published on Monday, Oct 19, 2026 by Pulumi
published on Monday, Oct 19, 2026 by Pulumi
Creates a new agent task for the specified organization. The request must include a prompt (the user event message) that initiates the task. The message is a user message, so its discriminator field must be set to “type”: “user_message”. Set the ‘permissionMode’ field in the request body to restrict the agent to read-only operations. Returns the created task details including task ID, name, status, and timestamp.
Create Task Resource
Resources are created with functions called constructors. To learn more about declaring and configuring resources, see Resources.
Constructor syntax
new Task(name: string, args: TaskArgs, opts?: CustomResourceOptions);@overload
def Task(resource_name: str,
args: TaskArgs,
opts: Optional[ResourceOptions] = None)
@overload
def Task(resource_name: str,
opts: Optional[ResourceOptions] = None,
org_name: Optional[str] = None,
agent_definition_id: Optional[str] = None,
approval_mode: Optional[str] = None,
cli_integrations: Optional[Sequence[Any]] = None,
enabled_integrations: Optional[Sequence[Any]] = None,
message: Optional[Any] = None,
permission_mode: Optional[str] = None,
plan_mode: Optional[bool] = None,
role: Optional[str] = None,
source: Optional[str] = None,
task_id: Optional[str] = None,
tool_execution_mode: Optional[str] = None)func NewTask(ctx *Context, name string, args TaskArgs, opts ...ResourceOption) (*Task, error)public Task(string name, TaskArgs args, CustomResourceOptions? opts = null)type: pulumiservice:api/agents:Task
properties: # The arguments to resource properties.
options: # Bag of options to control resource's behavior.
resource "pulumiservice_api_agents_task" "name" {
# resource properties
}Parameters
- name string
- The unique name of the resource.
- args TaskArgs
- The arguments to resource properties.
- opts CustomResourceOptions
- Bag of options to control resource's behavior.
- resource_name str
- The unique name of the resource.
- args TaskArgs
- The arguments to resource properties.
- opts ResourceOptions
- Bag of options to control resource's behavior.
- ctx Context
- Context object for the current deployment.
- name string
- The unique name of the resource.
- args TaskArgs
- The arguments to resource properties.
- opts ResourceOption
- Bag of options to control resource's behavior.
- name string
- The unique name of the resource.
- args TaskArgs
- The arguments to resource properties.
- opts CustomResourceOptions
- Bag of options to control resource's behavior.
- name String
- The unique name of the resource.
- args TaskArgs
- The arguments to resource properties.
- options CustomResourceOptions
- Bag of options to control resource's behavior.
Constructor example
The following reference example uses placeholder values for all input properties.
var taskResource = new PulumiService.Api.Agents.Task("taskResource", new()
{
OrgName = "string",
AgentDefinitionId = "string",
ApprovalMode = "string",
CliIntegrations = new[]
{
"any",
},
EnabledIntegrations = new[]
{
"any",
},
Message = "any",
PermissionMode = "string",
PlanMode = false,
Role = "string",
Source = "string",
TaskID = "string",
ToolExecutionMode = "string",
});
example, err := agents.NewTask(ctx, "taskResource", &agents.TaskArgs{
OrgName: pulumi.String("string"),
AgentDefinitionId: pulumi.String("string"),
ApprovalMode: pulumi.String("string"),
CliIntegrations: pulumi.Array{
pulumi.Any("any"),
},
EnabledIntegrations: pulumi.Array{
pulumi.Any("any"),
},
Message: pulumi.Any("any"),
PermissionMode: pulumi.String("string"),
PlanMode: pulumi.Bool(false),
Role: pulumi.String("string"),
Source: pulumi.String("string"),
TaskID: pulumi.String("string"),
ToolExecutionMode: pulumi.String("string"),
})
resource "pulumiservice_api_agents_task" "taskResource" {
lifecycle {
create_before_destroy = true
}
org_name = "string"
agent_definition_id = "string"
approval_mode = "string"
cli_integrations = ["any"]
enabled_integrations = ["any"]
message = "any"
permission_mode = "string"
plan_mode = false
role = "string"
source = "string"
task_id = "string"
tool_execution_mode = "string"
}
var taskResource = new Task("taskResource", TaskArgs.builder()
.orgName("string")
.agentDefinitionId("string")
.approvalMode("string")
.cliIntegrations("any")
.enabledIntegrations("any")
.message("any")
.permissionMode("string")
.planMode(false)
.role("string")
.source("string")
.taskID("string")
.toolExecutionMode("string")
.build());
task_resource = pulumiservice.api.agents.Task("taskResource",
org_name="string",
agent_definition_id="string",
approval_mode="string",
cli_integrations=["any"],
enabled_integrations=["any"],
message="any",
permission_mode="string",
plan_mode=False,
role="string",
source="string",
task_id="string",
tool_execution_mode="string")
const taskResource = new pulumiservice.api.agents.Task("taskResource", {
orgName: "string",
agentDefinitionId: "string",
approvalMode: "string",
cliIntegrations: ["any"],
enabledIntegrations: ["any"],
message: "any",
permissionMode: "string",
planMode: false,
role: "string",
source: "string",
taskID: "string",
toolExecutionMode: "string",
});
type: pulumiservice:api/agents:Task
properties:
agentDefinitionId: string
approvalMode: string
cliIntegrations:
- any
enabledIntegrations:
- any
message: any
orgName: string
permissionMode: string
planMode: false
role: string
source: string
taskID: string
toolExecutionMode: string
Task Resource Properties
To learn more about resource properties and how to use them, see Inputs and Outputs in the Architecture and Concepts docs.
Inputs
In Python, inputs that are objects can be passed either as argument classes or as dictionary literals.
The Task resource accepts the following input properties:
- Org
Name string - The organization name
- Agent
Definition stringId - Optional Custom Agent definition id to run this task as. When set, the task's role, permission mode, and enabled integrations are taken from the named agent definition instead of from this request, and the task is attributed to that agent. The agent's instructions and approval mode are not copied at creation: they are read from the definition on every turn, so editing the agent applies to the next turn of every run. A preview capability, gated per organization.
- Approval
Mode string - Optional approval mode override for this task. If omitted, org default is used.
- Cli
Integrations List<object> - Optional filter for CLI integrations to enable for this task. Semantics: omitted/null → enable all CLI integrations connected for the org; empty list → explicit opt-out (no CLI integrations for this task); populated list → whitelist by (catalogId, name) of the configured instances to enable. Entries with missing or unknown catalogId, missing name, or referencing a (catalogId, name) pair that is not connected for the organization are rejected with a 400 response. catalogId matching is case-insensitive.
- Enabled
Integrations List<object> - Optional list of integrations to enable for this task. Semantics: omitted/null → inherit all org-enabled integrations; empty list → explicit opt-out (no integration credentials for this task); populated list → whitelist of specific integrations by ID. Modeled as an object array rather than a bare string array so multi-instance support (instance_name, scope, etc.) can be added later without a wire break.
- Message object
- The message content. This is the first event in the conversation and must be a user message, so its discriminator field must be set to "type": "user_message".
- Permission
Mode string - Controls the permission scope for the task. When omitted, defaults to 'default' (the agent uses the creating user's full permissions).
- Plan
Mode bool - Whether to enable plan mode for this task.
- Role string
- Optional RBAC role the task assumes, identified by role id. When set, the task operates with that role's permissions in place of the creating user's own assignments. You may only assume a role that is assigned to you, directly or through a team; a role you do not hold is rejected with a 403. Omitted/null means no assumed role: the task uses the creating user's full permissions. Orthogonal to permissionMode, which controls the read-only behavior posture independently.
- Source string
- The origin that triggered this task. Defaults to 'api' if omitted.
- Task
ID string - The agent task identifier
- Tool
Execution stringMode - Where tools should be executed. Defaults to 'cloud' if omitted.
- Org
Name string - The organization name
- Agent
Definition stringId - Optional Custom Agent definition id to run this task as. When set, the task's role, permission mode, and enabled integrations are taken from the named agent definition instead of from this request, and the task is attributed to that agent. The agent's instructions and approval mode are not copied at creation: they are read from the definition on every turn, so editing the agent applies to the next turn of every run. A preview capability, gated per organization.
- Approval
Mode string - Optional approval mode override for this task. If omitted, org default is used.
- Cli
Integrations []interface{} - Optional filter for CLI integrations to enable for this task. Semantics: omitted/null → enable all CLI integrations connected for the org; empty list → explicit opt-out (no CLI integrations for this task); populated list → whitelist by (catalogId, name) of the configured instances to enable. Entries with missing or unknown catalogId, missing name, or referencing a (catalogId, name) pair that is not connected for the organization are rejected with a 400 response. catalogId matching is case-insensitive.
- Enabled
Integrations []interface{} - Optional list of integrations to enable for this task. Semantics: omitted/null → inherit all org-enabled integrations; empty list → explicit opt-out (no integration credentials for this task); populated list → whitelist of specific integrations by ID. Modeled as an object array rather than a bare string array so multi-instance support (instance_name, scope, etc.) can be added later without a wire break.
- Message interface{}
- The message content. This is the first event in the conversation and must be a user message, so its discriminator field must be set to "type": "user_message".
- Permission
Mode string - Controls the permission scope for the task. When omitted, defaults to 'default' (the agent uses the creating user's full permissions).
- Plan
Mode bool - Whether to enable plan mode for this task.
- Role string
- Optional RBAC role the task assumes, identified by role id. When set, the task operates with that role's permissions in place of the creating user's own assignments. You may only assume a role that is assigned to you, directly or through a team; a role you do not hold is rejected with a 403. Omitted/null means no assumed role: the task uses the creating user's full permissions. Orthogonal to permissionMode, which controls the read-only behavior posture independently.
- Source string
- The origin that triggered this task. Defaults to 'api' if omitted.
- Task
ID string - The agent task identifier
- Tool
Execution stringMode - Where tools should be executed. Defaults to 'cloud' if omitted.
- org_
name string - The organization name
- agent_
definition_ stringid - Optional Custom Agent definition id to run this task as. When set, the task's role, permission mode, and enabled integrations are taken from the named agent definition instead of from this request, and the task is attributed to that agent. The agent's instructions and approval mode are not copied at creation: they are read from the definition on every turn, so editing the agent applies to the next turn of every run. A preview capability, gated per organization.
- approval_
mode string - Optional approval mode override for this task. If omitted, org default is used.
- cli_
integrations list(any) - Optional filter for CLI integrations to enable for this task. Semantics: omitted/null → enable all CLI integrations connected for the org; empty list → explicit opt-out (no CLI integrations for this task); populated list → whitelist by (catalogId, name) of the configured instances to enable. Entries with missing or unknown catalogId, missing name, or referencing a (catalogId, name) pair that is not connected for the organization are rejected with a 400 response. catalogId matching is case-insensitive.
- enabled_
integrations list(any) - Optional list of integrations to enable for this task. Semantics: omitted/null → inherit all org-enabled integrations; empty list → explicit opt-out (no integration credentials for this task); populated list → whitelist of specific integrations by ID. Modeled as an object array rather than a bare string array so multi-instance support (instance_name, scope, etc.) can be added later without a wire break.
- message any
- The message content. This is the first event in the conversation and must be a user message, so its discriminator field must be set to "type": "user_message".
- permission_
mode string - Controls the permission scope for the task. When omitted, defaults to 'default' (the agent uses the creating user's full permissions).
- plan_
mode bool - Whether to enable plan mode for this task.
- role string
- Optional RBAC role the task assumes, identified by role id. When set, the task operates with that role's permissions in place of the creating user's own assignments. You may only assume a role that is assigned to you, directly or through a team; a role you do not hold is rejected with a 403. Omitted/null means no assumed role: the task uses the creating user's full permissions. Orthogonal to permissionMode, which controls the read-only behavior posture independently.
- source string
- The origin that triggered this task. Defaults to 'api' if omitted.
- task_
id string - The agent task identifier
- tool_
execution_ stringmode - Where tools should be executed. Defaults to 'cloud' if omitted.
- org
Name String - The organization name
- agent
Definition StringId - Optional Custom Agent definition id to run this task as. When set, the task's role, permission mode, and enabled integrations are taken from the named agent definition instead of from this request, and the task is attributed to that agent. The agent's instructions and approval mode are not copied at creation: they are read from the definition on every turn, so editing the agent applies to the next turn of every run. A preview capability, gated per organization.
- approval
Mode String - Optional approval mode override for this task. If omitted, org default is used.
- cli
Integrations List<Object> - Optional filter for CLI integrations to enable for this task. Semantics: omitted/null → enable all CLI integrations connected for the org; empty list → explicit opt-out (no CLI integrations for this task); populated list → whitelist by (catalogId, name) of the configured instances to enable. Entries with missing or unknown catalogId, missing name, or referencing a (catalogId, name) pair that is not connected for the organization are rejected with a 400 response. catalogId matching is case-insensitive.
- enabled
Integrations List<Object> - Optional list of integrations to enable for this task. Semantics: omitted/null → inherit all org-enabled integrations; empty list → explicit opt-out (no integration credentials for this task); populated list → whitelist of specific integrations by ID. Modeled as an object array rather than a bare string array so multi-instance support (instance_name, scope, etc.) can be added later without a wire break.
- message Object
- The message content. This is the first event in the conversation and must be a user message, so its discriminator field must be set to "type": "user_message".
- permission
Mode String - Controls the permission scope for the task. When omitted, defaults to 'default' (the agent uses the creating user's full permissions).
- plan
Mode Boolean - Whether to enable plan mode for this task.
- role String
- Optional RBAC role the task assumes, identified by role id. When set, the task operates with that role's permissions in place of the creating user's own assignments. You may only assume a role that is assigned to you, directly or through a team; a role you do not hold is rejected with a 403. Omitted/null means no assumed role: the task uses the creating user's full permissions. Orthogonal to permissionMode, which controls the read-only behavior posture independently.
- source String
- The origin that triggered this task. Defaults to 'api' if omitted.
- task
ID String - The agent task identifier
- tool
Execution StringMode - Where tools should be executed. Defaults to 'cloud' if omitted.
- org
Name string - The organization name
- agent
Definition stringId - Optional Custom Agent definition id to run this task as. When set, the task's role, permission mode, and enabled integrations are taken from the named agent definition instead of from this request, and the task is attributed to that agent. The agent's instructions and approval mode are not copied at creation: they are read from the definition on every turn, so editing the agent applies to the next turn of every run. A preview capability, gated per organization.
- approval
Mode string - Optional approval mode override for this task. If omitted, org default is used.
- cli
Integrations any[] - Optional filter for CLI integrations to enable for this task. Semantics: omitted/null → enable all CLI integrations connected for the org; empty list → explicit opt-out (no CLI integrations for this task); populated list → whitelist by (catalogId, name) of the configured instances to enable. Entries with missing or unknown catalogId, missing name, or referencing a (catalogId, name) pair that is not connected for the organization are rejected with a 400 response. catalogId matching is case-insensitive.
- enabled
Integrations any[] - Optional list of integrations to enable for this task. Semantics: omitted/null → inherit all org-enabled integrations; empty list → explicit opt-out (no integration credentials for this task); populated list → whitelist of specific integrations by ID. Modeled as an object array rather than a bare string array so multi-instance support (instance_name, scope, etc.) can be added later without a wire break.
- message any
- The message content. This is the first event in the conversation and must be a user message, so its discriminator field must be set to "type": "user_message".
- permission
Mode string - Controls the permission scope for the task. When omitted, defaults to 'default' (the agent uses the creating user's full permissions).
- plan
Mode boolean - Whether to enable plan mode for this task.
- role string
- Optional RBAC role the task assumes, identified by role id. When set, the task operates with that role's permissions in place of the creating user's own assignments. You may only assume a role that is assigned to you, directly or through a team; a role you do not hold is rejected with a 403. Omitted/null means no assumed role: the task uses the creating user's full permissions. Orthogonal to permissionMode, which controls the read-only behavior posture independently.
- source string
- The origin that triggered this task. Defaults to 'api' if omitted.
- task
ID string - The agent task identifier
- tool
Execution stringMode - Where tools should be executed. Defaults to 'cloud' if omitted.
- org_
name str - The organization name
- agent_
definition_ strid - Optional Custom Agent definition id to run this task as. When set, the task's role, permission mode, and enabled integrations are taken from the named agent definition instead of from this request, and the task is attributed to that agent. The agent's instructions and approval mode are not copied at creation: they are read from the definition on every turn, so editing the agent applies to the next turn of every run. A preview capability, gated per organization.
- approval_
mode str - Optional approval mode override for this task. If omitted, org default is used.
- cli_
integrations Sequence[Any] - Optional filter for CLI integrations to enable for this task. Semantics: omitted/null → enable all CLI integrations connected for the org; empty list → explicit opt-out (no CLI integrations for this task); populated list → whitelist by (catalogId, name) of the configured instances to enable. Entries with missing or unknown catalogId, missing name, or referencing a (catalogId, name) pair that is not connected for the organization are rejected with a 400 response. catalogId matching is case-insensitive.
- enabled_
integrations Sequence[Any] - Optional list of integrations to enable for this task. Semantics: omitted/null → inherit all org-enabled integrations; empty list → explicit opt-out (no integration credentials for this task); populated list → whitelist of specific integrations by ID. Modeled as an object array rather than a bare string array so multi-instance support (instance_name, scope, etc.) can be added later without a wire break.
- message Any
- The message content. This is the first event in the conversation and must be a user message, so its discriminator field must be set to "type": "user_message".
- permission_
mode str - Controls the permission scope for the task. When omitted, defaults to 'default' (the agent uses the creating user's full permissions).
- plan_
mode bool - Whether to enable plan mode for this task.
- role str
- Optional RBAC role the task assumes, identified by role id. When set, the task operates with that role's permissions in place of the creating user's own assignments. You may only assume a role that is assigned to you, directly or through a team; a role you do not hold is rejected with a 403. Omitted/null means no assumed role: the task uses the creating user's full permissions. Orthogonal to permissionMode, which controls the read-only behavior posture independently.
- source str
- The origin that triggered this task. Defaults to 'api' if omitted.
- task_
id str - The agent task identifier
- tool_
execution_ strmode - Where tools should be executed. Defaults to 'cloud' if omitted.
- org
Name String - The organization name
- agent
Definition StringId - Optional Custom Agent definition id to run this task as. When set, the task's role, permission mode, and enabled integrations are taken from the named agent definition instead of from this request, and the task is attributed to that agent. The agent's instructions and approval mode are not copied at creation: they are read from the definition on every turn, so editing the agent applies to the next turn of every run. A preview capability, gated per organization.
- approval
Mode String - Optional approval mode override for this task. If omitted, org default is used.
- cli
Integrations List<Any> - Optional filter for CLI integrations to enable for this task. Semantics: omitted/null → enable all CLI integrations connected for the org; empty list → explicit opt-out (no CLI integrations for this task); populated list → whitelist by (catalogId, name) of the configured instances to enable. Entries with missing or unknown catalogId, missing name, or referencing a (catalogId, name) pair that is not connected for the organization are rejected with a 400 response. catalogId matching is case-insensitive.
- enabled
Integrations List<Any> - Optional list of integrations to enable for this task. Semantics: omitted/null → inherit all org-enabled integrations; empty list → explicit opt-out (no integration credentials for this task); populated list → whitelist of specific integrations by ID. Modeled as an object array rather than a bare string array so multi-instance support (instance_name, scope, etc.) can be added later without a wire break.
- message Any
- The message content. This is the first event in the conversation and must be a user message, so its discriminator field must be set to "type": "user_message".
- permission
Mode String - Controls the permission scope for the task. When omitted, defaults to 'default' (the agent uses the creating user's full permissions).
- plan
Mode Boolean - Whether to enable plan mode for this task.
- role String
- Optional RBAC role the task assumes, identified by role id. When set, the task operates with that role's permissions in place of the creating user's own assignments. You may only assume a role that is assigned to you, directly or through a team; a role you do not hold is rejected with a 403. Omitted/null means no assumed role: the task uses the creating user's full permissions. Orthogonal to permissionMode, which controls the read-only behavior posture independently.
- source String
- The origin that triggered this task. Defaults to 'api' if omitted.
- task
ID String - The agent task identifier
- tool
Execution StringMode - Where tools should be executed. Defaults to 'cloud' if omitted.
Outputs
All input properties are implicitly available as output properties. Additionally, the Task resource produces the following output properties:
- Created
At string - When the task was created, in ISO 8601 format.
- Created
By object - Information about the user who created this task.
- Entities List<object>
- Pulumi entities (stacks, projects, etc.) that provide context for the agent.
- Id string
- The provider-assigned unique ID for this managed resource.
- bool
- Whether this task is shared with other org members.
- Name string
- Display name for the task, typically auto-generated from the initial user message.
- Status string
- Current execution status of the task.
- Task
Type string - Whether the task was started synchronously by a user or asynchronously by background automation.
- Tokens
Used int - Total Neo tokens consumed across all model invocations for this task. Neo tokens are the priced unit used for billing — distinct from the raw model input tokens surfaced in contextUsedTokens / contextWindowTokens.
- Async
Trigger stringType - The async trigger source for this task. Null for sync tasks.
- Context
Compaction intThreshold Percent - Percentage of the context window (1-100) at which the agent triggers conversation compaction. Populated alongside contextWindowTokens when token usage data is available; omitted otherwise.
- Context
Used intTokens - Total input tokens consumed across all model invocations for this task. Approximate context window usage.
- Context
Window intTokens - Maximum context window size in tokens for the primary model used by this task.
- Last
Heartbeat string - When the task runtime last reported a heartbeat. Null if the runtime has never checked in.
- Pending
Input intCount - The number of pending inputs waiting behind this task. Present for tasks that can receive asynchronous inputs.
- Role
Name string - The display name of the RBAC role this task assumes. Null when the task has no assumed role or the role no longer exists.
- Runtime
Phase string - The current runtime phase for this task. Null until the runtime checks in.
- string
- When the task was first shared. Null if never shared.
- Source
Automation stringID - The automation that spawned this task, if the task was created by an automation run.
- Vcs
Provider string - The version control system this task operates against. Set for tasks that target a specific VCS platform (e.g. code reviews). Null for tasks with no VCS context.
- Created
At string - When the task was created, in ISO 8601 format.
- Created
By interface{} - Information about the user who created this task.
- Entities []interface{}
- Pulumi entities (stacks, projects, etc.) that provide context for the agent.
- Id string
- The provider-assigned unique ID for this managed resource.
- bool
- Whether this task is shared with other org members.
- Name string
- Display name for the task, typically auto-generated from the initial user message.
- Status string
- Current execution status of the task.
- Task
Type string - Whether the task was started synchronously by a user or asynchronously by background automation.
- Tokens
Used int - Total Neo tokens consumed across all model invocations for this task. Neo tokens are the priced unit used for billing — distinct from the raw model input tokens surfaced in contextUsedTokens / contextWindowTokens.
- Async
Trigger stringType - The async trigger source for this task. Null for sync tasks.
- Context
Compaction intThreshold Percent - Percentage of the context window (1-100) at which the agent triggers conversation compaction. Populated alongside contextWindowTokens when token usage data is available; omitted otherwise.
- Context
Used intTokens - Total input tokens consumed across all model invocations for this task. Approximate context window usage.
- Context
Window intTokens - Maximum context window size in tokens for the primary model used by this task.
- Last
Heartbeat string - When the task runtime last reported a heartbeat. Null if the runtime has never checked in.
- Pending
Input intCount - The number of pending inputs waiting behind this task. Present for tasks that can receive asynchronous inputs.
- Role
Name string - The display name of the RBAC role this task assumes. Null when the task has no assumed role or the role no longer exists.
- Runtime
Phase string - The current runtime phase for this task. Null until the runtime checks in.
- string
- When the task was first shared. Null if never shared.
- Source
Automation stringID - The automation that spawned this task, if the task was created by an automation run.
- Vcs
Provider string - The version control system this task operates against. Set for tasks that target a specific VCS platform (e.g. code reviews). Null for tasks with no VCS context.
- created_
at string - When the task was created, in ISO 8601 format.
- created_
by any - Information about the user who created this task.
- entities list(any)
- Pulumi entities (stacks, projects, etc.) that provide context for the agent.
- id string
- The provider-assigned unique ID for this managed resource.
- bool
- Whether this task is shared with other org members.
- name string
- Display name for the task, typically auto-generated from the initial user message.
- status string
- Current execution status of the task.
- task_
type string - Whether the task was started synchronously by a user or asynchronously by background automation.
- tokens_
used number - Total Neo tokens consumed across all model invocations for this task. Neo tokens are the priced unit used for billing — distinct from the raw model input tokens surfaced in contextUsedTokens / contextWindowTokens.
- async_
trigger_ stringtype - The async trigger source for this task. Null for sync tasks.
- context_
compaction_ numberthreshold_ percent - Percentage of the context window (1-100) at which the agent triggers conversation compaction. Populated alongside contextWindowTokens when token usage data is available; omitted otherwise.
- context_
used_ numbertokens - Total input tokens consumed across all model invocations for this task. Approximate context window usage.
- context_
window_ numbertokens - Maximum context window size in tokens for the primary model used by this task.
- last_
heartbeat string - When the task runtime last reported a heartbeat. Null if the runtime has never checked in.
- pending_
input_ numbercount - The number of pending inputs waiting behind this task. Present for tasks that can receive asynchronous inputs.
- role_
name string - The display name of the RBAC role this task assumes. Null when the task has no assumed role or the role no longer exists.
- runtime_
phase string - The current runtime phase for this task. Null until the runtime checks in.
- string
- When the task was first shared. Null if never shared.
- source_
automation_ stringid - The automation that spawned this task, if the task was created by an automation run.
- vcs_
provider string - The version control system this task operates against. Set for tasks that target a specific VCS platform (e.g. code reviews). Null for tasks with no VCS context.
- created
At String - When the task was created, in ISO 8601 format.
- created
By Object - Information about the user who created this task.
- entities List<Object>
- Pulumi entities (stacks, projects, etc.) that provide context for the agent.
- id String
- The provider-assigned unique ID for this managed resource.
- Boolean
- Whether this task is shared with other org members.
- name String
- Display name for the task, typically auto-generated from the initial user message.
- status String
- Current execution status of the task.
- task
Type String - Whether the task was started synchronously by a user or asynchronously by background automation.
- tokens
Used Integer - Total Neo tokens consumed across all model invocations for this task. Neo tokens are the priced unit used for billing — distinct from the raw model input tokens surfaced in contextUsedTokens / contextWindowTokens.
- async
Trigger StringType - The async trigger source for this task. Null for sync tasks.
- context
Compaction IntegerThreshold Percent - Percentage of the context window (1-100) at which the agent triggers conversation compaction. Populated alongside contextWindowTokens when token usage data is available; omitted otherwise.
- context
Used IntegerTokens - Total input tokens consumed across all model invocations for this task. Approximate context window usage.
- context
Window IntegerTokens - Maximum context window size in tokens for the primary model used by this task.
- last
Heartbeat String - When the task runtime last reported a heartbeat. Null if the runtime has never checked in.
- pending
Input IntegerCount - The number of pending inputs waiting behind this task. Present for tasks that can receive asynchronous inputs.
- role
Name String - The display name of the RBAC role this task assumes. Null when the task has no assumed role or the role no longer exists.
- runtime
Phase String - The current runtime phase for this task. Null until the runtime checks in.
- String
- When the task was first shared. Null if never shared.
- source
Automation StringID - The automation that spawned this task, if the task was created by an automation run.
- vcs
Provider String - The version control system this task operates against. Set for tasks that target a specific VCS platform (e.g. code reviews). Null for tasks with no VCS context.
- created
At string - When the task was created, in ISO 8601 format.
- created
By any - Information about the user who created this task.
- entities any[]
- Pulumi entities (stacks, projects, etc.) that provide context for the agent.
- id string
- The provider-assigned unique ID for this managed resource.
- boolean
- Whether this task is shared with other org members.
- name string
- Display name for the task, typically auto-generated from the initial user message.
- status string
- Current execution status of the task.
- task
Type string - Whether the task was started synchronously by a user or asynchronously by background automation.
- tokens
Used number - Total Neo tokens consumed across all model invocations for this task. Neo tokens are the priced unit used for billing — distinct from the raw model input tokens surfaced in contextUsedTokens / contextWindowTokens.
- async
Trigger stringType - The async trigger source for this task. Null for sync tasks.
- context
Compaction numberThreshold Percent - Percentage of the context window (1-100) at which the agent triggers conversation compaction. Populated alongside contextWindowTokens when token usage data is available; omitted otherwise.
- context
Used numberTokens - Total input tokens consumed across all model invocations for this task. Approximate context window usage.
- context
Window numberTokens - Maximum context window size in tokens for the primary model used by this task.
- last
Heartbeat string - When the task runtime last reported a heartbeat. Null if the runtime has never checked in.
- pending
Input numberCount - The number of pending inputs waiting behind this task. Present for tasks that can receive asynchronous inputs.
- role
Name string - The display name of the RBAC role this task assumes. Null when the task has no assumed role or the role no longer exists.
- runtime
Phase string - The current runtime phase for this task. Null until the runtime checks in.
- string
- When the task was first shared. Null if never shared.
- source
Automation stringID - The automation that spawned this task, if the task was created by an automation run.
- vcs
Provider string - The version control system this task operates against. Set for tasks that target a specific VCS platform (e.g. code reviews). Null for tasks with no VCS context.
- created_
at str - When the task was created, in ISO 8601 format.
- created_
by Any - Information about the user who created this task.
- entities Sequence[Any]
- Pulumi entities (stacks, projects, etc.) that provide context for the agent.
- id str
- The provider-assigned unique ID for this managed resource.
- bool
- Whether this task is shared with other org members.
- name str
- Display name for the task, typically auto-generated from the initial user message.
- status str
- Current execution status of the task.
- task_
type str - Whether the task was started synchronously by a user or asynchronously by background automation.
- tokens_
used int - Total Neo tokens consumed across all model invocations for this task. Neo tokens are the priced unit used for billing — distinct from the raw model input tokens surfaced in contextUsedTokens / contextWindowTokens.
- async_
trigger_ strtype - The async trigger source for this task. Null for sync tasks.
- context_
compaction_ intthreshold_ percent - Percentage of the context window (1-100) at which the agent triggers conversation compaction. Populated alongside contextWindowTokens when token usage data is available; omitted otherwise.
- context_
used_ inttokens - Total input tokens consumed across all model invocations for this task. Approximate context window usage.
- context_
window_ inttokens - Maximum context window size in tokens for the primary model used by this task.
- last_
heartbeat str - When the task runtime last reported a heartbeat. Null if the runtime has never checked in.
- pending_
input_ intcount - The number of pending inputs waiting behind this task. Present for tasks that can receive asynchronous inputs.
- role_
name str - The display name of the RBAC role this task assumes. Null when the task has no assumed role or the role no longer exists.
- runtime_
phase str - The current runtime phase for this task. Null until the runtime checks in.
- str
- When the task was first shared. Null if never shared.
- source_
automation_ strid - The automation that spawned this task, if the task was created by an automation run.
- vcs_
provider str - The version control system this task operates against. Set for tasks that target a specific VCS platform (e.g. code reviews). Null for tasks with no VCS context.
- created
At String - When the task was created, in ISO 8601 format.
- created
By Any - Information about the user who created this task.
- entities List<Any>
- Pulumi entities (stacks, projects, etc.) that provide context for the agent.
- id String
- The provider-assigned unique ID for this managed resource.
- Boolean
- Whether this task is shared with other org members.
- name String
- Display name for the task, typically auto-generated from the initial user message.
- status String
- Current execution status of the task.
- task
Type String - Whether the task was started synchronously by a user or asynchronously by background automation.
- tokens
Used Number - Total Neo tokens consumed across all model invocations for this task. Neo tokens are the priced unit used for billing — distinct from the raw model input tokens surfaced in contextUsedTokens / contextWindowTokens.
- async
Trigger StringType - The async trigger source for this task. Null for sync tasks.
- context
Compaction NumberThreshold Percent - Percentage of the context window (1-100) at which the agent triggers conversation compaction. Populated alongside contextWindowTokens when token usage data is available; omitted otherwise.
- context
Used NumberTokens - Total input tokens consumed across all model invocations for this task. Approximate context window usage.
- context
Window NumberTokens - Maximum context window size in tokens for the primary model used by this task.
- last
Heartbeat String - When the task runtime last reported a heartbeat. Null if the runtime has never checked in.
- pending
Input NumberCount - The number of pending inputs waiting behind this task. Present for tasks that can receive asynchronous inputs.
- role
Name String - The display name of the RBAC role this task assumes. Null when the task has no assumed role or the role no longer exists.
- runtime
Phase String - The current runtime phase for this task. Null until the runtime checks in.
- String
- When the task was first shared. Null if never shared.
- source
Automation StringID - The automation that spawned this task, if the task was created by an automation run.
- vcs
Provider String - The version control system this task operates against. Set for tasks that target a specific VCS platform (e.g. code reviews). Null for tasks with no VCS context.
Package Details
- Repository
- pulumiservice pulumi/pulumi-pulumiservice
- License
- Apache-2.0
published on Monday, Oct 19, 2026 by Pulumi