Claude Skills by anubhavg-icpl

Production-ready Electron desktop app structure with main/renderer process separation and security best practices. Use when scaffolding, structuring, or architecting electron projects.

Expert in the Elegant design system - Graceful, refined aesthetic with delicate typography, minimal palettes, and polished layouts that exude sophistication. Use when building UI components, applying design tokens, or implementing visual styles for elegant.

Expert in the ElevenLabs design system - AI voice platform. Dark cinematic UI, audio-waveform aesthetics. Use when building UI components, applying design tokens, or implementing visual styles for elevenlabs.

ElevenLabs TTS, voice cloning, conversational AI, sound effects, music. Use when working with multimodal AI (images, audio, video) using elevenlabs.

Design the elicitation form schema and message template for a tool that needs mid-call user confirmation or disambiguation. Use when you need help with elicitation form designer.

Production-ready Elixir coding standards with OTP and Phoenix best practices. Use when enforcing elixir coding conventions and style rules.

Expert Elixir development with OTP, Phoenix, and functional programming mastery. Use when writing, reviewing, or refactoring elixir code.

Expert in Elasticsearch, Logstash, and Kibana for log management and analytics. Use when configuring, deploying, or managing elk stack infrastructure.

|. Use when you need help with email marketing.

Deep expertise in fine-tuning embedding models with sentence-transformers v3+ — SentenceTransformerTrainer, MultipleNegativesRankingLoss, Matryoshka, hard negatives mining. Use when implementing vector search, embeddings storage, or similarity queries with embedding fine tune.

Decision guide for picking embedding models in 2025-2026 — OpenAI, Cohere, Voyage, Jina, BGE, E5, Nomic, Stella, Snowflake Arctic; reading MTEB; trading dimension/cost/quality. Use when implementing vector search, embeddings storage, or similarity queries with embedding model picker.

Publish embedding models — sentence-transformers config, modules.json, 1_Pooling, MTEB submission, Matryoshka dims, embedding-specific model card. Use when creating, converting, or publishing model files with embedding model publish.

Pick embedding model, dimension, and retrieval mode for a given corpus and deployment. Use when you need help with embedding picker.

Inspect a word2vec model. Run analogies, find neighbors, diagnose quality. Use when you need help with embedding probe.

Audit a jailbreak-defense report across encoding-family attacks. Use when you need help with encoding audit.

Expert in the Energetic design system - Dynamic, vibrant style with thick borders, geometric shapes, high-contrast colors, and expressive typography conveying motion and vitality. Use when building UI components, applying design tokens, or implementing visual styles for energetic.

|. Use when you need help with eng runbook.

Pick a self-hosted LLM engine (llama.cpp, Ollama, TGI, vLLM, SGLang) given hardware, scale, and workload. Name 2026 TGI maintenance mode as a migration trigger. Use when you need help with engine picker.

Expert in the Enterprise design system - Clean, high-contrast enterprise design for data-driven workflows with intuitive drag-and-drop patterns and structured layouts. Use when building UI components, applying design tokens, or implementing visual styles for enterprise.

Design an entity linking pipeline — KB, candidate generator, disambiguator, evaluation. Use when you need help with entity linker.

Expert in Estuary Flow real-time CDC captures, collections, and materializations. Use when working with estuary flow for data processing, streaming, or analytics.

Expert in alignment, safety, red-teaming, and responsible AI deployment, from the AI Engineering from Scratch curriculum. Use when you need help with ethics safety alignment.

Design an LLM evaluation plan with calibrated judge and CI gates. Use when you need help with eval architect.

Plan a full generative-model evaluation: sample quality, adherence, preference, failure audit. Use when you need help with eval report.

Build a three-layer eval suite (static benchmarks, custom offline, online production) with evaluator-optimizer loop and CI gates. Use when you need help with eval suite.

Audit a proposed AlphaEvolve-style evolutionary coding loop's evaluator before committing any compute to the search. Use when you need help with evaluator rigor audit.

Expert in event-driven architecture, event sourcing, CQRS, and messaging patterns. Use when designing system architecture or making technical decisions about event driven architecture.

event-output. Use when you need help with event output.

Use when you have a written implementation plan to execute in a separate session with review checkpoints

Quantize and serve LLMs on consumer GPUs with ExLlamaV2/V3 (EXL2/EXL3), AWQ, and GPTQ. Use when deploying, running, or configuring local LLM inference with exllama awq gptq.

Expert in the Expo design system - React Native platform. Dark theme, tight letter-spacing, code-centric. Use when building UI components, applying design tokens, or implementing visual styles for expo.

Expert in the Expressive design system - Vibrant, personality-driven design with bold colors, playful graphics, and dynamic layouts that balance creativity with structure. Use when building UI components, applying design tokens, or implementing visual styles for expressive.

Fabrice Bellard — solo author of FFmpeg, QEMU, TCC, QuickJS; small, fast, complete, almost impossibly prolific. Use when you want code review, architecture advice, or opinions in the style of fabrice bellard.

Generate failure-mode detectors for agent traces, wired to a trace store, tagging the five industry-recurring modes plus domain-specific signatures. Use when you need help with failure detector.

Identify which fairness criterion a claim invokes and audit the associated assumptions. Use when you need help with fairness criterion.

fal.ai serverless inference for image/video models - queue + webhook patterns. Use when working with multimodal AI (images, audio, video) using fal ai.

Expert in the Fantasy design system - Game-inspired fantasy aesthetic with bold, premium visuals, rich color palettes, and immersive thematic elements. Use when building UI components, applying design tokens, or implementing visual styles for fantasy.

Expert in FastAPI async framework with Python type hints, Pydantic models, dependency injection, OAuth2, JWT auth, and production deployment. Use when you need deep expertise in fastapi.

Production-ready FastAPI project structure with async patterns, dependency injection, and modern Python practices. Use when scaffolding, structuring, or architecting fastapi projects.

Pick feature type, mel count, frame/hop, and normalization to match a downstream audio model. Use when you need help with feature extractor.

Expert in the Ferrari design system - Luxury automotive. Chiaroscuro editorial, Ferrari Red accents, cinematic black. Use when building UI components, applying design tokens, or implementing visual styles for ferrari.

|. Use when you need help with field notes editorial template.

Expert in the Figma design system - Collaborative design tool. Vibrant multi-color, playful yet professional. Use when building UI components, applying design tokens, or implementing visual styles for figma.

|. Use when you need help with finance report.

>. Use when you need help with find your level.

Evaluate fine-tuned LLMs — domain benchmarks, regression checks, catastrophic forgetting detection. Use when fine-tuning, training, or adapting language models with fine tune eval techniques.

Run a reproducible data-to-SFT-to-DPO-to-serve fine-tuning pipeline with ablations, quantization, and a 2026 Model Openness Framework model card. Use when you need help with finetuning pipeline.

Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup

Expert in FinOps for cloud financial management and cost optimization. Use when automating CI/CD, deployments, or operations with finops.

Design an LLM FinOps program — attribution schema (user/task/tenant + four token layers), three-tier enforcement ladder, and unit metric (cost per resolved / artifact). Use when you need help with finops plan.