AGENT SKILL

ui-ux

Apply UX psychology to design, critique, or redesign product screens and flows (onboarding, forms, signup walls, upgrades, pricing, conversion moments). Uses six evidence-backed principles: smart defaults, goal-gradient progress, reciprocity, IKEA/endowment ownership, loss aversion, and contrast/anchoring — with ethical anti-dark-pattern guardrails. Use when the user says "UI/UX", "UX psychology", "critique this screen", "improve conversion", "onboarding UX", "signup wall", "pricing screen", "upgrade modal", "design this flow", "/ui-ux", or wants before/after product UI recommendations grounded in how people actually decide.

License 许可
MIT
最近更新
2026-07-19
Stars
★ 1
Forks
1

01 前言 Front Matter

name
ui-ux
metadata
{ "source": "uxpeak — The UX Psychology Behind Apps People Can't Stop Using (https://youtu.be/2TlIg3VokY8)", "principles-ref": "references/principles.md", "short-description": "UX psychology for screens: defaults, progress, reciprocity, ownership, loss, contrast" }
description
Apply UX psychology to design, critique, or redesign product screens and flows (onboarding, forms, signup walls, upgrades, pricing, conversion moments). Uses six evidence-backed principles: smart defaults, goal-gradient progress, reciprocity, IKEA/endowment ownership, loss aversion, and contrast/anchoring — with ethical anti-dark-pattern guardrails. Use when the user says "UI/UX", "UX psychology", "critique this screen", "improve conversion", "onboarding UX", "signup wall", "pricing screen", "upgrade modal", "design this flow", "/ui-ux", or wants before/after product UI recommendations grounded in how people actually decide.
when-to-use
Product UI/UX review or design work where decision psychology, conversion, or retention matters more than pure visual polish. Screens, Figma/export descriptions, app flows, web forms, paywalls, checkouts.
argument-hint
<screen/flow description, screenshot path, or URL> [critique|redesign|checklist]

02 资源文件 1

03 SKILL.md 正文

UI/UX (Psychology-First Product Design)

Purpose

Help the user design, critique, or redesign product UI so the next action feels obvious, valuable, and worth completing. Visual polish alone is not the goal. Decision psychology is the goal.

Apps fail more often from ignoring how people decide than from ugly pixels.

Core Contract

  1. Ground every recommendation in one or more of the six principles below.
  2. Prefer concrete before → after screen changes over abstract advice.
  3. Prefer one primary action per moment of friction.
  4. Use psychology to reduce confusion and build trust, never to deceive.
  5. If the user only wants visual style (colors, type, glassmorphism), say so and either narrow to psychology or defer — this skill owns behavioral UX, not pure aesthetic art direction.

Six Principles (Quick Map)

# Principle User-job shift Primary surfaces
1 Smart defaults Fill from scratch → scan & adjust Forms, filters, booking, settings
2 Goal-gradient (never start at 0%) Standing still → already in motion Onboarding, checklists, profiles
3 Reciprocity (value before wall) Hostage → gift then invite Results gates, signup walls, freemium
4 IKEA + endowment (invest first) Disposable form → something that's mine Pre-auth setup, personalization
5 Loss aversion + status quo Pitch to ignore → cost of inaction Upgrades, storage limits, renewals
6 Contrast / anchoring Absolute price shock → relative fit Pricing, add-ons, protection plans

Full before/after rules, study anchors, and anti-patterns: references/principles.md (read when applying in depth or when the user wants sources).

Modes

Infer mode from the request; default to critique if ambiguous.

Mode When Output emphasis
critique Existing screen/flow Friction inventory + ranked principle hits
redesign User wants a better version Concrete before→after changes + copy + layout notes
checklist Audit many screens / ship readiness Pass/fail per principle with one fix each
design-from-scratch New flow, no UI yet Ordered flow + key screens using the six principles

Workflow

1. Load the target

  • Read attached screenshots/images with read_file (multimodal).
  • Read local design files, HTML/React screens, or copy docs.
  • If only a verbal description is given, restate the assumed screen layout before advising.
  • Do not invent product data the user never provided; mark assumptions explicitly.

2. Name the decision moment

In one sentence:

User is deciding whether to [action] after [context], with [stakes].

Examples: "continue onboarding", "create account", "upgrade now", "add protection plan".

3. Map friction to principles

For each blocking moment, ask:

  1. Are they forced to invent answers we already know? → defaults
  2. Does progress start at zero or feel endless? → goal-gradient
  3. Are we asking before giving anything real? → reciprocity
  4. Is there nothing yet that feels theirs? → IKEA / endowment
  5. Is the CTA a gain pitch with a free "maybe later"? → loss aversion
  6. Is a cost shown without a controlled first number? → contrast

Skip principles that do not apply. Do not force all six onto every screen.

4. Propose changes (smallest that move behavior)

For each high-impact issue:

  • Principle: name it
  • Current: what the UI does now
  • Change: exact UI/copy/structure change
  • Why it works: one sentence in behavioral terms
  • Risk / ethics check: any dark-pattern edge?

Prefer one decisive redesign over ten micro-nits when conversion is the goal.

5. Ethics gate (mandatory)

Reject or rewrite any recommendation that:

  • Fakes urgency, scarcity, progress, results, or social proof
  • Locks already-earned value behind a surprise wall after implying it was free
  • Uses loss framing for harm (fear, shame) without a truthful stake
  • Pre-selects paid plans, dark-pattern checkboxes, or hard-to-find dismiss controls
  • Mislabels secondary actions to guilt the user ("No, I hate saving money")

Allowed: honest defaults, real partial value, real progress, real cost of inaction, transparent relative pricing.

If the user's request is manipulative, refuse the dark version and offer an ethical equivalent that still uses the principle.

Output Shape

Use this structure unless the user asks for a different format:

# UI/UX Review: <screen or flow name>

## Decision moment
<one sentence>

## What's working
- ...

## Friction → principle map
| Moment | Principle | Severity (H/M/L) | Fix in one line |
|--------|-----------|------------------|-----------------|

## Recommended redesign (before → after)
### 1. <change title>
- **Before:** ...
- **After:** ...
- **Principle:** ...
- **Copy (if needed):** "..."

## Primary CTA & secondary
- Primary: ...
- Secondary: ... (honest, not a guilt trap)

## Ethics check
- Pass / concerns + mitigations

## Optional next probe
- One measurement or A/B that would falsify whether the change helped

For checklist mode, use a six-row table (principle / status / evidence / fix).

For design-from-scratch, output:

  1. User job + success metric
  2. Step sequence (ordered)
  3. Key screens with principle tags
  4. Empty states / failure / dismiss behavior
  5. What not to add (complexity that reintroduces decision fatigue)

Working With Visuals

When screenshots or mockups are present:

  1. Describe what you actually see (layout regions, hierarchy, CTAs) before advising.
  2. Point to specific UI regions ("top progress bar", "blurred results panel", "footer dismiss").
  3. Prefer overlay-style notes: where to change, not only that something should change.
  4. If generating a visual mock with image tools, keep labels legible and match the product's existing visual language unless redesigning brand.

Related Skills

  • playwright / safari-browser — observe live product behavior; this skill interprets UX intent.
  • hyperframes / imagegen — produce motion or mock visuals after the behavioral design is set.
  • user-story / technical-task — turn agreed UX changes into implementation work items.
  • method-simplification — compress overbuilt flows while preserving the conversion contract.

Success Criteria

  • Recommendations cite named principles, not vibes.
  • At least one concrete before→after per high-severity friction.
  • Ethics gate is explicit.
  • User can implement the top change without guessing copy or structure.
  • Unused principles are omitted, not shoehorned.

Invocation Examples

  • /ui-ux critique this onboarding screenshot
  • UI/UX redesign our signup wall for the SEO analyzer
  • /ui-ux checklist the upgrade modal in src/components/Paywall.tsx
  • Apply UX psychology to this checkout add-on pricing row
  • /skills ui-ux then attach a flow description

Source Note

Principles distilled from uxpeak's public lesson "The UX Psychology Behind Apps People Can't Stop Using" (https://youtu.be/2TlIg3VokY8), expanded into an agent-operable workflow. Treat cited classic studies (jam choice, car-wash stamps, Kahneman loss aversion, Cialdini reciprocity, IKEA effect, contrast effect) as design heuristics with famous anchors, not as peer-reviewed claims re-verified in this skill file. When the user needs academic precision, flag uncertainty and offer to verify sources.

04 同仓库更多 Skills

zfifteen
zfifteen/skills-catalog

Zotero

Use Zotero Desktop from Codex to enable/probe the local API, search a local Zotero library, list items/collections/tags, export BibTeX, insert citation keys into LaTeX or Markdown drafts, read indexed full text when requested, and import BibTeX/RIS records into Zotero through the connector server. Use when the user mentions Zotero, citations, references.bib, BibTeX export, local Zotero API, localhost:23119, or adding citations from a Zotero library.

★ 1
zfifteen
zfifteen/skills-catalog

zoom-apps-sdk

Use when using Apps SDK.

★ 1
zfifteen
zfifteen/skills-catalog

youtube

Obtain accurate, timestamped YouTube video transcripts (preferring manual captions over auto-generated) as the primary deliverable, then derive summaries, key claims, or notes from that transcript. Use when the user provides a YouTube URL (or video ID) and wants the transcript, a faithful summary, key claims with attribution, timestamped notes, or "what does this video actually say". The skill is transcript-first: never substitute metadata, title, description, or search snippets for the spoken content. Triggers: "youtube transcript", "get the transcript of this video", "summarize this youtube video", "what did they say in https://youtube.com/...", "key claims from this talk".

★ 1
zfifteen
zfifteen/skills-catalog

yeet

Publish local changes to GitHub by confirming scope, committing intentionally, pushing the branch, and opening a draft PR through the GitHub app from this plugin, with `gh` used only as a fallback where connector coverage is insufficient.

★ 1
zfifteen
zfifteen/skills-catalog

x-post-creator

Create X-ready post threads and co-located story packages from the active conversation or referenced project material. Use when the user says turn this into an X-post, /x-post, make this into an X thread, or wants plain-language social posts with supporting images saved to docs/x-posts/[story-name]/.

★ 1
zfifteen
zfifteen/skills-catalog

x-api

Search, read, and write X posts through the local X API MCP server (the same backend used by the Codex x-api skill). Use for searching X/x.com, reading threads and replies, creating lists, staging/publishing posts for @alltheputs, finding high-visibility reply targets for research outreach, ingesting public context, or any request that mentions /x-api or "search X".

★ 1