2.7K skills · sorted by relevance
Design and build backend services with API scaffolding, database migrations, authentication reviews, and load testing. Generate route handlers from OpenAPI specifications, identify PostgreSQL optimization opportunities, and compare architecture choices against traffic, data sensitivity, and reliability requirements.
Guides migration of existing codebases to Rust without unsafe code, covering architecture, implementation, testing, and output comparisons. It measures performance before and after optimization and produces an interactive HTML report documenting results and unresolved differences.
Assess Google Cloud workloads for environmental sustainability using the Well-Architected Framework. Get recommendations on lower-carbon regions, efficient computing and storage, AI energy use, and ongoing emissions measurement with Carbon Footprint.
Investigate failing or slow BigQuery jobs and unexpected cost increases. Compare execution details and resource usage with a normal baseline to isolate bottlenecks, identify likely causes, and determine practical remediation steps.
Learn how instructions, documents, tool results, and conversation history form an AI model’s context. Understand attention limits, why information in the middle can be missed, and how relevance and placement inform decisions about what to include.
Plan how an AI agent uses limited input space through reusable prompt prefixes, shorter tool outputs, summaries, and task splitting. Measure token use, response quality, and latency to choose optimizations that provide a net benefit.
Analyze ARIS usage logs to propose evidence-based changes to skill instructions, reviewer prompts, and workflow defaults. Receive a report and focused patch proposals, with selected patches staged for a separate application process.
Guide language model training with SimPO using pairs of preferred and rejected answers. Covers configuration, learning rate tuning, and memory troubleshooting for preference alignment, which adjusts model responses toward human preferences without a separate reference model.
Helps reduce GPU memory use by loading large language models in 4-bit or 8-bit formats, a technique called quantization. Covers bitsandbytes configuration, memory checks, QLoRA fine-tuning, and training with memory-saving optimizers.
Compress large language model weights to four bits with AWQ, a method that uses model activation patterns to limit quality loss. Follow workflows for quantization, loading compressed models, and serving them with vLLM using less GPU memory.
Compress large language models to 4-bit weights to reduce GPU memory needs and accelerate text generation. Covers loading compressed models, converting your own models, choosing settings, and preparing memory-efficient fine-tuning with transformers and PEFT.
Configure machine learning training and batch jobs across multiple clouds with SkyPilot. The skill guides GPU selection, lower-cost cloud and region choices, and automatic recovery when discounted spot instances are interrupted.
Write and refactor C++20/23 applications, addressing memory management, concurrency, and performance bottlenecks. Produce interfaces, implementations, tests, and relevant CMake updates, then check runtime errors and measure performance before applying targeted optimizations.
Design cloud architectures and migration plans for AWS, Azure, and GCP. Produce service diagrams, selection rationale, security designs, cost estimates, and recovery strategies, with deployment and rollback plans that account for availability and operational monitoring.
Develop microcontroller firmware for STM32 or ESP32, including peripheral drivers and FreeRTOS tasks. The skill guides interrupt handling, memory and power optimization, and checks that time-sensitive operations meet their deadlines.
Audit paid-ad landing pages for message consistency, mobile usability, accessibility, performance, forms, consent, tracking, and destination safety. Get evidence-based findings and recommendations ready for experiments to address obstacles to completing the intended action.
Diagnose paid advertising performance using account evidence, financial constraints, and experiments. Draft ranked changes to budgets, bids, or ad delivery, and apply supported operations only after approval and state checks, retaining a way to reverse them.
Audit C/C++ and Rust code for secrets left in memory. It checks missing or incomplete wipes, examines compiler output for removed cleanup operations, and produces reports with evidence and suggested fixes.
Apply Vercel Engineering’s performance guidelines when writing, reviewing, or refactoring React and Next.js code. Prioritize improvements to data fetching, downloaded code size, server processing, and unnecessary rendering using rules with explanations and code examples.
Screen an existing compound library against a protein target with Boltz. Rank candidates by binding confidence or lead optimization score, download predicted structures, and review solubility, permeability, and lipophilicity results alongside the ranked hits.
Measure website performance with Lighthouse, interpret scores and reports, and compare results across builds. Set limits for loading times and resource sizes, then integrate audits into CI, an automated workflow that checks code changes.
Create, audit, and improve pages for AI search engines by checking crawl access, metadata, clear answers, sources, and structured data that describes page content. It produces prioritized fixes and citation-friendly pages without guaranteeing inclusion in AI answers.
Improve how search engines discover and understand a website by reviewing titles, descriptions, links, and sitemaps. Add structured data, machine-readable descriptions of page content, and use checklists to identify technical and on-page SEO issues.
Write, review, and refactor React and Next.js code using performance guidelines from Vercel Engineering. Prioritize improvements to data fetching, downloaded code size, server processing, and unnecessary rendering, with rule files providing explanations and code examples.