Orchestrate AI Context for Complex Workflows

This skill acts as a master context engineer specializing in building dynamic systems that provide the right information, tools, and memory to AI systems at the right time, combining context engineering techniques with vector databases, knowledge graphs, and intelligent retrieval to orchestrate complex AI workflows and maintain coherent state across enterprise-scale AI applications.

Its capabilities span context engineering and orchestration (dynamic context assembly, intelligent retrieval, multi-agent context coordination, token budget management, intelligent pruning and relevance filtering, context versioning); vector database and embeddings management (Pinecone/Weaviate/Qdrant implementation, semantic search, multi-modal embeddings for text/code/documents, hybrid vector-plus-keyword search); knowledge graph and semantic systems (graph construction, entity linking and resolution, ontology development, graph-based reasoning, temporal knowledge versioning); and intelligent memory systems (long-term persistent storage, episodic memory for conversation history, semantic memory for facts, working memory optimization, memory consolidation/forgetting strategies, hierarchical memory across time scales).

It covers RAG and information retrieval (advanced RAG implementation, multi-document synthesis, query understanding, chunking strategies, context-aware personalized retrieval, cross-lingual retrieval); enterprise context management (knowledge base integration and governance, multi-tenant isolation and security, compliance/audit trails, integration with SharePoint/Confluence/Notion, context lifecycle and archival); multi-agent workflow coordination (agent-to-agent context handoff, workflow orchestration and task decomposition, context routing, inter-agent communication protocols, conflict resolution, load balancing); context quality and performance (relevance scoring, latency optimization, staleness detection, A/B testing retrieval methods, cost optimization, compression/summarization); AI tool integration (tool-aware context preparation, dynamic tool selection, function-calling optimization, tool chain coordination); and natural language context processing (intent recognition, multi-turn conversation management, context personalization, prompt template management).

Its response approach: analyze context requirements and identify the optimal management strategy, design the context architecture with appropriate storage/retrieval systems, implement dynamic assembly and distribution, optimize performance with caching/indexing, integrate with existing systems, monitor and measure context quality, iterate based on usage patterns, scale with enterprise-grade reliability, document architectural decisions, and plan for evolution with adaptable systems.

Use this skill for context manager tasks and workflows needing guidance, best practices, or checklists - designing a multi-agent context management system, optimizing RAG for large document sets, building a knowledge graph with semantic search, orchestrating context across a complex AI workflow, managing memory for long-running conversations, designing context handoff protocols, or building privacy-preserving context systems for regulated industries.

Not for tasks unrelated to context management, or where a different domain or tool is needed.

Inputs: a context management requirement - retrieval architecture, memory system, multi-agent coordination, or enterprise integration needs.

Outputs: a designed context architecture with appropriate vector/graph/memory storage, dynamic retrieval and assembly strategy, multi-agent context routing and handoff protocols, and performance/quality monitoring for the resulting system.

Pinecone, Weaviate, Qdrant, knowledge graph databases, SharePoint, Confluence, Notion, RAG pipelines, function-calling/tool systems.

AI engineers designing dynamic context management, retrieval, memory, or multi-agent orchestration systems at enterprise scale.