gamenumerics
Deterministic numeric engines for game design — exposed to your AI coding agent via MCP. Import xlsx config tables, audit growth curves, run battle/gacha simulations, and reverse-engineer the formulas behind game spreadsheets. Every numeric answer comes from zero-dependency pure-function engines — zero LLM guessing.
Who it's for: game designers and solo/small teams doing roguelike / deckbuilder balance work, spreadsheet-driven numerical design, or live-ops tuning audits — no engine integration required, just an xlsx export of your config tables.
Quick start (MCP server)
The fastest way to use these engines is through mcp-server-gamenumerics — a stdio MCP server exposing 17 read-only tools to Claude Code, Cursor, ZCode, or any MCP host:
npx -y mcp-server-gamenumerics
{ "mcpServers": { "gamenumerics": { "command": "npx", "args": ["-y", "mcp-server-gamenumerics"] } } }
Full per-host config (Claude Code / Cursor / ZCode), the 17-tool reference, and the performance baseline live in mcp-server/README.md.
What's inside
Tool surface (17 = 14 mapped + 3 meta)
import_xlsx · list_tables / read_table · battle_simulate / simulate_gacha / compute_power / power_curve / eval_formula / audit_column / infer_column_rule / infer_table_relation · grade_workspace / profile_table / infer_foreign_keys · list_workspaces / set_workspace / read_memory
Read-only by design — write tools stay behind the web workbench's human-in-the-loop confirm flow (separate product, not part of this repo).
Development
npm install
npm test # vitest: engine + table + mcp-server suites
npm run check # tsc --noEmit (lib + scripts)
npm run mcp:check # tsc for mcp-server (own tsconfig)
npm run mcp:build # esbuild bundle → mcp-server/dist/index.js