must only handle request parsing and response formatting.

3. Identify Monoliths

Flag massive files violating Single Responsibility Principle.

  • UI: > 500 lines (Medium), > 1,000 lines (Critical).
  • Backend Services: > 1,500 lines indicates "God Class".

See implementation examples for monolith detection scripts.

4. Audit Resource Performance

Check for large metadata or constants impacting IDE performance and binary size.

  • Resources > 1,000 lines require granulation.

See implementation examples for resource audit scripts.

Scoring Impact

  • Layer Violation: -15 per business logic instance in UI/Controller layer.
  • Structural Fragmentation: -10 per duplicated legacy entity.
  • Monoliths: -10 per unit > 1,000 lines.

Anti-Patterns

  • No applying generic patterns over project-specific rules: Respect existing architecture constraints.
  • No ignoring error handling or edge cases: Audit must cover boundary conditions.

References