Skill UI

This guide outlines a systematic workflow to eliminate copy-paste duplication (code clones) detected by tools like jscpd. It teaches developers how to refactor repeated code into shared functions, modules, constants, or base classes. Use this process when your codebase exhibits significant redundancy to improve maintainability, reduce technical debt, and enhance overall code quality.

This meta-skill scans a target codebase path, intelligently determines the most appropriate architectural diagram type (e.g., sequence, component, class), and concurrently generates the diagram in two standardized formats: PlantUML source code and draw.io XML. Finally, it composes these deliverables into a cohesive architecture document (DOCX). Ideal for creating comprehensive onboarding guides or technical RFCs that require both text-friendly and highly editable views of the system architecture.

Analyzes any frontend codebase (React, Vue, Svelte, Angular, plain CSS, Tailwind) to document the complete design system. It extracts crucial visual elements—including color palettes, typography, spacing scales, and component patterns—directly from source files, without needing to build, render, or run the application. Ideal for design audits, visual language documentation, or migrating a project's identity into a design tool.

Unship is a dedicated local workflow for front-end developers to compare multiple AI-generated UI alternatives (such as layouts, copy, or flows) directly within a running application environment. Instead of committing to one generated version, it adds temporary source-level variants, allowing for visual comparison. After the user selects the final version, the skill automatically cleans up all unused temporary code and attributes, ensuring a clean, production-ready codebase. Note: This tool is for local iteration only, not for production A/B testing or analytics.

This skill designs and audits the core engineering harness required for reliable AI agent collaboration within a codebase. It establishes critical infrastructure elements such as change tracking, linting rules, CI/CD gates, and standardized documentation, ensuring the repository remains the single source of truth for all intelligent agents. Ideal for setting up complex development workflows.

This guide helps developers audit and clean up Vite + React projects initialized using the Lovable scaffolding tool. It systematically removes all boilerplate artifacts, including placeholder assets, dedicated dependencies (like lovable-tagger), generated documentation, and unnecessary components (e.g., unused Radix deps). The goal is to deliver a clean, fully owned, and production-ready codebase.

This comprehensive guide details the idiomatic and efficient way to write Scala code. It covers crucial modern practices, such as utilizing functional transforms over traditional loops, embracing pattern matching, using Option/Either for robust error handling instead of null checks or exceptions, and managing concurrency safely. Essential reading for any Scala developer aiming for high-quality, maintainable, and performant codebases.

This skill generates a comprehensive and structured developer onboarding document for any service or codebase. It guides new engineers through the entire product lifecycle, covering architectural overview, local development setup, key coding patterns, testing strategies, CI/CD deployment processes, and team contacts, ensuring rapid productivity from day one.

An adversarial Quality Assurance (QA) pass designed to deep-dive into any codebase for logic errors, security flaws, and functional bugs. Instead of confirming that code 'looks right,' this skill actively attempts to break the code by simulating hostile inputs, tracing critical data paths, and checking for subtle, real-world correctness issues. Essential for high-stakes, critical code reviews.

A comprehensive guide and shared vocabulary for designing and structuring complex software modules. Use this when you need to refine an existing module's interface, identify opportunities for deepening its capabilities, determine optimal internal seams, or improve code testability and AI-navigability. It establishes clear principles for Module, Interface, Depth, and Seam.

This skill outlines a systematic, multi-phase workflow for diagnosing hard bugs and performance regressions in complex codebases. It emphasizes the critical necessity of building a tight, deterministic feedback loop (Phase 1) before moving to hypothesis generation. The process guides users through reproducing the failure, minimizing the scope of the issue, and forming actionable, ranked hypotheses.

This skill guides the systematic triage of project issues and external pull requests (PRs). It moves development items through a defined state machine using roles (e.g., bug, enhancement) and states (e.g., ready-for-agent, wontfix). The process involves context gathering, cross-codebase verification, deep questioning (grilling), and generating comprehensive, agent-ready briefs or marking them for human implementation.