Claude Skills

High-performance LLM inference with vLLM, quantization (AWQ, GPTQ, FP8), speculative decoding, and edge deployment. Use when optimizing inference latency, throughput, or memory.

HyDE (Hypothetical Document Embeddings) for improved semantic retrieval. Use when queries don't match document vocabulary, retrieval quality is poor, or implementing advanced RAG patterns.

Use this skill for internationalization (i18n) in React applications. Covers ALL user-facing strings, date/time handling, locale-aware formatting (useFormatting hook), ICU MessageFormat, Trans component, and RTL/LTR support.

Idempotency patterns for APIs and event handlers. Use when implementing exactly-once semantics, deduplicating requests, or building reliable distributed systems.

LLM observability platform for tracing, evaluation, prompt management, and cost tracking. Use when setting up Langfuse, monitoring LLM costs, tracking token usage, or implementing prompt versioning.

LangGraph checkpointing and persistence. Use when implementing fault-tolerant workflows, resuming interrupted executions, debugging with state history, or avoiding re-running expensive operations.

LangGraph Functional API with @entrypoint and @task decorators. Use when building workflows with the modern LangGraph pattern, enabling parallel execution, persistence, and human-in-the-loop.

LangGraph human-in-the-loop patterns. Use when implementing approval workflows, manual review gates, user feedback integration, or interactive agent supervision.

LangGraph parallel execution patterns. Use when implementing fan-out/fan-in workflows, map-reduce over tasks, or running independent agents concurrently.

Message queue patterns with RabbitMQ, Redis Streams, and Kafka. Use when implementing async communication, pub/sub systems, event-driven microservices, or reliable message delivery.

Use this skill for implementing Motion (Framer Motion) animations in React applications. Covers animation presets, page transitions, modal animations, list stagger effects, hover interactions, skeleton loaders, and RTL-aware animation patterns.

Mock Service Worker (MSW) 2.x for API mocking. Use when testing frontend components with network mocking, simulating API errors, or creating deterministic API responses in tests.

Multi-agent coordination and synthesis patterns. Use when orchestrating multiple specialized agents, implementing fan-out/fan-in workflows, or synthesizing outputs from parallel agents.

CLIP, SigLIP 2, Voyage multimodal-3 patterns for image+text retrieval, cross-modal search, and multimodal document chunking. Use when building RAG with images, implementing visual search, or hybrid retrieval.

Local LLM inference with Ollama. Use when setting up local models for development, CI pipelines, or cost reduction. Covers model selection, LangChain integration, and performance tuning.

Transactional outbox pattern for reliable event publishing. Use when implementing atomic writes with event delivery, ensuring exactly-once semantics, or building event-driven microservices.

OWASP Top 10 security vulnerabilities and mitigations. Use when conducting security audits, implementing security controls, or reviewing code for common vulnerabilities.

Automated accessibility testing with axe-core, Playwright, and jest-axe for WCAG compliance. Use when adding or validating a11y tests, running WCAG checks, or auditing UI accessibility.

Performance and load testing with k6 and Locust. Use when validating system performance under load, stress testing, identifying bottlenecks, or establishing performance baselines.

Production hybrid search combining PGVector HNSW with BM25 using Reciprocal Rank Fusion. Use when implementing hybrid search, semantic + keyword retrieval, vector search optimization, metadata filtering, or choosing between HNSW and IVFFlat indexes.

Unit testing patterns and best practices. Use when writing isolated unit tests, implementing AAA pattern, designing test isolation, or setting coverage targets for business logic.

VCR.py HTTP recording for Python tests. Use when testing Python code making HTTP requests, recording API responses for replay, or creating deterministic tests for external services.

Reranking patterns for improving search precision. Use when implementing cross-encoder reranking, LLM-based relevance scoring, or improving retrieval quality in RAG pipelines.

Production-grade fault tolerance for distributed systems. Use when implementing circuit breakers, retry with exponential backoff, bulkhead isolation patterns, or building resilience into LLM API integrations.