What problem does it solve? When a department runs multiple epics against finite capacity, work starts because it can rather than because it should — there is no prioritization, no WIP discipline, and no cross-epic dependency view. This Skill provides the management layer above a single epic: a Portfolio Kanban that ranks epics economically and controls how many run at once. ## Core Features & Use Cases - WSJF Prioritization: Computes Weighted Shortest Job First (cost of delay divided by job size) from relative Fibonacci estimates of business value, time criticality, risk reduction, and job size, so the highest-value epic is always pulled first. - WIP Limits via Little's Law: Enforces work-in-progress caps on every Kanban state except Funnel, keeping cycle time short even when AI makes starting new work cheap. - Dependency DAG & OKR Linkage: Tracks cross-epic dependencies as a cycle-free directed graph and requires every epic past Funnel to name the OKR it advances. - Use Case: A department has eight proposed epics and capacity for two. The Skill maintains .ai/portfolio/manifest.md, ranks all eight by WSJF, blocks epics with unfinished dependencies, pulls the top two into the epic lane, and dispatches portfolio-reviewer to validate the ranking before acting on it. ## Quick Start Ask the AI to run portfolio-management to build a WSJF-ranked Portfolio Kanban from your current list of epics and pull the top-ranked unblocked epic into the epic lane.
