Claude Skills by anubhavg-icpl

Expert in Kubernetes multi-tenancy patterns, virtual clusters, namespace isolation, and tenant workload management. Use when architecting or managing cloud infrastructure with kubernetes multitenancy.

Kubernetes operations including manifests, Helm charts, operators, troubleshooting, and resource management. Use when you need help with kubernetes operations.

Expert KQL assistant for live Azure Data Explorer analysis via Azure MCP server. Use when you need help with kusto assistant.

Expert in AWS Lambda 2025 patterns, SnapStart, Layers, and provisioned concurrency. Use when deploying to or building on lambda edge/serverless platform.

Expert in the Lamborghini design system - Supercar brand. True black surfaces, gold accents, dramatic uppercase typography. Use when building UI components, applying design tokens, or implementing visual styles for lamborghini.

Deep expertise in LanceDB — Lance columnar format, embedded + serverless modes, S3-backed tables, full-text search, versioning, and multimodal lakehouse. Use when implementing vector search, embeddings storage, or similarity queries with lancedb.

Expert in LangChain and LlamaIndex for building LLM-powered applications. Use when you need help with langchain llamaindex.

Self-hostable open-source LLM observability with tracing, scoring, datasets, and prompt management. Use when evaluating, monitoring, or observing LLM performance with langfuse.

Build stateful, durable agent graphs with checkpointing, human-in-the-loop, and time-travel debugging. Use when building AI applications with langgraph.

LangChain's hosted LLM observability and evaluation platform — traces, datasets, evaluators, hub. Use when evaluating, monitoring, or observing LLM performance with langsmith.

|. Use when you need help with last30days.

Jina's late chunking — embed long context first, then chunk the embeddings. Use when building or optimizing retrieval-augmented generation pipelines with late chunking.

Expert in Layer 2 scaling solutions, rollups, and blockchain performance. Use when building blockchain, DeFi, or Web3 applications with layer2 scaling.

Expert in modernizing legacy codebases with safe, incremental refactoring strategies. Use when you need help with legacy code modernizer.

Build stateful agents with persistent memory blocks and self-editing context using Letta (formerly MemGPT). Use when building AI applications with letta.

Expert in the Levels design system - Conversion-focused design that removes friction and guides users toward action through clarity, trust, and speed. Use when building UI components, applying design tokens, or implementing visual styles for levels.

Expert in the Linear design system - Project management. Ultra-minimal, precise, purple accent. Use when building UI components, applying design tokens, or implementing visual styles for linear app.

Expert in the Lingo design system - Playful, minimal design with bright colors, rounded shapes, tactile 3D borders, and friendly illustrations for approachable interfaces. Use when building UI components, applying design tokens, or implementing visual styles for lingo.

Run LiteLLM as a unified gateway over local + cloud LLMs with router config, virtual keys, budgets, fallbacks, and Redis caching. Use when deploying, running, or configuring local LLM inference with litellm proxy.

Expert in LitmusChaos - CNCF graduated Kubernetes-native chaos engineering platform. Use when writing, running, or improving tests with litmus chaos.

|. Use when you need help with live artifact.

|. Use when you need help with live dashboard.

Build, run, and tune llama.cpp for local LLM inference across CUDA, ROCm, Metal, Vulkan, and SYCL. Use when deploying, running, or configuring local LLM inference with llama cpp.

Run llama.cpp's HTTP server with OpenAI-compatible endpoints, slots, multimodal, and reverse proxies. Use when deploying, running, or configuring local LLM inference with llama cpp server.

Build and run Mozilla llamafile single-file LLM executables with Cosmopolitan Libc / APE. Use when deploying, running, or configuring local LLM inference with llamafile.

Build agentic document workflows, RAG, and query engines with LlamaIndex 2025. Use when building AI applications with llamaindex.

Run a 10-prompt vibes-eval on a LLaVA-family VLM and produce a human-readable scorecard. Use when you need help with llava vibes eval.

Token economics, prompt caching, model routing — engineering LLM apps for sustainable spend. Use when evaluating, monitoring, or observing LLM performance with llm cost.

Expert in putting LLMs to work in production applications, from the AI Engineering from Scratch curriculum. Use when you need help with llm engineering.

Expert in Large Language Model development, fine-tuning, and deployment. Use when you need deep expertise in llm.

LLM integration patterns including API usage, streaming, function calling, RAG pipelines, and cost optimization. Use when you need help with llm integration.

Build a self-hosted LLM observability dashboard that ingests OpenTelemetry GenAI spans, runs evals, and catches injected regressions in under five minutes. Use when you need help with llm observability.

Review an end-to-end LLM training pipeline manifest before a multi-million-dollar run. Use when you need help with llm pipeline reviewer.

Produce an LLM security plan covering secrets vault, PII scrubbing with consistent tokenization, network egress allowlist, audit log retention, and zero-trust posture. Use when you need help with llm security plan.

Expert in building, training, and understanding large language models end-to-end, from the AI Engineering from Scratch curriculum. Use when you need help with llms from scratch.

EleutherAI lm-evaluation-harness — MMLU, ARC, HellaSwag, GSM8K, IFEval, BBH benchmarks. Use when evaluating, monitoring, or observing LLM performance with lm eval harness.

Run LM Studio with the lms CLI, headless llmster daemon, REST API, and MLX backend on Apple Silicon. Use when deploying, running, or configuring local LLM inference with lm studio.

Design a realistic LLM load test — pick tool (LLMPerf, k6, GenAI-Perf, guidellm), build four patterns (steady, ramp, spike, soak), and gate in CI. Use when you need help with load test plan.

Wire local-only agentic stacks — Continue.dev, Cline, Aider, Open Interpreter, Goose — to Ollama, LM Studio, llama-server, and Jan. Use when deploying, running, or configuring local LLM inference with local agent runtime.

Build end-to-end local RAG with Chroma/LanceDB/Qdrant + nomic-embed/bge-m3/FastEmbed + llama-cpp-server or Ollama, all in Docker Compose. Use when deploying, running, or configuring local LLM inference with local rag stack.

Self-host LocalAI (mudler) as an OpenAI/Anthropic/ElevenLabs drop-in for LLMs, vision, audio, image and embeddings on any hardware. Use when deploying, running, or configuring local LLM inference with localai.

Expert in log analysis, pattern recognition, and debugging through log investigation. Use when diagnosing, troubleshooting, or fixing bugs with log analysis.

Mobile login and authentication flow screens. Use when you need help with login flow.

Design a long-context evaluation battery for a given model and use case. Use when you need help with long context eval.

Pick brute-context, ring-attention, token-compression, or agentic-retrieval for a long-video understanding task and compute latency + recall expectations. Use when you need help with long video strategy planner.

Package and publish LoRA adapters — HF Hub layout, vLLM dynamic loading, llama.cpp LoRA GGUF, Ollama ADAPTER directive, Replicate Cog. Use when creating, converting, or publishing model files with lora adapter publish.

Low-Rank Adaptation for parameter-efficient fine-tuning of LLMs. Use when fine-tuning, training, or adapting language models with lora techniques.

Expert in the Lovable design system - AI full-stack builder. Playful gradients, friendly dev aesthetic. Use when building UI components, applying design tokens, or implementing visual styles for lovable.

Expert Lua development for scripting, game development, and embedded systems. Use when writing, reviewing, or refactoring lua code.

Expert in the Luxury design system - High-end dark aesthetic with bold headings, monochromatic palette, and premium feel for luxury brand experiences. Use when building UI components, applying design tokens, or implementing visual styles for luxury.