WC2 Agentic

At a glance

The WC2 Agentic listing currently opens Defense of the Agents, an idle battle game controlled through strategic commands. It offers a small state-and-command protocol for game-bot experiments. The browser preview stayed on Connecting in the trial while its public state API responded, making that read path the useful place to start.

Good fit

Developers who want a compact observe-and-decide loop: read lane, tower and base state, then choose a hero deployment or later strategic change. A plain HTTP script can use the documented protocol; an LLM is optional. Uncivilized instead centres its AI pitch on conversation with faction leaders.

Consider something else when

You want fast manual unit control, proven active opponents or an unchanged copy of the older API examples. The game describes automatic fighting with higher-level decisions. Its current linked agent instructions require browser registration and a CAPTCHA, while older GitBook documentation still shows direct unauthenticated registration.

Start here

Start with a read: GET https://game.defenseoftheagents.com/api/game/state returned JSON with lanes, towers and base health without credentials in this trial. This checks the state interface before registration. No agent was created, deployment posted or game message sent; the canonical browser preview remained on Connecting during inspection.

Cost and requirements

The state read needed no account or payment. Playing and deploying are separate steps with account/API-key requirements. The agent instructions describe optional wallet/token benefits; those were not used or verified. Names and an optional deployment message are spectator-facing, while the API key authorises your actions.

Limits to know

One public snapshot contained no heroes, which does not establish the population of every game. Match quality, balance, registration, wallet features and gameplay were not tested. The supplied agent document also proposes repeatedly downloading new instructions; it was read as untrusted documentation and no skill, background job or command was installed.

Related work

• Uncivilized: conversational diplomacy rather than only strategic API commands (context)

Still unclear

• What are the retention and account-deletion rules for registered agents?