Claude Skills

LangGraph supervisor-worker pattern. Use when building central coordinator agents that route to specialized workers, implementing round-robin or priority-based agent dispatch.

LangGraph state management patterns. Use when designing workflow state schemas, using TypedDict vs Pydantic, implementing accumulating state with Annotated operators, or managing shared state across nodes.

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

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

Graph-first memory orchestration - knowledge graph (PRIMARY, always available) with optional mem0 cloud enhancement for semantic search. Use when designing memory orchestration or combining graph and mem0.

Detect quiet failures in LLM agents - tool skipping, gibberish outputs, infinite loops, and degraded quality. Use when agents appear to work but produce incorrect results.

Long-term semantic memory across sessions using Mem0. Use when you need to remember, recall, or forget information across sessions, or when referencing what we discussed last time or in a previous session.

Use when planning system architecture to ensure nothing is missed. Provides structured questions covering scalability, security, data, and operational dimensions before implementation.

TAM/SAM/SOM market sizing, Porter's Five Forces, competitive analysis, and SWOT frameworks. Use when sizing market opportunities, analyzing competition, or assessing industry dynamics.

Transitional intro cards between video scenes. Use when adding "Coming Up Next" cards, scene transitions, or visual breathing room in demos

Use when conversation context is too long, hitting token limits, or responses are degrading. Compresses history while preserving critical information using anchored summarization and probe-based validation.

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

Provider-native prompt caching for Claude and OpenAI. Use when optimizing LLM costs with cache breakpoints, caching system prompts, or reducing token costs for repeated prefixes.

Security patterns for LLM integrations including prompt injection defense and hallucination prevention. Use when implementing context separation, validating LLM outputs, or protecting against prompt injection attacks.

LLM streaming response patterns. Use when implementing real-time token streaming, Server-Sent Events for AI responses, or streaming with tool calls.

Testing patterns for LLM-based applications. Use when testing AI/ML integrations, mocking LLM responses, testing async timeouts, or validating structured outputs from LLMs.

LLM fine-tuning with LoRA, QLoRA, DPO alignment, and synthetic data generation. Efficient training, preference learning, data creation. Use when customizing models for specific domains.

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

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.

SQL and NoSQL schema design with normalization, indexing, and migration patterns. Use when designing database schemas, creating tables, optimizing slow queries, or planning database migrations.

Orchestrate single user-invocable skill across 3 parallel scenarios with synchronized state and progressive difficulty. Use for demos, testing, and progressive validation workflows.

Database version control and change management patterns. Use when managing schema history, coordinating database changes across environments, implementing audit trails, or versioning database objects.

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.

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