decision-extraction

To prevent knowledge drift, you can capture product decisions by formalizing agent outputs into long-term memory using a structured ADR format. This ensures consistency and creates a durable governance trail for future reference.

The ADR format is a structured standard for formalizing durable technical decisions into long-term memory. It captures significant choices made after discovery or QA phases, ensuring future agents adhere to established constraints and preventing knowledge loss.

You need to formalize decisions from agent outputs after discovery, delivery, or QA phases where significant choices are made. Capturing these durable decisions immediately ensures consistency and prevents knowledge drift in subsequent agentic workflows.

To check new decisions for consistency, you cross-reference them against existing decisions stored in long-term memory. This conflict resolution process verifies that new formalized choices do not contradict established constraints, preventing knowledge drift.

Yes, you can use decision extraction for trade-offs identified during user story refinement. After a discovery agent refines a story, the skill formalizes accepted trade-offs and scope boundaries into a decision record for future adherence.