compound-capture

To preserve cross-sprint learnings, extract evidence from sprint outcomes and patch durable lessons into shared memory only when provenance is guaranteed by decisive artifacts. This ensures reusable guidance survives state reconciliation without regressions.

The best way to link sprint outcomes to durable memory is to apply a durable-learning workflow that grounds lessons in decisive sprint evidence. You must reference the exact artifact paths proving provenance before updating memory or reference docs.

You distill evidence into actionable guidance by identifying the queued feature in tracked-work.json and running a compounding workflow. This extracts verified outcomes into durable lessons that inform future contracts, proposals, or design decisions.

You should update docs/live/memory.md only after verifying that evidence supports a lasting, reusable lesson. This prevents active sprint state regressions while ensuring shared memory reflects proven, decisive sprint artifacts.

Yes, you can capture durable lessons without affecting active sprint state by using a compounding workflow. It updates shared memory and linked records based on proven evidence while explicitly avoiding regressions in your current tracked work.

Cross-sprint memory requires provenance from decisive artifacts to ensure lessons are grounded in verified evidence rather than noise. Referencing artifact paths proves the learning is lasting and reusable for future design decisions.