What problem does it solve?
This Skill prevents EOS SaaS teams from reinventing fragile table logic by providing a robust, state-driven way to implement sorting, filtering, pagination, and row selection without UI glitches or re-render loops.
Core Features & Use Cases
- Headless table state engine (@tanstack/react-table v8): centralizes row-model pipelines (core, sorted, filtered, paginated) so the UI only renders state outputs.
- EOS SaaS integration pattern: wires TanStack Query server-driven pagination/sorting/filtering into controlled table state (manualPagination/manualSorting/manualFiltering) with stable memoization.
- Production-grade correctness & performance: supports faceted filters, stable row selection via getRowId, expanding rows, and row virtualization using @tanstack/react-virtual for large datasets.
- Bug prevention: detects common failure modes like unstable data/columns references, wrong accessorKey vs accessorFn usage, missing matching row models for enabled state, and incorrect flexRender usage.
Quick Start
Use the tanstack_table skill to implement a typed Leads table with server-driven pagination/sorting/filtering and virtualized rows, ensuring stable column definitions and getRowId-based row selection.