Skill UI

Operates as an end-to-end GitHub repository steward, triaging issues, reviewing PRs, analyzing contributor activity, and generating maintenance guidance so the human can simply ask “what’s next” while the agent keeps UX, docs, and backlog healthy.

An expert skill designed for writing, reviewing, and refactoring SwiftUI code for iOS and macOS. It provides deep insights into modern state management (@Observable), optimizing view composition, handling data flow, and improving performance. The skill also assists with interpreting Instruments traces and adhering to Apple's latest Human Interface Guidelines.

This comprehensive guide outlines a rigorous, human-in-the-loop engineering workflow for leveraging advanced AI agents to complete complex implementation tasks. It mandates defining clear completion criteria, decomposing work into manageable units, and strategically routing models (Haiku, Sonnet, Opus) based on task complexity. The workflow emphasizes continuous evaluation, regression testing, security review of invariants and edge cases, and strict cost discipline to ensure robust and reliable software development using AI assistance.

This skill provides a comprehensive framework for designing and reviewing native iOS applications. It covers adherence to Apple's Human Interface Guidelines (HIG), ensuring consistency across layout, typography, safe areas, and system components (e.g., Dynamic Island). Use this when building or reviewing SwiftUI/UIKit code to guarantee the app feels intuitive, highly polished, and genuinely native to the Apple ecosystem.

A powerful code review utility designed for SwiftUI projects. It comprehensively analyzes code quality, ensuring adherence to modern Swift APIs, optimal performance, robust data flow management, and compliance with Apple's Human Interface Guidelines and accessibility standards (like VoiceOver and Dynamic Type). Use this tool when writing, refactoring, or reviewing complex SwiftUI implementations to maintain high code hygiene and maintainability.

End-to-end autonomous research workflow covering idea discovery, experiments, review, and paper submission. Supports resumable runs, optional Lean formalization, and configurable checkpoints for human oversight.

Vibers provides a human code review workflow for AI-generated GitHub projects; add the collaborator, GitHub Action, and commit guidance to receive spec- and security-focused feedback plus follow-up PR fixes.

This skill involves creating structured specifications before writing any code to serve as a shared source of truth between developers and engineers. It is used when starting new projects, features, or significant changes where requirements are unclear or ambiguous. The workflow includes a scope check, specifying objectives, planning, task breakdown, and implementation with human reviews at each phase. It helps decompose complex requirements into capability maps and ensures clarity before coding begins.

This meta-skill automates the complex process of modernizing legacy PDF documents (e.g., historical contracts, scanned manuals, legal documents). It follows a structured four-step pipeline: structural extraction, natural-language rewriting of problem pages, generating an audit summary, and finally re-merging the content into a single, refined PDF. This ensures content is modernized while maintaining fidelity to the original meaning, providing a human review summary before finalization.

Rewrites overly formal or vague code review comments generated by AI or corporate tools. It cuts out filler phrases like "I noticed..." or "It might be worth considering...", transforming them into direct, actionable feedback. The output uses a clear format (Location: Severity: Observation. Fix.) that sounds like it came from a helpful, experienced teammate, significantly improving the quality and reception of PR feedback.

This comprehensive guide outlines a rigorous, human-in-the-loop engineering workflow for leveraging advanced AI agents to complete complex implementation tasks. It mandates defining clear completion criteria, decomposing work into manageable units, and strategically routing models (Haiku, Sonnet, Opus) based on task complexity. The workflow emphasizes continuous evaluation, regression testing, security review of invariants and edge cases, and strict cost discipline to ensure robust and reliable software development using AI assistance.

A formal framework designed for evaluating AI-assisted code generation sessions. It implements Development Oriented Evals (EDD) principles, allowing users to define capability and regression criteria. The system supports multiple evaluation types—code-based, model-based, and human review—and tracks reliability using advanced metrics like pass@k, ensuring robust quality control for AI agents.