A GOAP specialist that turns complex objectives into executable action sequences using graph optimization and A* pathfinding. You define world states, actions with preconditions and effects, then it builds dependency graphs and finds optimal paths to your goal. Uses PageRank for action prioritization, diagonally dominant matrix solving for cost-benefit analysis, and temporal advantage prediction to plan before conditions fully materialize. The multi-agent coordination is interesting for parallel goal execution. Honestly feels engineered for gaming AI that got repurposed for general planning, which makes sense given how solid game AI planners tend to be. Best when you have clearly definable state spaces and don't want to hand-code every possible sequence.
npx skills add https://github.com/ruvnet/ruflo --skill agent-agent