Free Claude Code Skills

38 AI UX Skills for Claude Code

Design judgment your coding agent applies on its own. Install once, no prompting.

$npx skills add imsaif/aiux-skills

New to skills? Read how skills work (6 lessons, about 20 minutes).

Adaptive Interfaces

Use when the UI should adjust to how each person uses it: personalized layouts, reordering features by usage, 'the app should learn what I use most'. Adaptive Interfaces tune layout and functionality to individual behavior.

Adaptive & Intelligent Systems

Used by:

Ambient Intelligence

aiux-ambient-intelligence

Use when help should appear from context without being asked: proactive background assistance, sensing situation or environment, 'it should just know', assistance while attention is elsewhere. Ambient Intelligence assists without explicit interaction.

Adaptive & Intelligent Systems

Used by:

Guided Learning

aiux-guided-learning

Use when a complex AI feature overwhelms newcomers: step-by-step walkthroughs, wizards, 'users do not know where to start', tutorials that adapt to skill level. Guided Learning breaks the task into steps matched to the user's knowledge.

Adaptive & Intelligent Systems

Used by:

Predictive Anticipation

aiux-predictive-anticipation

Use when the product should act before the user asks: suggesting next actions, pre-loading likely content, 'it should know what I need next', smart defaults from behavior. Predictive Anticipation moves the product from reactive to anticipatory.

Adaptive & Intelligent Systems

Used by:

Augmented Creation

aiux-augmented-creation

Use when AI helps people make content rather than making it for them: co-writing, draft generation with human editing, creative tools, AI-assisted design or writing. Augmented Creation keeps the user as the author with AI as collaborator.

Human-AI Collaboration

Used by:

Autonomy Spectrum

aiux-autonomy-spectrum

Use when deciding how independently the AI may act: autonomy levels, 'ask before doing vs just do it', per-task permissions, supervised vs autonomous modes. Autonomy Spectrum gives users a dial instead of an on/off switch.

Human-AI Collaboration

Used by:

Collaborative AI

aiux-collaborative-ai

Use when several people and an AI share one workflow: team workspaces with AI, shared documents or channels, 'AI in our group project', coordinating human and AI contributions. Collaborative AI keeps teams coordinated while AI participates.

Human-AI Collaboration

Used by:

Contextual Assistance

aiux-contextual-assistance

Use when help should arrive at the moment of need: proactive tips, inline suggestions, 'users do not know what to ask', onboarding hints, help without interrupting the workflow. Contextual Assistance offers timely guidance from context.

Human-AI Collaboration

Used by:

Escalation Pathways

aiux-escalation-pathways

Use when an agent should stop and ask a human: 'the agent should check with me when unsure', handoff on ambiguity or missing permissions, approval requests mid-task, avoiding both constant interruptions and silent guessing. Escalation Pathways structure when and how agents ask for guidance.

Human-AI Collaboration

Used by:

Feedback Loops

aiux-feedback-loops

Use when user corrections should improve the AI: thumbs up or down, edit-and-learn, 'it keeps repeating the same mistake', preference learning, ratings that actually change behavior. Feedback Loops turn corrections into cumulative improvement.

Human-AI Collaboration

Used by:

Graceful Handoff

aiux-graceful-handoff

Use when control passes between AI and human: taking over from automation, 'let me finish this manually', resuming automation after manual edits, no lost progress at the switch. Graceful Handoff makes transitions seamless in both directions.

Human-AI Collaboration

Used by:

Human-in-the-Loop

aiux-human-in-the-loop

Use when AI output needs human review, approval, or sign-off before it takes effect or reaches someone: approval queues, 'a person checks before send', moderation of AI answers, override and reject controls. Human-in-the-Loop keeps automation accountable to human judgment.

Human-AI Collaboration

Used by:

Intent Preview

aiux-intent-preview

Use when an agent is about to act and the user should see the plan first: 'show me what it will do before it does it', dry-run summaries, previewing steps and reversibility, approve or edit before execution. Intent Preview earns informed consent for actions.

Human-AI Collaboration

Used by:

Mixed-Initiative Control

aiux-mixed-initiative-control

Use when human and AI work on the same thing at the same time: co-editing with an agent, interrupting or redirecting mid-task, 'I want to jump in without stopping it', fluid turn-taking. Mixed-Initiative Control lets control flow both ways.

Human-AI Collaboration

Used by:

Workspace-Native Agent Integration

aiux-workspace-native-agents

Use when AI should live inside an existing tool rather than beside it: 'stop making users switch to a chat app', embedded copilots in editors or CRMs, in-context actions, meeting users where they work. Workspace-Native Agent Integration removes context-switching overhead.

Human-AI Collaboration

Used by:

Action Audit Trail

aiux-action-audit-trail

Use when users need to see or undo what an AI agent did: activity logs, 'what did the agent change', audit history, rollback of agent actions, reviewing multi-step runs. Action Audit Trail makes agent behavior reviewable and correctable after the fact.

Trustworthy & Reliable AI

Used by:

Agent Reflection & Learning

aiux-agent-reflection-learning

Use when users should see that the AI learns from their corrections: 'why does it keep making the same mistake', showing learned preferences, memory of feedback, improvement over time. Agent Reflection and Learning builds trust through visible improvement.

Trustworthy & Reliable AI

Used by:

Confidence Visualization

aiux-confidence-visualization

Use when users need to know how sure the AI is: confidence scores, uncertainty indicators, 'should I trust this answer', flagging low-certainty output, deciding when to verify. Confidence Visualization shows reliability so trust is calibrated.

Trustworthy & Reliable AI

Used by:

Error Recovery & Graceful Degradation

aiux-error-recovery

Use when the AI fails or hits something it cannot handle: error states, fallbacks, 'what happens when the model is wrong or down', retry paths, degraded modes. Error Recovery and Graceful Degradation fails clearly with a way forward.

Trustworthy & Reliable AI

Used by:

Explainable AI (XAI)

aiux-explainable-ai

Use when users ask why the AI decided something: 'show its reasoning', explanations for recommendations or scores, black-box complaints, debugging model decisions, transparency requirements. Explainable AI makes decisions understandable.

Trustworthy & Reliable AI

Used by:

Plan Summary

aiux-plan-summary

Use when users should evaluate an agent's approach before it runs: 'why this plan', goal interpretation, subtask checklists, stated assumptions, strategy review before execution. Plan Summary exposes the reasoning so plans can be judged.

Trustworthy & Reliable AI

Used by:

Responsible AI Design

aiux-responsible-ai-design

Use when fairness, bias, or accountability is at stake: bias audits, 'could this harm or exclude someone', ethical review of AI decisions, accountability for automated outcomes. Responsible AI Design builds these checks into the lifecycle.

Trustworthy & Reliable AI

Used by:

Safe Exploration

aiux-safe-exploration

Use when users fear breaking something by trying the AI: sandbox or preview modes, 'let them experiment without consequences', undo-everything trials, test environments. Safe Exploration makes trying risk-free.

Trustworthy & Reliable AI

Used by:

Trust Calibration

aiux-trust-calibration

Use when users trust the AI too much or too little: over-reliance on wrong output, micromanaging a capable agent, per-domain track records, 'earn more autonomy over time'. Trust Calibration aligns perceived reliability with actual performance.

Trustworthy & Reliable AI

Used by:

Context Switching

aiux-context-switching

Use when users move between tasks, topics, or projects with an AI: 'it forgets when I change topics', resuming earlier threads, multiple parallel conversations, preserving context across switches. Context Switching maintains continuity.

Natural Interaction

Used by:

Conversational UI

aiux-conversational-ui

Use when designing chat or voice interaction: chatbots, assistants, message threads, composer design, 'talk to the app in plain language', turn-taking and context in conversation. Conversational UI makes dialogue feel natural instead of robotic.

Natural Interaction

Used by:

Multimodal Interaction

aiux-multimodal-interaction

Use when one input mode is not enough: combining voice, touch, text, images, or gesture, 'talk or type or point', accessibility through alternate modes, camera plus chat. Multimodal Interaction lets people interact the way the moment demands.

Natural Interaction

Used by:

Progressive Disclosure

aiux-progressive-disclosure

Use when a UI shows too much at once or advanced options bury the basics: overwhelming settings, cluttered toolbars, 'simplify this screen', show-more reveals, casual vs power users. Progressive Disclosure reveals complexity only as it is needed.

Natural Interaction

Used by:

Agent Status & Monitoring

aiux-agent-status-monitoring

Use when users need to follow long-running AI work without watching it: progress for background agents, 'what is the agent doing now', status badges, task dashboards, notifications when input is needed. Agent Status and Monitoring keeps users informed at the right level of attention.

Performance & Efficiency

Used by:

Intelligent Caching

aiux-intelligent-caching

Use when AI responses feel slow or repeated queries recompute: latency complaints, 'answers take too long', pre-fetching likely requests, caching common results. Intelligent Caching makes frequent AI results instant.

Performance & Efficiency

Used by:

Progressive Enhancement

aiux-progressive-enhancement

Use when quality output takes time: streaming rough answers that improve, 'users wait with no feedback', instant draft then detail, skeleton results while the model works. Progressive Enhancement gives something useful immediately and better progressively.

Performance & Efficiency

Used by:

Privacy-First Design

aiux-privacy-first-design

Use when handling user data in an AI product: consent, data minimization, 'what does the AI store about me', transparent privacy controls, trust concerns about collection. Privacy-First Design collects less and explains clearly.

Privacy & Control

Used by:

Selective Memory

aiux-selective-memory

Use when users should control what the AI remembers: memory settings, 'forget what I said earlier', viewing and deleting stored context, outdated or sensitive info resurfacing. Selective Memory makes remembering transparent and editable.

Privacy & Control

Used by:

Universal Access Patterns

aiux-universal-access-patterns

Use when the AI product must work for everyone: screen readers, low literacy, language diversity, motor differences, 'is this accessible', assistive technology support. Universal Access Patterns remove ability and language barriers.

Accessibility & Inclusion

Used by:

Anti-Manipulation Safeguards

aiux-anti-manipulation-safeguards

Use when users may disguise harmful requests: jailbreak attempts, 'it is just fiction or roleplay or hypothetical' framing, prompt manipulation, safety bypasses. Anti-Manipulation Safeguards detect real intent behind the framing.

Safety & Harm Prevention

Used by:

Crisis Detection & Escalation

aiux-crisis-detection-escalation

Use when users might express self-harm or crisis: suicide or harm signals in messages, 'what if a user says something alarming', hotline resources, escalation to humans. Crisis Detection and Escalation routes people to real help immediately.

Safety & Harm Prevention

Used by:

Session Degradation Prevention

aiux-session-degradation-prevention

Use when long conversations erode safety or quality: multi-hour chats, 'the AI gets too agreeable over time', boundary drift, session limits and refreshed checks. Session Degradation Prevention keeps late-session behavior as safe as the first message.

Safety & Harm Prevention

Used by:

Vulnerable User Protection

aiux-vulnerable-user-protection

Use when minors or at-risk users may be present: age gating, dependency risks, mental-health-sensitive contexts, 'what if a teenager uses this', graduated protections. Vulnerable User Protection detects vulnerability and adjusts safeguards.

Safety & Harm Prevention

Used by:

Want several at once?

Save patterns as you browse and download them as one pack from your dashboard. Not sure which patterns your product needs? Run the free audit. First time using skills? Learn how skills work.