bmad-spike

To run a time-boxed spike, define a crisp investigation question, set a strict deadline, and follow structured activation steps to produce evidence-backed artifacts. This workflow converts technical uncertainty into actionable ADRs, PoC findings, and trade-off analyses for the backlog.

The best way to trace spike investigations to backlog stories is by producing tracker artifacts with front-matter defined deliverables. This structured workflow links spike outcomes directly to downstream implementation items, ensuring full traceability from research to execution.

Yes, you can use a time-boxed investigation workflow for functional domain research and process improvements. The structured spike workflow applies to both technical and functional questions, producing repeatable steps and traceable knowledge artifacts regardless of the research domain.

No specific dependencies are required to start a structured spike workflow. You simply need to define a crisp spike question, set a timebox, and begin the initialization step to produce evidence-backed knowledge artifacts like PoC findings and ADRs.

A spike should produce evidence-backed knowledge artifacts such as Architecture Decision Records, PoC findings, and findings summaries. These deliverables are defined in the front-matter and ensure the time-boxed investigation converts uncertainty into actionable, traceable knowledge.