wf-codebase-health

Guides module design and scores codebase health against structural principles with file:line evidence.

Updated May 9, 2026
One-click install
npx skills add https://github.com/23min/aiwf --skill wf-codebase-health-23min
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: wf-codebase-health
Source: https://github.com/23min/aiwf/tree/main/internal/skills/embedded-rituals/plugins/wf-rituals/skills/wf-codebase-health
Command: npx skills add https://github.com/23min/aiwf --skill wf-codebase-health-23min

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Codebases decay through accumulated small decisions — blurred module boundaries, duplicated logic, untyped contracts, tests that pin implementation instead of behavior — until a senior engineer needs weeks instead of hours to understand the code. This Skill provides a stack-agnostic field guide of code-health principles to consult while designing new modules and to score existing code against, catching structural decay before it compounds. ## Core Features & Use Cases - Design priming: Consult the principles while designing a new module, planning a refactor, or writing a spec that introduces new boundaries, so structure comes out right the first time. - Codebase scoring: Score an inherited or vibe-coded codebase Strong / Weak / Missing per principle with file:line evidence, then fix the one or two worst failures and re-score. - Eight principle families: Module boundaries, contracts, data discipline, behavior-pinning tests, errors/logs/audit, reasoning aids, operational properties, and economy — each with smells, moves, and explicit tradeoffs. - Use Case: When reviewing a large diff that introduces a new boundary, run this Skill alongside the per-diff review gate to audit whole-codebase structural properties a single-diff review skips. ## Quick Start Ask the assistant to score this codebase against the wf-codebase-health principles and report Strong, Weak, or Missing per principle with file:line evidence.

Frequently Asked Questions about wf-codebase-health

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

FAQPage Schema
How do I score a codebase for structural health?

Score each principle in the rubric as Strong, Weak, or Missing with file:line evidence for every judgment. Then pick the one or two principles failing hardest, fix those, and re-score rather than attempting to fix everything at once.

What code review checklist covers whole-codebase design instead of a single diff?

This rubric covers module boundaries, contracts, data discipline, tests, errors and audit, reasoning aids, operational properties, and economy. It complements per-diff review gates, which check correctness and conventions on the change in front of you.

When should I use this guide versus a per-diff code review?

Use the per-diff gate for correctness, acceptance-criteria coverage, and conventions on a specific change. Use this whole-codebase guide when designing new modules, planning refactors, or reviewing large diffs that introduce new boundaries — on such diffs, run both.

Does this code-health rubric enforce rules automatically?

No, the principles are advisory forces, not pass/fail gates, and the Skill enforces nothing. If a project wants mechanical enforcement, that project owns the decision and implements it as a linter rule, CI check, or test in its own terms.

Is this rubric tied to a specific language or framework?

No, it is stack-agnostic — it names no specific libraries, frameworks, or methodologies. Examples reference general concepts like dataclasses, JSON Schema, and structured logging that apply across languages, and the project's own conventions always win.