UX CORE (UXC) — Bob

A Claude plugin that surfaces relevant cognitive biases when you're building user-facing products. Powered by 105 biases and 63 curated product questions from UX CORE by Wolf Alexanyan / KeepSimple.

What it does

Describe what you're building or the problem you're facing — and Bob tells you which psychological patterns apply, what to do about them, and what could backfire. With before/after visual demos.

Bob works two ways:

• Building something? (pricing page, onboarding, checkout, email) → Bob picks the most relevant biases and gives concrete actions
• Diagnosing a problem? ("why aren't our promotions working?", "users blame us for their mistakes") → Bob matches your problem to 63 curated product questions with pre-mapped bias recommendations

Install

Skills.sh (recommended):

npx skills add aram-m/uxcore

Claude Code (CLI):

claude plugin marketplace add github:aram-m/uxcore
claude plugin install uxcore

Claude.ai: Zip the skills/bob/ directory and upload via Settings > Skills > Upload Skill.

Usage

Invoke directly:

/bob I'm designing a cancellation flow for our SaaS

Or just describe a product/UX task — Bob auto-triggers on pricing, onboarding, checkout, copy, notifications, forms, A/B tests, and other user-facing work.

Repo Structure

uxcore/
  .claude-plugin/
    plugin.json              <- Plugin manifest
  skills/
    bob/
      SKILL.md               <- Bob's personality, workflow, and instructions
      references/
        question-index.md    <- Index of 63 curated product questions with keywords
        questions.md         <- Full question entries with pre-mapped biases and answers
        bias-index.md        <- Quick-reference index of all 105 biases
        biases.md            <- Full bias descriptions with product/UX applications
        demo-recipes.md      <- 89 before/after visual scenarios
  SPEC.md                    <- Design spec (historical)

Examples

Building something:

"I'm designing a pricing page with three tiers: $9, $29, $99/mo"

Bob responds with relevant biases (Anchoring, Decoy Effect, Contrast Effect...), each with a concrete action for your specific tiers, an ASCII or HTML before/after demo, and a "Watch out" section flagging risks.

Diagnosing a problem:

"Why aren't our promotions working? We've been running discounts but conversion is flat."

Bob matches this to a curated question, pulls pre-mapped biases (Curse of Knowledge, Conjunction Fallacy, Self-Reference Effect...) with contextual explanations of why each one applies to your situation, and adapts recommendations to your specific context.

Credits

Biases and question dataset sourced from UX CORE (UXC) by Wolf Alexanyan / KeepSimple.

License

MIT