context-engineering-advisor

Context engineering designs bounded memory and retrieval rules to keep AI focused, whereas context stuffing hoards unstructured data, causing context rot. This Skill diagnoses these symptoms and prescribes a structured Context Manifest to optimize AI-driven product management work.

To fix context rot in multi-agent RAG systems, implement a reset-then-implement cycle using a two-layer memory architecture. This Skill guides you through boundary ownership and a Context Manifest to ensure persistent memory remains reliable.

Structure memory architecture by defining a two-layer model separating short-term conversational state from long-term persistent memory. This Skill provides diagnostic frameworks to establish boundary ownership and retrieval strategies for product planning workflows.

Yes, you can use this context engineering framework for backlog prioritization and discovery planning. It guides AI-driven product management workflows by applying a Research, Plan, Reset, Implement cycle to maintain bounded, relevant context.

Implement a reset cycle when teams experience context rot or AI outputs lose relevance during complex tasks. This Skill uses a structured reset-then-implement approach to clear short-term conversational memory and re-apply boundaries before continuing.