v0.2.5x Agent System
Status: Planned for v0.2.5x.
Agents are intended to help keep arenas active while introducing a second form of player skill: training and strategy design. They are not presented as active Ranked participants today.
Native Agents
Section titled “Native Agents”The initial plan introduces 12 native starter Agents with different behavior, intelligence, and play styles. They form the starting point for later community-trained Agent NFTs and their royalty model.
Bot Versus Agent
Section titled “Bot Versus Agent”A traditional bot follows fixed reactions. An Agent is intended to evaluate context: current HP, accumulated value, nearby threats, objectives, and whether a fight is worth taking.
Examples of trainable decisions include:
- when to attack or retreat;
- which target to prioritize;
- when to protect accumulated value;
- how aggressively to contest an objective; and
- when to reposition instead of taking a risky fight.
Planned Arenas
Section titled “Planned Arenas”- Arenagent: trained Agent versus Agent competition.
- Mixed arenas: human and Agent participants in the same supported ruleset.
Agent ownership, training, minting, COLA operating requirements, settlement, and human-versus-Agent eligibility remain planned design. They require explicit rules before any real-token version is enabled.