Use this agent when you need to maintain continuity and prevent duplicate work across your project. This agent should be consulted: (1) At the start of each new development session to review what has been completed; (2) Before implementing new features or changes to verify they haven't already been done; (3) When making architectural decisions to ensure alignment with the long-term roadmap; (4) Whenever the project structure or file organization changes; (5) When reviewing work from other AI models to verify no duplication has occurred; (6) Periodically to audit current work against historical records and catch scope creep or misalignment. Example: User starts a new session and says 'I want to add authentication to the API.' The assistant should use the project-historian agent to check if authentication work has already been completed, then report findings before proceeding. Another example: An AI model completes a feature and the assistant uses project-historian to verify the feature wasn't already implemented in a previous session.

Use this agent when architecting Next.js applications, designing app structure with App Router, implementing data fetching strategies, optimizing performance with streaming and Suspense, designing middleware patterns, handling authentication flows, or planning full-stack Next.js systems. Examples: (1) "How do we structure a multi-tenant app in Next.js?" - design workspace isolation, RLS integration, route organization. (2) "Real-time features are slow to load" - implement streaming, Suspense, progressive rendering. (3) "Which data fetching strategy for cost dashboard?" - Server components, revalidation, ISR patterns.