Skill UI

This skill analyzes the predictable ways language models lose performance as context length increases. It details core patterns, including 'lost-in-middle,' context poisoning, and distraction. Use this knowledge to diagnose unpredictable agent failures, design highly reliable systems for long document processing, and implement advanced context engineering strategies.

This skill analyzes existing codebases to create comprehensive, long-form technical documentation. It captures the architecture, design patterns, implementation details, and rationale (the 'why') behind complex systems. Ideal for creating definitive technical manuals, system overviews, and internal knowledge bases for onboarding and maintenance.

A comprehensive guide covering Internationalization (i18n) and Localization (L10n) best practices. It details how to build applications that can support multiple languages, regions, and text directions (RTL). Covers implementation patterns for React, Next.js, and Python, along with essential guidelines for handling date/number formats and managing translation keys.

A comprehensive guide for building professional, high-performance native iOS applications. Covers modern frameworks like SwiftUI and Swift 6, alongside deep expertise in UIKit, Core Data, networking (REST/GraphQL), performance profiling, and full App Store lifecycle management. Ideal for developers needing guidance on modern Apple ecosystem best practices.

This skill addresses the challenge of handling temporary CDN links for videos generated by Kling AI. It provides a comprehensive, multi-cloud solution to download the video asset and persistently upload it to major cloud storage platforms, including AWS S3, Google Cloud Storage (GCS), and Azure Blob Storage. It facilitates building robust, long-term media pipelines and ensures metadata preservation.

Extends transformer context windows using RoPE, YaRN, ALiBi, and position interpolation so long documents (32k-128k+ tokens) can be processed, pre-trained models lengthened, and efficient positional encodings or extrapolation strategies deployed.

A comprehensive guide to implementing robust, multi-layered API response caching using distributed stores like Redis and Memcached, alongside CDN integration. Learn advanced strategies such as tag-based cache invalidation, read-through/write-through patterns, TTL management, and serving stale-while-revalidate content to drastically boost application performance and scalability.

This skill provides methodologies for designing sophisticated, layered memory architectures for AI agents. It moves beyond simple context windows and basic vector stores by incorporating short-term, long-term, and graph-based memory systems. Implementing this is crucial for agents that require persistence across sessions, maintaining complex entity consistency, and performing advanced, relationship-driven reasoning over accumulated knowledge bases, including time-aware fact tracking.

Flash Attention optimizer delivers 2-4x speedup and 10-20x memory savings for transformer attention, ideal when training or running long-context models that hit GPU memory or latency limits; supports PyTorch native scaled_dot_product_attention, the flash-attn library, H100 FP8 precision, and sliding-window attention.

Conducts in-depth security audits across code, infrastructure, and configurations using the security-pro-pack toolkit to surface vulnerabilities, compliance gaps, and cryptographic weaknesses along with actionable remediation guidance.

This skill provides comprehensive, production-grade patterns for asynchronous programming in Rust using the Tokio runtime. It covers essential concepts like tasks, channels, and streams, along with robust error handling techniques. Use it when developing concurrent network services, building complex async applications, or optimizing performance in Rust.

RWKV is an RNN+Transformer hybrid architecture with linear-time inference, infinite context, and GPT-style training. It delivers constant-memory, streaming generation on CUDA and CPU, ideal for long-context, memory-constrained deployments.