claude-plugins

Generates a complete, production-ready UI from a natural-language prompt — React + Tailwind, grounded in all 8 UIX reference books + WCAG 2.2, ROIN baseline always on. Asks 4 targeted questions then writes a self-contained .tsx file and compiles it to a browser-openable preview.html. Auto-activates on "design a [X] from scratch", "build me a UI for [X]", "create a full [X] page", "generate a UI for [X]", or /uix:generate-design.

This skill should be used when the user asks to "analyze this file", "check this code", "scan for violations", "find clean code issues", "review code quality", "check file length", "audit this project", "what's wrong with this code", or when a file looks messy or overly long. Scans files for clean code violations and outputs a grouped report of critical issues, warnings, and style suggestions.

This skill should be used when the user asks to "refactor this file", "clean up this code", "extract this function", "break this function apart", "pull this out into a method", "simplify this switch", "turn this into a table", "inline this", "replace magic number", or mentions a specific refactoring by name. Applies one named, surgical refactoring from the textbook catalog (Extract Method, Table-Driven, Replace Conditional with Polymorphism, etc.) and shows exactly what changed.

This skill should be used when the user asks to "clean my code", "make it human readable", "make this readable", "rewrite this cleanly", "refactor this", "clean up this function", "simplify this code", "make this code better", "fix code quality", "clean this file", "this code is messy", "improve this code", or when code has visible violations like long functions, deep nesting, poor names, or missing interfaces. Produces a clean version preserving all behavior, with reasoning for each change.

This skill should be used when the user asks to "check error handling", "find silent failures", "fix silent errors", "add input checks", "fail fast", "make this safer", "check for hidden errors", "defensive programming", or mentions empty catch blocks, swallowed exceptions, or missing validation. Detects and optionally fixes Rule 12 violations — silent error swallowing and missing guard clauses.

This skill should be used when the user asks to "suggest a better structure", "apply a design pattern", "this switch is too big", "simplify this class", "apply strategy pattern", "apply command pattern", "apply factory pattern", "fix this structure", "replace switch with polymorphism", or mentions restructuring a class or giant switch statement. Detects and optionally applies well-known design patterns (Strategy, Command, Factory, State) from OOP vs. Functional Programming.

This skill should be used when the user asks to "check my tests", "clean up my tests", "fix my tests", "are my tests clean", "review this test file", "AAA pattern", "tests should be clean", or mentions messy tests, test naming, or tests that have loops or conditionals. Detects and optionally fixes Rule 14 violations — tests missing the Arrange / Act / Assert structure, bad test names, or control flow inside tests.

This skill should be used when the user asks to "reduce coupling", "untangle this code", "fix method chains", "too many imports", "this file depends on too much", "long method chains", "reaching through objects", "law of demeter", or mentions tangled dependencies between files. Detects and optionally fixes Rule 11 violations — long method chains, high fan-out imports, and circular dependencies.

This skill should be used when the user asks to "set up clean code rules", "initialize clean code", "init cleancode", "add clean code to this project", "set up code standards", "configure clean code rules", or when a new project is opened and no .cleancode-rules.md file exists. Generates the canonical ruleset and platform config files so all AI assistants follow clean code throughout the project.

Extract misplaced code from any folder (backend or frontend) and move it to a properly-named folder that fits this project's existing conventions. Plan-first - reads the codebase to infer style, proposes a move-plan grounded in clean-code principles (Rule 15, Martin ch 11, McConnell ch 5), waits for a single confirmation, then moves files + rewrites imports atomically with git-aware history preservation. Handles backend layouts (services, repositories, controllers, route handlers), frontend layouts (components, pages, features, hooks, stores, ui/design-system), and mixed monorepos. Auto-chains from /cleancode:rewrite when a file split would benefit from cross-folder relocation. Triggers on phrases like "restructure my project", "move this code to a proper folder", "extract this from routes/", "my routes folder is doing too much", "where should this code live", "reorganize my folders", "split this across modules", "extract feature folder", "move business logic out of pages/components", "promote this shared cod

Audits a UI against WCAG 2.2 (Perceivable/Operable/Understandable/Robust), covering every AA criterion and all nine criteria new in 2.2 — focus not obscured, focus appearance, dragging movements, target size 24×24, consistent help, redundant entry, accessible authentication. Auto-activates on "accessibility audit", "WCAG check", "a11y review", "is this accessible", "check contrast", "keyboard accessible", or /uix:accessibility-audit. Produces a per-criterion pass/fail/NA report with fixes.

Recommends the right UI pattern from Tidwell's "Designing Interfaces" catalog for a given interaction problem — navigation, layout, lists/data, forms, controls, mobile. Auto-activates on "what pattern should I use for X", "how do I let users [do Y]", "what's the best way to show [Z]", or /uix:choose-pattern. Returns primary pattern, 1-2 alternatives, and why each fits or doesn't.

Builds a complete UI color palette — 8-10 grey shades, 1-2 primary hues with 5-10 shades each, accent colors for success/warning/danger/info — following Refactoring UI's HSL method plus Albers's relativity and Colour and Light's luminance rules. Auto-activates on "build a color palette", "design a color system", "pick UI colors", "how many shades do I need", "should I use this color", "dark mode palette", or /uix:color-system. Outputs palette with HSL values, contrast ratios, and usage rules.

Proposes a component design (button, card, modal, form, nav, table, dashboard widget, etc.) grounded in Tidwell's pattern catalog plus Refactoring UI's hierarchy/spacing/color/type rules. Auto-activates when the user says "design a component", "propose a [component] design", "how should I build a [X]", "spec a [X]", or invokes /uix:design-component. Returns pattern choice, states, spacing/type decisions, accessibility notes, and a reference implementation sketch.

Explains and applies UI/UX design principles — visual hierarchy, spacing systems, typography rules, Krug's laws of usability, Hick's/Fitts's/Miller's laws, satisficing, scent, Tidwell's user behaviors (safe exploration, satisficing, habituation, spatial memory), empty states, de-emphasis, action hierarchy, dark patterns. Auto-activates when the user asks "why does X feel wrong", "what's the principle behind Y", "how should I hierarchy / space / align this", or invokes /uix:design-principles. Grounded in Refactoring UI, Don't Make Me Think, Designing Interfaces, and UX for Beginners.

Re-scaffolds or customizes `.uix-config.md` for the uix plugin. Note that the plugin auto-creates this file via its SessionStart hook on first session — manual init is only needed to reset to defaults, force-overwrite, or walk through tuning options interactively. Auto-activates on "reset uix config", "re-init uix", "customize uix settings", "tune uix validator", "uix exclude this rule/path", or /uix:init.

Reviews an existing UI (file, URL, screenshot, code, or description) against the full UI_MASTER_GUIDE.md — hierarchy, spacing, typography, color, depth, forms, patterns, mobile, copy, and WCAG 2.2 accessibility. Auto-activates on "review my UI", "critique this design", "audit this page", "is this design good", "what's wrong with this screen", or /uix:review-ui. Produces a categorized findings list with severity, source citations, and fix recommendations.

Generates Mom-Test-compliant user interview plans, drafts screener questions, reviews planned questions for "bad data" signals (compliments/fluff/ideas), and analyzes interview notes to extract real signals (pain, commitments, workarounds) from noise. Auto-activates on "write interview questions", "plan user interviews", "review my questions", "analyze user interview notes", "talk to users", or /uix:user-research. Grounded in The Mom Test + UX for Beginners.

Explains distributed-systems and database principles — CAP, PACELC, consistency models (linearizable/causal/eventual), consensus (Raft/Paxos/quorums), ACID vs BASE, idempotency, Bloom/HyperLogLog/Count-Min/Merkle. Auto-activates when the user mentions these terms or asks "why does X happen in distributed systems". Grounded in DDIA ch. 5, 7, 9.

Proposes a database schema (SQL or NoSQL) for a given domain. Use when the user asks to "design a database", "model data for X", "build an ER", pick SQL vs NoSQL, or invokes /sdd:design-database. Grounded in DDIA and Kimball.

Produces a system design doc for a given requirement. Use when the user asks to "design a system", "architect a service", "design X at scale", or invokes /sdd:design-system. Grounded in DDIA and Fundamentals of Software Architecture.

Generates Mermaid, Excalidraw JSON, or DBML diagrams. Use when the user asks to "diagram X", "draw the architecture", "generate an ER diagram", "create a sequence diagram", or invokes /sdd:diagram. Accepts a free-text description with `--format` and `--save` flags.

Reviews an architecture (file, repo, design doc, or pasted text) against scalability, security, reliability, compliance, and anti-pattern checklists plus fitness functions. Use when the user asks to "review my architecture", "critique this design", "audit this system", or invokes /sdd:review-architecture.

This skill should be used when the user asks to "fix everything", "fix all violations", "clean this up", "auto-fix", "apply all fixes", "clean my code", "fix the whole project", or wants a one-shot fixer that runs every cleancode fixer in order across the codebase. Orchestrates analyze + each fixer skill (safety, untangle, test, refactor) in severity order across every source file, produces a full project-wide plan, asks for confirmation, and applies fixes.

This skill should be used when the user asks for "project health", "code health", "clean code score", "cleanliness score", "how bad is my code", "show metrics", "code quality dashboard", or wants a top-level view of the codebase's cleanliness. Produces a read-only dashboard of violation counts, worst offenders, and a 0-100 cleanliness score. Read-only by design — for fixing, use the specific skills.

This skill should be used when the user asks to "set up cursor rules", "add clean code to cursor", "configure codex", "write cursorrules", "set up AGENTS.md", "add rules for codex", "set up all platforms", "configure all AI assistants", "sync rules to cursor", or when clean code rules need to be written for a specific AI coding assistant platform. Generates platform-native config files from .cleancode-rules.md.

This skill should be used when the user asks to "explain this violation", "teach me about clean code", "why is this bad", "what's wrong with long functions", "explain single responsibility", "what does DRY mean", "teach me about interfaces", "explain this principle", "why do we need interfaces", "what is OOP", or when the user wants to understand the reasoning behind a clean code rule. Provides a clear explanation with a book citation, counter-example, and how to fix it.

This skill should be used when the user asks to "track to-fix items", "track technical debt", "scan for tech debt", "add a debt item", "what needs cleaning", "list known issues", "close a tech debt item", or mentions keeping a running list of cleanup work. Maintains a persistent .cleancode-todo.md file at the project root with entries for each known violation.