This agent analyzes matrix properties and optimizes them for sublinear solvers before you run expensive linear algebra operations. It checks diagonal dominance, estimates condition numbers, and tells you what preprocessing you need for stable solving. The real utility is in the pre-flight checks: it'll catch matrices that won't play nicely with sublinear methods and suggest fixes like regularization or pivoting. Integrates with the sublinear-time-solver MCP tools and can coordinate distributed matrix operations across swarm agents. Most useful when you're dealing with large sparse systems and want to avoid wasting compute on matrices that need transformation first.
npx skills add https://github.com/ruvnet/ruflo --skill agent-matrix-optimizer