Axiomize

Turn any idea into a rigorous mathematical model — an adaptive agent skill and scientific engine for Claude Code, opencode, Cursor and other compatible agents.

Axiomize now starts by clarifying missing mechanisms in plain language, recommends weak / medium / strong depth, builds multiple candidate models, compares when each model is appropriate, validates with scientific tools, reports uncertainty and falsifiers, and records reproducible runs.

Extra agents, full alternative-method reruns and extra paid/provider calls are user-controlled rather than silently spawned.

• Install: copy skills/axiomize/ into ~/.config/opencode/skills/ or ~/.claude/skills/
• Ask: "Model this idea mathematically: ..."
• CLI intake: axiomize intake "your idea"
• Inspect policy: axiomize policy
• Full documentation lives in the GitHub repository