review-shipped-code

Reviews React, Next.js, and TypeScript diffs against standards and spec axes before merge.

1|Updated Jul 13, 2026
One-click install
npx skills add https://github.com/dineshrevunuru/SuperSkills --skill review-shipped-code-dineshrevunuru
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-shipped-code
Source: https://github.com/dineshrevunuru/SuperSkills/tree/main/review-shipped-code
Command: npx skills add https://github.com/dineshrevunuru/SuperSkills --skill review-shipped-code-dineshrevunuru

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Design and rendered-UI reviews never read the source code, so production front-end changes can ship with inaccessible controls, hardcoded values, dead code, or drift from the original spec while still looking fine. This Skill closes that gap with a source-level code review for production React/Next/TypeScript apps. ## Core Features & Use Cases - Two isolated review axes: A STANDARDS axis checks accessibility, design-token conformance, motion register, React/Next idioms, type safety, and dead code; a SPEC axis checks fidelity to the originating spec for missing requirements, scope creep, and misread implementations. - Mode gate and intake gate: Routes prototypes out, downgrades pure refactors or small diffs to a quick pass, and defaults features to a full pass with parallel isolated sub-agents. - AND-gate ship verdict: Reports each axis separately with severity-rated findings and blocks shipping if either axis carries a critical issue, with the shortest path to unblock. - Use Case: Before merging a PR that adds a booking-flow button, run the review to catch a non-semantic clickable div, an ungated spring animation, and hardcoded hex values that a rendered-UI check would miss. ## Quick Start Review the diff of my current branch against main and tell me if this code is ready to ship.

Frequently Asked Questions about review-shipped-code

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

FAQPage Schema
How do I review a React or Next.js pull request before merging?

Run a source-level review of the diff against a fixed point such as main...HEAD. The review checks accessibility, design-token usage, React and Next idioms, type safety, and dead code on a standards axis, plus fidelity to the spec on a separate axis, then issues a ship verdict.

What is the difference between a standards review and a spec review?

The standards axis judges code craft: accessibility, tokens, motion, idioms, and dead code. The spec axis judges fidelity to the originating requirements: missing items, scope creep, and misread implementations. A change can pass one and fail the other, so they run isolated and are never merged.

Can I use this code review on prototypes or throwaway demos?

No. Prototypes and throwaway experiments are reviewed for feel and learning, not code craft, so hardcoded values and console logs are acceptable there. This review applies only to production code that real users will touch.

What happens if there is no spec for the change being reviewed?

The spec axis reports 'no spec available' and is skipped rather than inventing acceptance criteria. The standards axis still runs normally, and the ship verdict is based on the axes that produced findings.

Why does the review skip issues that ESLint or TypeScript already catch?

Tooling like ESLint, tsc, Prettier, and jsx-a11y already guarantees a baseline layer, so re-reporting it wastes the review. The review focuses on what tooling cannot see: token conformance, motion register, real accessibility, and design judgment.