Skill UI

This skill performs comprehensive, feature-level integration validation after all associated development tasks are implemented. It goes beyond checking individual tasks to ensure cross-task consistency, full test suite passage, and complete specification coverage. It validates data flow, architectural adherence, and performs critical security scans for residual issues (e.g., TODOs, hardcoded secrets), providing a holistic GO/NO-GO assessment for deployment. Use this skill as a final quality gate to verify that the completed feature set respects the designed boundaries and dependencies end-...

This advanced code reviewer analyzes your codebase through the principles of 12 classic software engineering books (e.g., Clean Code, DDIA, Mythical Man-Month). It detects conceptual flaws, design smells, tight coupling, missing abstractions, and architectural risks—issues that standard linters and basic AI tools often miss. Essential for major refactoring and pre-design validation.

This skill integrates BuyWhere's product catalog surface, allowing AI agents to perform structured product searches, compare prices across various merchants, and discover deals. It is designed for developers building advanced e-commerce shopping agents, guiding them through the necessary API key setup and MCP configuration.

This skill generates a complete, standalone, and highly structured pricing page in standalone HTML. It automatically respects the active design system (colors, typography, components) and supports defining 2-4 plan tiers, including feature cards, comprehensive comparison tables, and an FAQ section. Ideal for product marketing and sales-focused websites.

Generates a complete, single-page SaaS landing page structure. This tool produces a fully contained HTML file, integrating required sections such as Hero, Features, Social Proof, Pricing, and Call-to-Action (CTA). It strictly adheres to an existing design system for colors, typography, and layout, making it ideal for product launches and marketing campaigns.

Squirrel is an advanced, full-cycle AI coding skill designed to guide software development from concept to production. It automatically detects the current project state (greenfield, in-progress, or mature) and adapts an 8-phase engineering pipeline. This comprehensive workflow covers planning, building, rigorous testing, bug hunting, code polishing, documentation, and deployment readiness, ensuring high-quality, maintainable results.

Cavecrew provides a set of specialized subagents (investigator, builder, reviewer) designed for efficient code development delegation. Unlike vanilla agents, Cavecrew compresses output results, drastically reducing the token usage injected back into the main context. This allows users to maintain context over much longer, complex coding sessions, effectively preventing context exhaustion during large-scale refactoring, debugging, or multi-step feature development.

This comprehensive, zero-trust security auditor scans third-party AI agent skills before installation. It performs a rigorous 6-phase review process, analyzing surface patterns, script integrity, permissions, and supply chain risks. It is designed to detect malicious payloads, credential theft, instruction overrides, and prompt injection attacks, ensuring secure AI development and deployment.

This skill generates comprehensive, tailored instruction files for AI coding assistants like GitHub Copilot and Claude. It supports multiple strategies (flat or nested) and is specifically designed for managing complex monorepos by creating area-scoped instructions using `applyTo` globs. Use it to standardize AI tooling usage across large codebases.

This skill provides a structured workflow for developing and modifying content management systems (CMS), applicable to platforms like WordPress, Shopify, and Drupal. It covers crucial architectural seams, including theme and template design, plugin/module development, content modeling, media asset pipelines, and static export processes, ensuring changes maintain structural integrity.

A deep-dive session designed to stress-test a technical plan against an existing domain model. This process forces the user to resolve ambiguities, sharpen terminology, and validate decisions by cross-referencing code and existing documentation (CONTEXT.md, ADRs). It ensures architectural rigor and completeness before implementation.

This pipeline orchestrates the process of adapting a highly polished, existing academic paper for resubmission to a new conference or journal. It specializes in text-only microedits, integrating prior reviewer feedback (review corpus) while strictly forbidding new experiments, new citations, or major structural changes. It is designed for high-stakes, time-constrained resubmissions.