Pokémon Showdown
Not yet importedCompetitive turn-based battles with structured state, legal-action constraints, and objective match outcomes.
1 task previewagent · sandboxedMITversion upstream
Moderate import. Requires a versioned battle server sandbox, structured state and actions, opponent policy, and match replay.
What it measures
Pokémon Showdown is a strong simulation candidate because battles are browser-accessible and can expose structured observations and legal actions without requiring real-time visual control.
The import still needs a reproducible server, seeded teams, opponent policy, turn budgets, battle logs, and a clearly versioned ruleset.
not yet imported
Not yet imported. A first release should use structured turn-based control and pinned teams rather than visual browser automation.
Tasks
1 previewed| id | task | runtime |
|---|---|---|
| showdown/battle | Pinned-team battleSelect legal actions against a fixed opponent policy and ruleset.tools: battle state, battle action | sandbox |
Tools
2 requested| tool | description | kind | required |
|---|---|---|---|
| battle state | Returns public battle state and the agent's private team state. | sandbox | yes |
| battle action | Submits one legal move or switch for the current turn. | sandbox | yes |