Skip to main content
Manage individual tasks within a benchmark that define what your agent should accomplish.

Actions

ActionDescription
listList all tasks in a benchmark
createCreate a new task
getGet a task by ID
updateUpdate a task
deleteDelete a task

Parameters

list

ParameterTypeRequiredDescription
benchmarkIdstringyesBenchmark ID to list tasks for

create

ParameterTypeRequiredDescription
benchmarkIdstringyesBenchmark ID to add the task to
namestringyesName of the task
timeoutnumbernoTask-specific timeout in seconds
taskItemobjectnoTask item configuration (patient, instruction, trigger)
scenarioobjectnoScenario configuration for the task
tagsstring[]noTags for filtering and grouping

get

ParameterTypeRequiredDescription
idstringyesTask ID

update

ParameterTypeRequiredDescription
idstringyesTask ID
namestringnoUpdated name
timeoutnumber | nullnoUpdated timeout, or null to clear
taskItemobject | nullnoUpdated task item, or null to clear
scenarioobject | nullnoUpdated scenario, or null to clear
tagsstring[]noUpdated tags

delete

ParameterTypeRequiredDescription
idstringyesTask ID

Examples

Create a prior auth submission task

{
  "action": "create",
  "benchmarkId": "bm_abc123",
  "name": "Submit prior auth for MRI",
  "timeout": 120,
  "taskItem": {
    "patient": "Maria Chen",
    "instruction": "Submit a prior authorization request for a lumbar MRI"
  },
  "tags": ["prior-auth", "imaging"]
}

List tasks in a benchmark

{
  "action": "list",
  "benchmarkId": "bm_abc123"
}

Update task tags

{
  "action": "update",
  "id": "task_xyz789",
  "tags": ["prior-auth", "imaging", "urgent"]
}