Skill UI

This command guides the entire lifecycle of major feature development, providing a structured, multi-phase workflow. It helps users move a feature from initial concept (research) through planning, implementation, and progress tracking. Use it when you need to coordinate status, next steps, and phased delivery for complex features.

Analyzes a code repository to generate a recommended, minimal set of read-only shell command permissions for Claude Code's settings.json. It automatically detects the project's tech stack (e.g., Python, Node.js, Go), build tools, and monorepo structure, ensuring a safe and evidence-based security baseline for new projects or auditing existing permissions.

Transforms vague UI ideas into structured, Stitch-optimized prompts by adding platform context, UI/UX keywords, design system cues, and detailed page structure guidance so generated interfaces stay consistent and polished.

Provides guidelines for auditing and fixing HTML metadata, ensuring titles, descriptions, canonical URLs, Open Graph/Twitter tags, favicons, JSON-LD structured data, and robots directives stay consistent for shareable or SEO-sensitive pages.

Master common data transformations using functional programming patterns in TypeScript. This guide covers essential techniques for arrays, objects, and complex data structures, including mapping, filtering, reducing, grouping, and safely accessing nested values. It teaches how to replace verbose, imperative loops with concise, declarative functional code, improving readability and maintainability.

Automates the generation of production-ready Playwright end-to-end (E2E) tests. By accepting high-level inputs such as user stories, feature descriptions, or component names, this tool analyzes the codebase, selects appropriate testing templates (e.g., CRUD, login, checkout), and generates structured, best-practice test files. It ensures adherence to modern web testing standards, including proper locator strategies (getByRole), assertion techniques, and project conventions (fixtures, page objects).

Structured process for reviewing GitHub Actions workflows to identify exploitable vulnerabilities by tracing concrete attack paths, focusing on external attackers and evidence-backed findings, and ensuring every reported issue includes a reproducible exploitation scenario.

A lightweight integration providing full read/write access to Google Slides presentations. This tool allows users to programmatically create, inspect, modify, and extract content (including text and tables) from Google Workspace documents. Perfect for local automation tasks requiring full control over presentation structure and content without needing a dedicated server.

Hosted Agent Infrastructure orchestrates remote sandboxed execution environments so coding agents can run with unlimited concurrency, reproducibility, and shared state while minimizing session startup time.

Adds and manages structured evaluation data in Hugging Face model cards, extracting README tables, importing Artificial Analysis benchmarks, and running vLLM/lighteval evaluations to emit model-index metadata.

Sets up a new AgentHub collaboration session by scaffolding the .agenthub directory, generating session IDs, configuring evaluation criteria, and optionally capturing a baseline metric so teams can immediately spawn agents and run structured evaluations.

This assistant streamlines the workflow for managing GitHub issues using the GitHub CLI. It guides users through structured processes to create, list, and view issues. When creating, it prompts for detailed information (including issue type, reproduction steps, and expected vs. actual behavior for bugs), ensuring high-quality issue reports. It also allows filtering and listing of existing issues.