108 件 ・ 関連度順
概要と使いどころ
Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.
日本語の概要は準備中です。原文の説明を表示しています。
Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace. Use when training large-scale models with limited compute (5× cost reduction vs dense models), implementing sparse architectures like Mixtral 8x7B or DeepSeek-V3, or scaling model capacity without proportional compute increase. Covers MoE architectures, routing mechanisms, load balancing, expert parallelism, and inference optimization.
日本語の概要は準備中です。原文の説明を表示しています。
Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use when optimizing inference speed (1.5-3.6× speedup), reducing latency for real-time applications, or deploying models with limited compute. Covers draft models, tree-based attention, Jacobi iteration, parallel token generation, and production deployment strategies.
日本語の概要は準備中です。原文の説明を表示しています。
Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use when optimizing inference speed (1.5-3.6× speedup), reducing latency for real-time applications, or deploying models with limited compute. Covers draft models, tree-based attention, Jacobi iteration, parallel token generation, and production deployment strategies.
日本語の概要は準備中です。原文の説明を表示しています。
Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace. Use when training large-scale models with limited compute (5× cost reduction vs dense models), implementing sparse architectures like Mixtral 8x7B or DeepSeek-V3, or scaling model capacity without proportional compute increase. Covers MoE architectures, routing mechanisms, load balancing, expert parallelism, and inference optimization.
日本語の概要は準備中です。原文の説明を表示しています。
Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.
日本語の概要は準備中です。原文の説明を表示しています。
Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.
日本語の概要は準備中です。原文の説明を表示しています。
Achieve aggressive v3 performance targets: 2.49x-7.47x Flash Attention speedup, 150x-12,500x search improvements, 50-75% memory reduction. Comprehensive benchmarking and optimization suite.
日本語の概要は準備中です。原文の説明を表示しています。
Image SEO audit — make a site's images discoverable in Google Images and stop them from dragging down Core Web Vitals. Audits alt text quality and coverage, descriptive file names, modern formats (WebP/AVIF), dimensions and compression, responsive srcset, lazy-loading, explicit width/height (CLS), image sitemaps, and ImageObject structured data. Use this skill whenever the user asks about image SEO, alt text, Google Images ranking, image optimization for search, image file size hurting page speed, WebP/AVIF, or image structured data. Trigger on: "image SEO", "alt text", "rank in Google Images", "optimize my images for search", "image alt tags", "WebP", "image compression SEO", "image sitemap", "ImageObject schema", or any image-discoverability / image-performance question. For overall page speed use /seo-analysis; for generating new images use the brand image skills.
日本語の概要は準備中です。原文の説明を表示しています。
Search Experience Optimization (SXO), the bridge between SEO and UX/CRO. Audits the journey from the SERP click to the on-page goal: click-through factors (title, meta, rich results), then post-click signals that keep users and drive conversions: above-the-fold relevance and intent match, page speed / Core Web Vitals, readability, clear CTAs, mobile usability, and what causes pogo-sticking back to Google. Use when the user has decent rankings but poor CTR or poor on-page conversion, wants better dwell time / engagement, less bounce, or a page experience aligned with search intent. Trigger on: "SXO", "search experience optimization", "good rankings but low CTR", "high bounce rate from search", "pogo-sticking", "improve dwell time", "SEO and conversion", "ranking but not converting", "engagement signals", "make this page convert search traffic". For pure landing-page conversion in ads use /google-ads-landing; for content quality use /seo-page.
日本語の概要は準備中です。原文の説明を表示しています。
Guided journey from a leaking conversion flow - landing page, signup, checkout, or in-app onboarding - to a measured, tested funnel. Orchestrates six skills phase by phase - lean-analytics, cro-methodology, storybrand-messaging, hundred-million-offers, influence-psychology, design-everyday-things - each phase carries its full method inline so it runs standalone, asking the user questions at every decision point and recording results in the project docs/ folder (FUNNEL.md, METRICS.md, EXPERIMENTS.md, CONVERSION-OPTIMIZATION-PLAN.md) so the journey resumes across sessions. Use when the user wants to raise conversion on a specific flow, fix checkout or signup abandonment, diagnose onboarding drop-off, or says 'people start but never finish'. With no site yet, use create-website; for a whole-site look, message, and speed audit, improve-website; if the flow converts but needs traffic, grow-website; for in-product engagement and retention, improve-app. For one framework in isolation, invoke that skill directly.
日本語の概要は準備中です。原文の説明を表示しています。
Guided journey from a working codebase grown slow and tangled to one measurably fast, cleanly bounded, and readable. Orchestrates eight skills phase by phase - working-with-legacy-code, clean-architecture, software-design-philosophy, refactoring-patterns, system-design, ddia-systems, release-it, pragmatic-programmer - every phase carries its method inline so it runs standalone, asking the user questions at every decision point and recording results in the project docs/ folder (PERFORMANCE.md, ARCHITECTURE.md, ARCHITECTURE-OPTIMIZATION-PLAN.md) so the journey resumes across sessions. Use when the user wants to make an app faster, untangle drifted boundaries, fix slow endpoints and queries, or says 'it works but it is slow and getting worse'. For an untested prototype, improve-code-quality; for an aged codebase you fear to touch, remove-technical-debt; for greenfield structure, design-code-architecture; for marketing-site page speed, improve-website. For one framework in isolation, invoke that skill directly.
日本語の概要は準備中です。原文の説明を表示しています。
Analyze Swift and mixed-language compile hotspots using build timing summaries and Swift frontend diagnostics, then produce a recommend-first source-level optimization plan. Use when a developer reports slow compilation, type-checking warnings, expensive clean-build compile phases, long CompileSwiftSources tasks, warn-long-function-bodies output, or wants to speed up Swift type checking.
日本語の概要は準備中です。原文の説明を表示しています。
Theme speed and UX optimization — Core Web Vitals, Liquid code, image loading, mobile responsiveness
日本語の概要は準備中です。原文の説明を表示しています。
Achieve aggressive v3 performance targets: 2.49x-7.47x Flash Attention speedup, 150x-12,500x search improvements, 50-75% memory reduction. Comprehensive benchmarking and optimization suite.
日本語の概要は準備中です。原文の説明を表示しています。
Achieve aggressive v3 performance targets: 2.49x-7.47x Flash Attention speedup, 150x-12,500x search improvements, 50-75% memory reduction. Comprehensive benchmarking and optimization suite.
日本語の概要は準備中です。原文の説明を表示しています。
Achieve aggressive v3 performance targets: 2.49x-7.47x Flash Attention speedup, 150x-12,500x search improvements, 50-75% memory reduction. Comprehensive benchmarking and optimization suite.
日本語の概要は準備中です。原文の説明を表示しています。
Safely diagnose and improve local network speed, latency, jitter, DNS, Wi-Fi, Ethernet, macOS network services, and bufferbloat. Use this skill whenever the user asks to optimize internet/network speed, make the network faster, diagnose slow Wi-Fi, latency, packet loss, DNS delay, unstable Codex/AI tool connectivity, or asks about the viral Codex network optimization workflow. Always protect VPN/proxy tools such as Clash Verge, Mihomo, Shadowrocket, Tailscale, V2Ray, Surge, and corporate VPNs; do not modify or disable them unless the user explicitly asks for that specific change.
日本語の概要は準備中です。原文の説明を表示しています。
Performance optimization mindset - knowing when to optimize, how to measure, where bottlenecks hide, and when "fast enough" is the right answerUse when "slow, performance, optimize, profiling, benchmark, latency, throughput, cache, n+1, bottleneck, memory leak, too slow, speed up, response time, performance, optimization, profiling, caching, latency, throughput, big-o, benchmarking" mentioned.
日本語の概要は準備中です。原文の説明を表示しています。
Use when reducing model size, improving inference speed, or deploying to edge devices - covers quantization, pruning, knowledge distillation, ONNX export, and TensorRT optimizationUse when ", " mentioned.
日本語の概要は準備中です。原文の説明を表示しています。
Deep technical SEO auditing covering crawlability, indexation, Core Web Vitals, site architecture, canonicalization, structured data validation, and performance optimization. Use when the user asks about technical SEO, site audits, crawl issues, indexation problems, Core Web Vitals, page speed, or site architecture optimization. For content-focused SEO, see the seo-content-writer skill.
日本語の概要は準備中です。原文の説明を表示しています。
Achieve aggressive v3 performance targets: 2.49x-7.47x Flash Attention speedup, 150x-12,500x search improvements, 50-75% memory reduction. Comprehensive benchmarking and optimization suite.
日本語の概要は準備中です。原文の説明を表示しています。
Comprehensive SEO analysis for any website or business type. Performs full site audits, single-page deep analysis, technical SEO checks (crawlability, indexability, Core Web Vitals with INP), schema markup detection/validation/generation, content quality assessment (E-E-A-T framework per Dec 2025 update extending to all competitive queries), image optimization, sitemap analysis, and Generative Engine Optimization (GEO) for AI Overviews, ChatGPT, and Perplexity citations. Analyzes AI crawler accessibility (GPTBot, ClaudeBot, PerplexityBot), llms.txt compliance, brand mention signals, and passage-level citability. Industry detection for SaaS, e-commerce, local business, publishers, agencies. Triggers on: "SEO", "audit", "schema", "Core Web Vitals", "sitemap", "E-E-A-T", "AI Overviews", "GEO", "technical SEO", "content quality", "page speed", "structured data".
日本語の概要は準備中です。原文の説明を表示しています。
"Optimize Core Web Vitals (LCP, INP, CLS) for better search rankings and user experience. Use this skill when the user needs to diagnose page speed issues, improve Largest Contentful Paint, reduce layout shift, or pass Google's page experience signals — even if they say 'my site is slow', 'Core Web Vitals failing', or 'page speed optimization'.".
日本語の概要は準備中です。原文の説明を表示しています。