UX Skills for Claude, ChatGPT, Gemini, and other AI agents
A curated selection of frameworks and patterns I use to design robust AI interactions and user experiences.
ai governors
Apply the Governors framework to design or audit human-in-the-loop features that keep users informed, in control, and safe as AI acts autonomously. Use this skill whenever someone is designing or reviewing AI product features involving oversight, trust, transparency, or control — including "how do I keep users in the loop", "how should I handle risky AI actions", "users don't trust the AI", "how do I prevent costly AI mistakes", "should I ask for confirmation before this action", "how do I show AI reasoning", "users are scared the AI will overwrite their data", "how do I handle AI memory and privacy", or any request about making an AI feature feel safe and controllable. Trigger even when the user doesn't say "governor" or "human-in-the-loop" — if they're designing any AI feature and the question touches on control, trust, transparency, cost, risk, or oversight, use this skill.
ai identifiers
Apply the AI Identifiers framework to design or audit the distinct, brand-level qualities that define how an AI presents itself across a product. Use this skill whenever someone is designing or reviewing the visual, verbal, or behavioral identity of an AI — including questions like "what should we call our AI", "how should our AI look", "what color should we use for AI features", "how do we make our AI feel distinct", "what icons should represent AI actions", "how do we give our AI a personality", "should our AI have an avatar", or any request about making an AI feel coherent, recognizable, and on-brand. Also trigger when the user is building a new AI feature and hasn't yet thought about how it should present itself — proactively raising identifiers as a design consideration is part of this skill's job.
ai inputs
Design and implement AI input patterns for products. Use this skill whenever the user wants to add an AI-powered input mechanism to their product, improve how users interact with AI features, decide which input pattern fits a use case, or audit existing AI input UX. Trigger on phrases like "how should users prompt this", "add AI input to", "let users control the AI with", "what input pattern should I use", "design an AI prompt experience", "how do I let users fill fields with AI", "add a regenerate button", "inline AI actions", or any request about how users should interact with or direct AI in the product. Always use this skill before designing or recommending any AI interaction surface.
ai trust builders
Apply AI Trust Builder design patterns to give users confidence that an AI product's results are ethical, accurate, and trustworthy. Use this skill whenever a designer, PM, or developer wants to make their AI product feel safer, more transparent, or more accountable. Trigger on: "make users feel safe", "add a disclaimer", "handle user data", "label AI-generated content", "privacy mode", "disclose AI is being used", "watermark AI outputs", "make the AI more transparent", "audit trail for AI", "user consent for recording", or any request touching AI accountability, privacy, explainability, or honest representation of what AI is doing. Also use when auditing an existing AI product for trust signals or when building new AI features into a non-AI-native product. Covers seven patterns: Caveat, Consent, Data Ownership, Disclosure, Footprints, Incognito Mode, and Watermark.
ai tuners
Apply AI Tuner design patterns when adding or improving AI features in a product. Tuners are the controls that let users shape how AI interprets input and produces output — before, during, or after generation. Use this skill whenever the user wants to add AI configuration UI to a product, improve how users control AI behavior, design prompt controls, model selectors, filters, style systems, voice/tone settings, or any mechanism that lets users influence what the AI does. Trigger on phrases like "let users control the AI", "add model switching", "prompt settings", "AI configuration", "let users set tone or style", "negative prompting", "AI filters", "mode switching", "AI parameter controls", or any request to give users more agency over AI output. This skill covers eight tuner patterns: Attachments, Connectors, Filters, Model Management, Modes, Parameters, Preset Styles, Saved Styles, and Voice & Tone.
ai wayfinders
Apply Wayfinder patterns to design or improve AI onboarding, discoverability, and first-interaction flows in any product. Use this skill whenever the user wants to add AI to a product surface, reduce blank-slate anxiety, help users discover what the AI can do, improve an initial CTA or prompt input, add suggestions or templates, design a gallery, add nudges, or generally reduce friction at the start of an AI interaction. Trigger even on vague requests like "make it easier to get started with AI", "users don't know what to type", "how do we show what the AI can do", "add some example prompts", or "improve onboarding to our AI feature". Wayfinders are: Initial CTA, Example Gallery, Suggestions, Templates, Nudges, Follow-ups, Prompt Details, and Randomize.
cognitive load conversion
Audit UI designs, flows, copy, and layouts to reduce cognitive load and maximize conversion. Apply this skill whenever a user shares a screen, mockup, flow, form, landing page, onboarding step, or any UI element and asks how to improve it — even if they don't say "cognitive load" or "conversion". Trigger on phrases like "why aren't users converting", "improve this flow", "reduce friction", "simplify this", "make this easier to use", "review this UI", "why do users drop off", "improve this form", "critique this design", "make this clearer", or any open-ended "improve this" request about a product surface. Always use this skill before giving UX or conversion improvement advice.
double diamond
Guide teams and individuals through the Double Diamond design thinking framework (Discover, Define, Develop, Deliver). Use this skill whenever someone mentions design process, product discovery, problem framing, ideation, prototyping, or user research — even without saying "Double Diamond". Trigger on: "where do I start with this design problem", "how do we approach building X", "help me structure our design process", "we need to do discovery", "we're in the ideation phase", "how do we validate our solution", "I have a design challenge", "I'm not sure what problem we're solving", "walk me through the design process", "help me run a design sprint". Also trigger when a non-designer or stakeholder wants to understand the design process or where they fit in. If someone describes a product problem without knowing how to approach it, proactively offer to guide them through this framework.
empathy mapping
Create, facilitate, and critique empathy maps for UX research and design thinking. Use this skill whenever a user wants to build an empathy map, understand their users more deeply, synthesize qualitative research into a shared team artifact, or translate user interviews into structured insights. Trigger on phrases like "create an empathy map", "map out what users think and feel", "help me understand my users", "synthesize these interviews", "what do my users say vs think", "build user empathy", or any request to structure user research into Says/Thinks/Does/Feels quadrants. Also trigger when users share raw interview transcripts, survey responses, or user research notes and want to make sense of them. Always use this skill before attempting to create any empathy map content from scratch.
feature prioritization
Apply structured prioritization matrix techniques to rank features, ideas, or design decisions by two weighted criteria (e.g. user impact vs. effort, feasibility vs. ROI). Use this skill whenever a user wants to prioritize features, compare design options, rank backlog items, decide what to build next, run a prioritization workshop, or make a structured UX or product decision. Trigger on phrases like "help me prioritize", "what should we build first", "rank these features", "should we focus on X or Y", "prioritize the backlog", "run a prioritization exercise", "impact vs effort", or any request to choose between competing options in a structured way.
journey mapping
Create, structure, and facilitate user journey maps from scratch or from existing research. Use this skill whenever a user wants to map a user experience, visualize a customer flow, identify pain points across a process, or build a journey map artifact. Trigger on phrases like "create a journey map", "map out the user experience", "visualize a user flow", "identify pain points in our process", "map the customer journey", "help me understand our user's experience", or any request involving understanding or documenting how a person moves through a product, service, or scenario — even if they don't say "journey map" explicitly. Also trigger when someone wants to understand the difference between journey maps, experience maps, service blueprints, or user story maps.
persuasive ux
Apply BJ Fogg's 7 Persuasive Technology Tools (Captology) to UX analysis and design. Use this skill whenever the user wants to improve a UI, flow, or feature using persuasion principles — including prompts like "how can I make this more engaging", "why aren't users completing this flow", "improve this onboarding", "make this CTA better", "reduce drop-off", "nudge users toward X", or any open-ended "improve this UI/UX" request. Also trigger when the user shares a screenshot, mockup, or describes a feature and wants design recommendations. Always use this skill before giving UX improvement advice — even if the user doesn't explicitly mention persuasion, Fogg, or Captology.
skill router
Route any UX, product, or AI design question to the right skill file. Use this skill first when the request is ambiguous or spans multiple skills, or when you need to identify which skill to apply. Acts as a decision tree across all 16 available skills.
ux heuristics review
Apply the 10 Usability Heuristics to critique existing UI or guide new product design. Use this skill whenever the user shares a screenshot, mockup, or written description of a feature or flow and wants UX feedback, a heuristic audit, design critique, or recommendations for a new product. Also trigger when the user asks things like "is this good UX?", "review this design", "what's wrong with this flow", "how should I design X", or "critique this UI". Always apply this skill before giving any UX or product design recommendations — even if the request seems simple.
ux personas
Create detailed, research-based UX personas for product and experience design. Use this skill whenever a user wants to create, improve, or critique personas — including requests like "help me define our users", "create a persona for X", "build user profiles for our product", "we need to define our target audience", or "turn our research into personas". Also trigger when a user shares user research, interview notes, survey data, or user segments and wants to turn that into design-ready personas. Always use this skill before attempting to write any persona content from scratch.
ux research methods
Guide selecting the right UX research method for a given situation. Use this skill whenever the user asks which research method to use, how to plan UX research, what research to do at a given product stage, how to study user behavior vs. attitudes, how to pick between qualitative and quantitative approaches, or whether to run interviews, usability tests, surveys, A/B tests, or any other UX research technique. Also trigger when the user describes a research question and wants a recommendation, or when they ask about the tradeoffs between specific methods. Trigger even if the user just says "what research should I do" or "how do I learn more about my users" without naming specific methods.
ux storyboard
Create UX storyboards from scratch, from user research, or from existing journey maps. Use this skill whenever a user wants to create a storyboard, visualize a user scenario, illustrate how a user interacts with a product, or communicate a UX story to a team or stakeholders. Also trigger when the user asks to "sketch a user flow", "show how a user would use X", "create a scenario illustration", "map out a use case visually", or wants to present research findings in a visual, narrative format. Even if the user doesn't say "storyboard" explicitly — if they want to show a sequence of steps a user takes, trigger this skill.
Hey 👋
My name is Tommy. Im a Product designer and developer from Copenhagen, Denmark.
Connected with me on LinkedIn ✌️