Skill UI

This skill defines canonical, ordered middleware patterns for LangChain and LangGraph agents. It addresses critical production-level LLM deployment issues, including PII leakage across tenants, token budget overconsumption, prompt injection defense (Guardrails), and cache key corruption. Use this to build secure, scalable, and cost-aware AI applications.

This skill provides a comprehensive framework for security auditing LLM agents that possess external tools (file system access, command execution, network calls). It employs advanced dialogue probing techniques—such as command injection, path traversal, and SSRF attempts—to identify vulnerabilities where the agent might be induced to execute unexpected or malicious code, or access sensitive resources. Essential for red teaming and securing agent functionalities.

This document outlines strict, mandatory rules for AI agents regarding the invocation of specialized skills (Superpowers). It establishes that agents must check for and proactively use relevant skills in every response or operation, regardless of perceived simplicity. It details platform-specific tool usage and priority rules, ensuring user instructions always take precedence.

Provides comprehensive migration patterns for converting existing v5 Webiny code to the v6 architecture. It details key architectural shifts, including replacing ContextPlugins with Dependency Injection (DI) Services, updating event subscriptions to specialized EventHandler Features, and adopting the feature-based container registration model. Essential for developers and AI agents performing major framework upgrades.

This tool allows users to execute complex, multi-agent workflows defined in YAML format directly within the AI environment. It functions as an orchestration engine, sequentially invoking various specialized roles (agents) without requiring manual API key configuration. Ideal for structured tasks requiring collaboration between multiple personas.

This comprehensive guide serves as a decision flowchart for developers using Claude Code. It details when to use specific slash commands (/plan, /compact, /loop), select specialized agents (planner, code-reviewer, build-error-resolver), or invoke workflow skills (e2e-testing, tdd-workflow). It helps users select the optimal tool combination for complex tasks across the entire software development lifecycle, including feature planning, bug fixing, testing, and architectural design.

This skill analyzes your entire codebase to ensure it is optimized for AI contributions and developer tooling. It automatically generates essential configuration files such as `AGENTS.md`, Copilot instructions, CI workflows, and customizable issue templates. It learns from your existing PR review patterns to tailor the setup precisely to your project's stack. Use it when you need to prepare a repository for advanced AI-assisted development.

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 pulls the latest GEO-SEO components, including skills, agents, scripts, and schema templates, from the official upstream repository. It performs a detailed comparison (diff) to show all changes—new, modified, or removed—before applying the update. It ensures local customizations are preserved while keeping the toolkit current. Use this whenever you suspect your local setup might be outdated.

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.

The Desktop Commander MCP grants agents deep reach across the user's entire operating environment, extending far beyond the typical IDE sandbox. Key capabilities include maintaining persistent shell sessions, managing long-running background processes, accessing arbitrary filesystems, and natively handling structured data formats (Excel, PDF, DOCX). It also enables remote operations via SSH and large-scale content search, making it ideal for complex, multi-turn development, DevOps, and data analysis workflows.

This tool searches a comprehensive database of over 2500 curated open-source repositories for tools related to Large Language Models (LLMs), ChatGPT, RAG, agents, and NLP. It interprets natural language queries, converts non-English inputs to core English keywords, and systematically filters results across multiple specialized categories (e.g., LangChain, Chatbots, CLIs). Use it whenever the user needs to find specific AI development libraries or open-source AI tooling.