plan-design-review

Reviews UI/UX design plans across seven scored passes and fixes gaps interactively.

1|Updated Jul 23, 2026
One-click install
npx skills add https://github.com/sanjanb/my-agent-harness --skill plan-design-review-sanjanb
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plan-design-review
Source: https://github.com/sanjanb/my-agent-harness/tree/main/skills/plan-design-review
Command: npx skills add https://github.com/sanjanb/my-agent-harness --skill plan-design-review-sanjanb

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Design plans often ship with vague UI descriptions, missing interaction states, and generic AI-generated patterns that only surface as problems during implementation. This Skill audits a design plan before any code is written, scoring each design dimension and walking the user through fixes interactively. ## Core Features & Use Cases - Seven scored design passes: Rates information architecture, interaction states, user journey, AI slop risk, design system alignment, responsive/accessibility, and unresolved decisions on a 0-10 scale, then fixes each gap to reach 10. - AI slop detection: Flags recognizable generic patterns like 3-column feature grids, purple gradients, and placeholder-as-label forms against an explicit blacklist. - Interactive decision workflow: Raises one AskUserQuestion per issue with recommendations, effort estimates, and completeness scores, writing approved decisions back into the plan file. - Use Case: Before implementing a new dashboard, run this Skill on the plan to catch missing empty states, unspecified mobile behavior, and generic card layouts, then produce a task list and review report for downstream shipping. ## Quick Start Ask the agent to review the design plan for the current feature before implementation begins.

Frequently Asked Questions about plan-design-review

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

FAQPage Schema
How do I review a design plan before implementation?

Invoke the plan-design-review skill while in plan mode. It runs seven scored passes covering information architecture, states, journey, AI slop, design system fit, responsiveness, and open decisions, asking one question per finding before editing the plan.

What design issues does a plan review catch?

It catches missing loading/empty/error states, vague UI descriptions, generic AI-generated patterns like 3-column feature grids and purple gradients, missing accessibility specs, and unresolved decisions such as mobile navigation behavior.

Can I use this for live site visual audits?

No, this skill works on design plans in plan mode, not live sites. For visual audits of a running site, the documentation directs you to use the separate /design-review skill instead.

Does the design review work without user interaction?

It is interactive by design and requires AskUserQuestion responses for each finding. In headless sessions it blocks; in spawned sessions it auto-chooses recommended options and reports results as prose.

What happens if the review finds zero issues?

If every pass has zero findings, the review states so per section and proceeds directly to the completion summary and review log. Any non-trivial finding must go through a user question before the plan is finalized.