Skill UI

Provides WCAG 2.1 based accessibility audit guidance for web interfaces, covering perceivable, operable, understandable, and robust principles along with code snippets for alt text, contrast, keyboard, focus, and media alternatives.

This skill automates the setup and configuration of Ollama, allowing users to deploy Large Language Models (LLMs) locally. It performs system assessment (OS, GPU, RAM), selects optimal models based on hardware availability, and handles installation across macOS, Linux, and Docker. Use this when you require free, self-hosted AI alternatives, need offline inference capabilities, or wish to eliminate reliance on costly hosted API services.

Systematically examines project dependencies, using GitHub data to highlight single-maintainer, unmaintained, unpopular, or otherwise risky packages before audits so teams can focus security efforts and propose better-maintained alternatives.

This skill helps developers structure and record critical architectural decisions (ADRs) made during coding sessions. It automatically detects moments of decision, capturing the context, alternatives considered, and the rationale (trade-offs and consequences). This maintains a living log that explains 'why' the code base was built the way it is, preventing knowledge loss and improving project maintainability.

This skill acts as an expert guide for the Go ecosystem, recommending production-ready libraries and frameworks. It follows best practices by prioritizing the standard library first. Use it when you need to compare alternatives, select the best library for a specific task, or add a new dependency to your Go project.

A comprehensive guide for migrating search functionality from alternatives like Elasticsearch, Typesense, or Meilisearch to Algolia. It details the necessary steps, including data migration, query translation (e.g., ES to Algolia), and implementing a zero-downtime swap using the Strangler Fig pattern.

This skill helps users discover alternatives, equivalents, or related tools/models/datasets for a specified Japanese NLP resource (GitHub repo, Hugging Face model, or tool). It first checks a bundled dataset and then expands the search via web research across GitHub and Hugging Face, providing a comprehensive list of related resources.

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.

A comprehensive skill for generating a structured Request for Comments (RFC) document. Use this to formally propose technical decisions, architectural changes, or significant system implementations. The output ensures all critical aspects are covered, including problem statements, proposed solutions, rejected alternatives, success metrics, and migration plans, facilitating rigorous technical review and decision-making.