What problem does it solve? Backlogs decay into vague wish-lists or duplicate graveyards when left unrefined, forcing teams to re-litigate intent before every iteration. This Skill keeps the top of the backlog execution-ready by decomposing oversized items, consolidating duplicates, and verifying readiness before iteration planning. ## Core Features & Use Cases - Backlog Triage & De-duplication: Sweep the backlog for stale, duplicated, or superseded items and consolidate or retire them with documented rationale. - Decomposition & Detailing: Split oversized items into deliverable work items with traceability from parent to children, confirming intent, boundaries, and dependencies. - Readiness Verification & Cohort Sizing: Check scope alignment, terminology, and acceptance-criteria presence, then apply cohort-level sizing suitable for sequencing without false precision. - Use Case: Before sprint planning, run refinement across the backlog so the AI Program Manager and Product Owner hand Forge-Task-Routing a clean, execution-ready set of work items with current metadata. ## Quick Start Review the current backlog, decompose any oversized work items, consolidate duplicates, and flag which items are execution-ready for the next iteration.
