# Shared Review Workflow

> Severity definitions, approval gate protocol, and fix constraints shared across all review and test skills.

- Skill: `majiayu000/shared-review-workflow` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add majiayu000/shared-review-workflow`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/shared-review-workflow/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/shared-review-workflow

---


## Purpose

Define the common review workflow shared by code-review, design-review, and unit-test skills across all platforms.

## Severity definitions

| Severity     | Meaning                                                         |
| ------------ | --------------------------------------------------------------- |
| must-fix     | Security vulnerability, broken functionality, blocks deployment or App Store submission |
| should-fix   | Standards violation, maintainability concern, tech debt         |
| nice-to-have | Style preference, minor improvement, optional optimization      |

## Approval gate

If issues are found and `--no-fix` is not set:

> "Found X issues (Y must-fix). Approve fixes? (yes/no)"

Do not modify code before approval.

## Fix and confirm (if approved)

- Apply fixes
- Re-run automated gates (lint, format, typecheck)
- Run unit tests to confirm no regressions
- Report final status

## Fix constraints

- Don't add dependencies unless required
- Don't weaken assertions to make tests pass
- Prefer targeted fixes over broad refactors
- Don't disable or skip tests to make them pass