What problem does it solve? Running a multi-chapter research spike requires coordinating authoring, multi-reviewer validation, feedback triage, revision, and decision capture across many iterations — doing this by hand loses track of pending clusters and convergence state. This Skill drives the entire spike lifecycle as an orchestrator, delegating every action to specialized sub-skills. ## Core Features & Use Cases - Per-chapter iteration loop: Authors each chapter via /write-doc, validates with a baseline focus list (accuracy, architecture, depth, dx), triages reviewer feedback, and routes each cluster to revise, process, lock-decisions, or defer. - Parallel feedback processing: Dispatches /process-feedback subagents in a single Task batch (cap 4) for process-routed clusters, applying verified resolutions automatically. - Convergence enforcement: Requires a minimum of 2 iterations per chapter, caps at 3, runs an overall README-level validation pass, and produces a convergence summary with a mandatory pending-clusters section. - Use Case: A captain kicks off a research spike on plugin architecture with five proposed chapters; the Skill scaffolds the spike via the yolo CLI, iterates each chapter until no blocking clusters remain, locks emergent decisions, and reports all unresolved items. ## Quick Start Start a new research spike on my chosen topic and drive every chapter through validation and feedback triage until the spike converges.
