AI Agent Economic Layer v1.0-Sim

The default is a hypothetical one dollar task across four equally weighted agent records. Gas per hop is point zero zero one dollars times forty divided by ten, or point zero zero four. Four hops consume point zero one six, leaving point nine eight four distributable, with no protocol fee. At three hundred pixels per dollar the budget bar is three hundred, gas four point eight and distributable two hundred ninety five point two. Each quarter share is point two four six. This is assigned arithmetic, not real gas estimation or an on-chain transfer. The fictional stake and trust fields do not invoke wallets, networks, models or verified capabilities. The slashing preset marks the third agent malicious and sets its payout to zero. The other three quarter shares remain point two four six each, totaling point seven three eight. At three hundred pixels per dollar the unslashed payout is two hundred ninety five point two, remaining payout two hundred twenty one point four and removed share seventy three point eight. The slash penalty percentage slider is stored but never used in this calculation. Fifteen percent escrow divides one dollar by four then multiplies point fifteen, yielding point zero three seven five per agent. Escrow is reported only; it is not subtracted from distributable funds or transferred, and removed payouts are not redistributed. High gas sets the assumed price to one hundred eighty, giving point zero one eight per hop and point zero seven two over four hops. Payout becomes point nine two eight, lower than default point nine eight four by point zero five six. At three hundred pixels per dollar default payout measures two hundred ninety five point two, high gas two hundred seventy eight point four and difference sixteen point eight. These synthetic gas units do not represent current blockchain costs. The graph initializer safely returns when Cytoscape is absent. Sliders, presets, reset and actual JSON or CSV export operate locally without that graph. Initialization fills the four ledger rows after a scheduled callback, so comparisons wait for actual rows rather than a guessed delay. The export recomputes current arithmetic. No actual settlement, security or yield investment is proved.