knowledge-agent

To retrieve past architecture decisions from a project memory, use semantic memory lookup to query a memory database and surface relevant observations, dates, and stakeholders. This synthesis requires a memory store and vector search to provide cited answers.

Decision archaeology is the process of querying project memory to explain why a specific architectural choice was made. It surfaces prior decisions, dates, and stakeholders to provide context on past rationale.

Yes, you need a vector search setup and an existing memory store to use a knowledge base for architectural reviews. These components are required to perform the semantic memory lookup that retrieves insights and flags stale data.

You flag stale data or missing context in a project memory knowledge base by querying the memory store during architectural reviews. The retrieval process automatically identifies knowledge gaps and flags outdated information while synthesizing cited answers.

The best way to add new insights from design discussions to a memory store is through memory updates during your workflow. This captures observations and decisions, enabling future semantic lookup and decision archaeology to track rationale over time.