2026/05/13 - Amazon Bedrock AgentCore Control - 6 new12 updated api methods
Changes Adds support for read-only summary APIs for Policy Engine, Policy, and Policy Generation resources, enabling metadata retrieval without KMS decryption for AWS Config integration.
Retrieves a metadata-only summary of a specific policy without decrypting customer content. This lightweight read operation returns resource identifiers, status, and timestamps, but does not include the policy definition, description, or status reasons. Because this operation does not require access to the customer's KMS key, it is suitable for resource discovery, inventory, and integration scenarios where only metadata is needed.
See also: AWS API Documentation
Request Syntax
client.get_policy_summary(
policyEngineId='string',
policyId='string'
)
string
[REQUIRED]
The identifier of the policy engine that manages the policy to retrieve the summary for.
string
[REQUIRED]
The unique identifier of the policy to retrieve the summary for. This must be a valid policy ID that exists within the specified policy engine.
dict
Response Syntax
{
'policyId': 'string',
'name': 'string',
'policyEngineId': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'policyArn': 'string',
'status': 'CREATING'|'ACTIVE'|'UPDATING'|'DELETING'|'CREATE_FAILED'|'UPDATE_FAILED'|'DELETE_FAILED'
}
Response Structure
(dict) --
policyId (string) --
The unique identifier of the policy.
name (string) --
The customer-assigned name of the policy.
policyEngineId (string) --
The identifier of the policy engine that manages this policy.
createdAt (datetime) --
The timestamp when the policy was originally created.
updatedAt (datetime) --
The timestamp when the policy was last modified.
policyArn (string) --
The Amazon Resource Name (ARN) of the policy.
status (string) --
The current status of the policy.
Retrieves a paginated list of metadata-only policy summaries within a policy engine without decrypting customer content. This lightweight read operation returns resource identifiers, status, and timestamps for each policy, but does not include policy definitions, descriptions, or status reasons. Because this operation does not require access to the customer's KMS key, it is suitable for resource discovery, inventory, and integration scenarios where only metadata is needed.
See also: AWS API Documentation
Request Syntax
client.list_policy_summaries(
nextToken='string',
maxResults=123,
policyEngineId='string',
targetResourceScope='string'
)
string
A pagination token returned from a previous ListPolicySummaries call. Use this token to retrieve the next page of results when the response is paginated.
integer
The maximum number of policy summaries to return in a single response.
string
[REQUIRED]
The identifier of the policy engine whose policy summaries to retrieve.
string
Optional filter to list policy summaries that apply to a specific resource scope or resource type. This helps narrow down results to those relevant for particular Amazon Web Services resources, agent tools, or operational contexts within the policy engine ecosystem.
dict
Response Syntax
{
'policies': [
{
'policyId': 'string',
'name': 'string',
'policyEngineId': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'policyArn': 'string',
'status': 'CREATING'|'ACTIVE'|'UPDATING'|'DELETING'|'CREATE_FAILED'|'UPDATE_FAILED'|'DELETE_FAILED'
},
],
'nextToken': 'string'
}
Response Structure
(dict) --
policies (list) --
An array of policy summary objects that match the specified criteria. Each summary contains resource identifiers, status, and timestamps without customer-encrypted content.
(dict) --
Represents a metadata-only summary of a policy resource. This structure contains resource identifiers, status, and timestamps without customer-encrypted fields such as definition, description, or status reasons. Policy summaries are returned by operations that do not require access to the customer's KMS key.
policyId (string) --
The unique identifier for the policy.
name (string) --
The customer-assigned name of the policy.
policyEngineId (string) --
The identifier of the policy engine that manages this policy.
createdAt (datetime) --
The timestamp when the policy was originally created.
updatedAt (datetime) --
The timestamp when the policy was last modified.
policyArn (string) --
The Amazon Resource Name (ARN) of the policy.
status (string) --
The current status of the policy.
nextToken (string) --
A pagination token that can be used in subsequent ListPolicySummaries calls to retrieve additional results. This token is only present when there are more results available.
Retrieves a paginated list of metadata-only policy generation summaries within a policy engine without decrypting customer content. This lightweight read operation returns resource identifiers, status, timestamps, and findings for each policy generation, but does not include status reasons. Because this operation does not require access to the customer's KMS key, it is suitable for resource discovery, inventory, and integration scenarios where only metadata is needed.
See also: AWS API Documentation
Request Syntax
client.list_policy_generation_summaries(
nextToken='string',
maxResults=123,
policyEngineId='string'
)
string
A pagination token returned from a previous ListPolicyGenerationSummaries call. Use this token to retrieve the next page of results when the response is paginated.
integer
The maximum number of policy generation summaries to return in a single response.
string
[REQUIRED]
The identifier of the policy engine whose policy generation summaries to retrieve.
dict
Response Syntax
{
'policyGenerations': [
{
'policyEngineId': 'string',
'policyGenerationId': 'string',
'name': 'string',
'policyGenerationArn': 'string',
'resource': {
'arn': 'string'
},
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'status': 'GENERATING'|'GENERATED'|'GENERATE_FAILED'|'DELETE_FAILED',
'findings': 'string'
},
],
'nextToken': 'string'
}
Response Structure
(dict) --
policyGenerations (list) --
An array of policy generation summary objects that match the specified criteria. Each summary contains resource identifiers, status, timestamps, and findings without customer-encrypted content.
(dict) --
Represents a metadata-only summary of a policy generation resource. This structure contains resource identifiers, status, timestamps, and findings without customer-encrypted fields such as status reasons. Policy generation summaries are returned by operations that do not require access to the customer's KMS key.
policyEngineId (string) --
The identifier of the policy engine associated with this generation request.
policyGenerationId (string) --
The unique identifier for this policy generation request.
name (string) --
The customer-assigned name for this policy generation request.
policyGenerationArn (string) --
The ARN of this policy generation request.
resource (dict) --
The resource information associated with this policy generation.
arn (string) --
The Amazon Resource Name (ARN) of the resource. This globally unique identifier specifies the exact resource that policies will be evaluated against for access control decisions.
createdAt (datetime) --
The timestamp when this policy generation request was created.
updatedAt (datetime) --
The timestamp when this policy generation was last updated.
status (string) --
The current status of this policy generation request.
findings (string) --
Findings and insights from this policy generation process.
nextToken (string) --
A pagination token that can be used in subsequent ListPolicyGenerationSummaries calls to retrieve additional results. This token is only present when there are more results available.
Retrieves a metadata-only summary of a specific policy generation request without decrypting customer content. This lightweight read operation returns resource identifiers, status, timestamps, and findings, but does not include status reasons. Because this operation does not require access to the customer's KMS key, it is suitable for resource discovery, inventory, and integration scenarios where only metadata is needed.
See also: AWS API Documentation
Request Syntax
client.get_policy_generation_summary(
policyGenerationId='string',
policyEngineId='string'
)
string
[REQUIRED]
The unique identifier of the policy generation request to retrieve the summary for.
string
[REQUIRED]
The identifier of the policy engine associated with the policy generation request.
dict
Response Syntax
{
'policyEngineId': 'string',
'policyGenerationId': 'string',
'name': 'string',
'policyGenerationArn': 'string',
'resource': {
'arn': 'string'
},
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'status': 'GENERATING'|'GENERATED'|'GENERATE_FAILED'|'DELETE_FAILED',
'findings': 'string'
}
Response Structure
(dict) --
policyEngineId (string) --
The identifier of the policy engine associated with this policy generation.
policyGenerationId (string) --
The unique identifier of the policy generation request.
name (string) --
The customer-assigned name for the policy generation request.
policyGenerationArn (string) --
The Amazon Resource Name (ARN) of the policy generation request.
resource (dict) --
The resource information associated with the policy generation.
arn (string) --
The Amazon Resource Name (ARN) of the resource. This globally unique identifier specifies the exact resource that policies will be evaluated against for access control decisions.
createdAt (datetime) --
The timestamp when the policy generation request was created.
updatedAt (datetime) --
The timestamp when the policy generation was last updated.
status (string) --
The current status of the policy generation request.
findings (string) --
The findings from the policy generation process, if available.
Retrieves a paginated list of metadata-only policy engine summaries without decrypting customer content. This lightweight read operation returns resource identifiers, status, and timestamps for each policy engine, but does not include descriptions or status reasons. Because this operation does not require access to the customer's KMS key, it is suitable for resource discovery, inventory, and integration scenarios where only metadata is needed.
See also: AWS API Documentation
Request Syntax
client.list_policy_engine_summaries(
nextToken='string',
maxResults=123
)
string
A pagination token returned from a previous ListPolicyEngineSummaries call. Use this token to retrieve the next page of results when the response is paginated.
integer
The maximum number of policy engine summaries to return in a single response.
dict
Response Syntax
{
'policyEngines': [
{
'policyEngineId': 'string',
'name': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'policyEngineArn': 'string',
'status': 'CREATING'|'ACTIVE'|'UPDATING'|'DELETING'|'CREATE_FAILED'|'UPDATE_FAILED'|'DELETE_FAILED',
'encryptionKeyArn': 'string'
},
],
'nextToken': 'string'
}
Response Structure
(dict) --
policyEngines (list) --
An array of policy engine summary objects that exist in the account. Each summary contains resource identifiers, status, and timestamps without customer-encrypted content.
(dict) --
Represents a metadata-only summary of a policy engine resource. This structure contains resource identifiers, status, and timestamps without customer-encrypted fields such as description or status reasons. Policy engine summaries are returned by operations that do not require access to the customer's KMS key.
policyEngineId (string) --
The unique identifier for the policy engine.
name (string) --
The customer-assigned name of the policy engine.
createdAt (datetime) --
The timestamp when the policy engine was originally created.
updatedAt (datetime) --
The timestamp when the policy engine was last modified.
policyEngineArn (string) --
The Amazon Resource Name (ARN) of the policy engine.
status (string) --
The current status of the policy engine.
encryptionKeyArn (string) --
The Amazon Resource Name (ARN) of the KMS key used to encrypt the policy engine data.
nextToken (string) --
A pagination token that can be used in subsequent ListPolicyEngineSummaries calls to retrieve additional results. This token is only present when there are more results available.
Retrieves a metadata-only summary of a specific policy engine without decrypting customer content. This lightweight read operation returns resource identifiers, status, timestamps, and the encryption key ARN, but does not include the description or status reasons. Because this operation does not require access to the customer's KMS key, it is suitable for resource discovery, inventory, and integration scenarios where only metadata is needed.
See also: AWS API Documentation
Request Syntax
client.get_policy_engine_summary(
policyEngineId='string'
)
string
[REQUIRED]
The unique identifier of the policy engine to retrieve the summary for. This must be a valid policy engine ID that exists within the account.
dict
Response Syntax
{
'policyEngineId': 'string',
'name': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'policyEngineArn': 'string',
'status': 'CREATING'|'ACTIVE'|'UPDATING'|'DELETING'|'CREATE_FAILED'|'UPDATE_FAILED'|'DELETE_FAILED',
'encryptionKeyArn': 'string'
}
Response Structure
(dict) --
policyEngineId (string) --
The unique identifier of the policy engine.
name (string) --
The customer-assigned name of the policy engine.
createdAt (datetime) --
The timestamp when the policy engine was originally created.
updatedAt (datetime) --
The timestamp when the policy engine was last modified.
policyEngineArn (string) --
The Amazon Resource Name (ARN) of the policy engine.
status (string) --
The current status of the policy engine.
encryptionKeyArn (string) --
The Amazon Resource Name (ARN) of the KMS key used to encrypt the policy engine data.
{'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}
Creates an Amazon Bedrock AgentCore Runtime.
See also: AWS API Documentation
Request Syntax
client.create_agent_runtime(
agentRuntimeName='string',
agentRuntimeArtifact={
'containerConfiguration': {
'containerUri': 'string'
},
'codeConfiguration': {
'code': {
's3': {
'bucket': 'string',
'prefix': 'string',
'versionId': 'string'
}
},
'runtime': 'PYTHON_3_10'|'PYTHON_3_11'|'PYTHON_3_12'|'PYTHON_3_13'|'PYTHON_3_14'|'NODE_22',
'entryPoint': [
'string',
]
}
},
roleArn='string',
networkConfiguration={
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
clientToken='string',
description='string',
authorizerConfiguration={
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
requestHeaderConfiguration={
'requestHeaderAllowlist': [
'string',
]
},
protocolConfiguration={
'serverProtocol': 'MCP'|'HTTP'|'A2A'|'AGUI'
},
lifecycleConfiguration={
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
environmentVariables={
'string': 'string'
},
filesystemConfigurations=[
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
],
tags={
'string': 'string'
}
)
string
[REQUIRED]
The name of the AgentCore Runtime.
dict
[REQUIRED]
The artifact of the AgentCore Runtime.
containerConfiguration (dict) --
The container configuration for the agent artifact.
containerUri (string) -- [REQUIRED]
The ECR URI of the container.
codeConfiguration (dict) --
The code configuration for the agent runtime artifact, including the source code location and execution settings.
code (dict) -- [REQUIRED]
The source code location and configuration details.
s3 (dict) --
The Amazon Amazon S3 object that contains the source code for the agent runtime.
bucket (string) -- [REQUIRED]
The name of the Amazon S3 bucket. This bucket contains the stored data.
prefix (string) -- [REQUIRED]
The prefix for objects in the Amazon S3 bucket. This prefix is added to the object keys to organize the data.
versionId (string) --
The version ID of the Amazon Amazon S3 object. If not specified, the latest version of the object is used.
runtime (string) -- [REQUIRED]
The runtime environment for executing the agent code. Specify the programming language and version to use for the agent runtime. For valid values, see the list of supported runtimes.
entryPoint (list) -- [REQUIRED]
The entry point for the code execution, specifying the function or method that should be invoked when the code runs.
(string) --
string
[REQUIRED]
The IAM role ARN that provides permissions for the AgentCore Runtime.
dict
[REQUIRED]
The network configuration for the AgentCore Runtime.
networkMode (string) -- [REQUIRED]
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) -- [REQUIRED]
The security groups associated with the VPC configuration.
(string) --
subnets (list) -- [REQUIRED]
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
string
A unique, case-sensitive identifier to ensure idempotency of the request.
This field is autopopulated if not provided.
string
The description of the AgentCore Runtime.
dict
The authorizer configuration for the AgentCore Runtime.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) -- [REQUIRED]
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) -- [REQUIRED]
The name of the custom claim field to check.
inboundTokenClaimValueType (string) -- [REQUIRED]
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) -- [REQUIRED]
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) -- [REQUIRED]
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) -- [REQUIRED]
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) -- [REQUIRED]
The domain to override with a private endpoint.
privateEndpoint (dict) -- [REQUIRED]
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
dict
Configuration for HTTP request headers that will be passed through to the runtime.
requestHeaderAllowlist (list) --
A list of HTTP request headers that are allowed to be passed through to the runtime.
(string) --
dict
The protocol configuration for an agent runtime. This structure defines how the agent runtime communicates with clients.
serverProtocol (string) -- [REQUIRED]
The server protocol for the agent runtime. This field specifies which protocol the agent runtime uses to communicate with clients.
dict
The life cycle configuration for the AgentCore Runtime.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
dict
Environment variables to set in the AgentCore Runtime environment.
(string) --
(string) --
list
The filesystem configurations to mount into the AgentCore Runtime. Use filesystem configurations to provide persistent storage to your AgentCore Runtime sessions.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) -- [REQUIRED]
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
dict
A map of tag keys and values to assign to the agent runtime. Tags enable you to categorize your resources in different ways, for example, by purpose, owner, or environment.
(string) --
(string) --
dict
Response Syntax
{
'agentRuntimeArn': 'string',
'workloadIdentityDetails': {
'workloadIdentityArn': 'string'
},
'agentRuntimeId': 'string',
'agentRuntimeVersion': 'string',
'createdAt': datetime(2015, 1, 1),
'status': 'CREATING'|'CREATE_FAILED'|'UPDATING'|'UPDATE_FAILED'|'READY'|'DELETING'
}
Response Structure
(dict) --
agentRuntimeArn (string) --
The Amazon Resource Name (ARN) of the AgentCore Runtime.
workloadIdentityDetails (dict) --
The workload identity details for the AgentCore Runtime.
workloadIdentityArn (string) --
The ARN associated with the workload identity.
agentRuntimeId (string) --
The unique identifier of the AgentCore Runtime.
agentRuntimeVersion (string) --
The version of the AgentCore Runtime.
createdAt (datetime) --
The timestamp when the AgentCore Runtime was created.
status (string) --
The current status of the AgentCore Runtime.
{'networkConfiguration': {'vpcConfig': {'requireServiceS3Endpoint': 'boolean'}}}
Creates a custom browser.
See also: AWS API Documentation
Request Syntax
client.create_browser(
name='string',
description='string',
executionRoleArn='string',
networkConfiguration={
'networkMode': 'PUBLIC'|'VPC',
'vpcConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
recording={
'enabled': True|False,
's3Location': {
'bucket': 'string',
'prefix': 'string',
'versionId': 'string'
}
},
browserSigning={
'enabled': True|False
},
enterprisePolicies=[
{
'location': {
's3': {
'bucket': 'string',
'prefix': 'string',
'versionId': 'string'
}
},
'type': 'MANAGED'|'RECOMMENDED'
},
],
certificates=[
{
'location': {
'secretsManager': {
'secretArn': 'string'
}
}
},
],
clientToken='string',
tags={
'string': 'string'
}
)
string
[REQUIRED]
The name of the browser. The name must be unique within your account.
string
The description of the browser.
string
The Amazon Resource Name (ARN) of the IAM role that provides permissions for the browser to access Amazon Web Services services.
dict
[REQUIRED]
The network configuration for the browser. This configuration specifies the network mode for the browser.
networkMode (string) -- [REQUIRED]
The network mode for the browser. This field specifies how the browser connects to the network.
vpcConfig (dict) --
The VPC configuration for the browser. This configuration is required when the network mode is set to VPC.
securityGroups (list) -- [REQUIRED]
The security groups associated with the VPC configuration.
(string) --
subnets (list) -- [REQUIRED]
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
dict
The recording configuration for the browser. When enabled, browser sessions are recorded and stored in the specified Amazon S3 location.
enabled (boolean) --
Indicates whether recording is enabled for the browser. When set to true, browser sessions are recorded.
s3Location (dict) --
The Amazon S3 location where browser recordings are stored. This location contains the recorded browser sessions.
bucket (string) -- [REQUIRED]
The name of the Amazon S3 bucket. This bucket contains the stored data.
prefix (string) -- [REQUIRED]
The prefix for objects in the Amazon S3 bucket. This prefix is added to the object keys to organize the data.
versionId (string) --
The version ID of the Amazon Amazon S3 object. If not specified, the latest version of the object is used.
dict
The browser signing configuration that enables cryptographic agent identification using HTTP message signatures for web bot authentication.
enabled (boolean) -- [REQUIRED]
Specifies whether browser signing is enabled. When enabled, the browser will cryptographically sign HTTP requests to identify itself as an AI agent to bot control vendors.
list
A list of enterprise policy files for the browser.
(dict) --
Browser enterprise policy configuration.
location (dict) -- [REQUIRED]
The location of the enterprise policy file.
s3 (dict) --
The Amazon S3 location for storing data. This structure defines where in Amazon S3 data is stored.
bucket (string) -- [REQUIRED]
The name of the Amazon S3 bucket. This bucket contains the stored data.
prefix (string) -- [REQUIRED]
The prefix for objects in the Amazon S3 bucket. This prefix is added to the object keys to organize the data.
versionId (string) --
The version ID of the Amazon Amazon S3 object. If not specified, the latest version of the object is used.
type (string) --
The type of browser enterprise policy. Available values are MANAGED and RECOMMENDED.
list
A list of certificates to install in the browser.
(dict) --
A certificate to install in the browser or code interpreter.
location (dict) -- [REQUIRED]
The location of the certificate.
secretsManager (dict) --
The Amazon Web Services Secrets Manager location of the certificate.
secretArn (string) -- [REQUIRED]
The ARN of the Amazon Web Services Secrets Manager secret containing the certificate.
string
A unique, case-sensitive identifier to ensure that the operation completes no more than one time. If this token matches a previous request, Amazon Bedrock AgentCore ignores the request but does not return an error.
This field is autopopulated if not provided.
dict
A map of tag keys and values to assign to the browser. Tags enable you to categorize your resources in different ways, for example, by purpose, owner, or environment.
(string) --
(string) --
dict
Response Syntax
{
'browserId': 'string',
'browserArn': 'string',
'createdAt': datetime(2015, 1, 1),
'status': 'CREATING'|'CREATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED'|'DELETED'
}
Response Structure
(dict) --
browserId (string) --
The unique identifier of the created browser.
browserArn (string) --
The Amazon Resource Name (ARN) of the created browser.
createdAt (datetime) --
The timestamp when the browser was created.
status (string) --
The current status of the browser.
{'networkConfiguration': {'vpcConfig': {'requireServiceS3Endpoint': 'boolean'}}}
Creates a custom code interpreter.
See also: AWS API Documentation
Request Syntax
client.create_code_interpreter(
name='string',
description='string',
executionRoleArn='string',
networkConfiguration={
'networkMode': 'PUBLIC'|'SANDBOX'|'VPC',
'vpcConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
certificates=[
{
'location': {
'secretsManager': {
'secretArn': 'string'
}
}
},
],
clientToken='string',
tags={
'string': 'string'
}
)
string
[REQUIRED]
The name of the code interpreter. The name must be unique within your account.
string
The description of the code interpreter.
string
The Amazon Resource Name (ARN) of the IAM role that provides permissions for the code interpreter to access Amazon Web Services services.
dict
[REQUIRED]
The network configuration for the code interpreter. This configuration specifies the network mode for the code interpreter.
networkMode (string) -- [REQUIRED]
The network mode for the code interpreter. This field specifies how the code interpreter connects to the network.
vpcConfig (dict) --
The VPC configuration for the code interpreter. This configuration is required when the network mode is set to VPC.
securityGroups (list) -- [REQUIRED]
The security groups associated with the VPC configuration.
(string) --
subnets (list) -- [REQUIRED]
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
list
A list of certificates to install in the code interpreter.
(dict) --
A certificate to install in the browser or code interpreter.
location (dict) -- [REQUIRED]
The location of the certificate.
secretsManager (dict) --
The Amazon Web Services Secrets Manager location of the certificate.
secretArn (string) -- [REQUIRED]
The ARN of the Amazon Web Services Secrets Manager secret containing the certificate.
string
A unique, case-sensitive identifier to ensure that the operation completes no more than one time. If this token matches a previous request, Amazon Bedrock AgentCore ignores the request but does not return an error.
This field is autopopulated if not provided.
dict
A map of tag keys and values to assign to the code interpreter. Tags enable you to categorize your resources in different ways, for example, by purpose, owner, or environment.
(string) --
(string) --
dict
Response Syntax
{
'codeInterpreterId': 'string',
'codeInterpreterArn': 'string',
'createdAt': datetime(2015, 1, 1),
'status': 'CREATING'|'CREATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED'|'DELETED'
}
Response Structure
(dict) --
codeInterpreterId (string) --
The unique identifier of the created code interpreter.
codeInterpreterArn (string) --
The Amazon Resource Name (ARN) of the created code interpreter.
createdAt (datetime) --
The timestamp when the code interpreter was created.
status (string) --
The current status of the code interpreter.
{'environment': {'agentCoreRuntimeEnvironment': {'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}}}
Response {'harness': {'environment': {'agentCoreRuntimeEnvironment': {'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}}}}
Operation to create a Harness.
See also: AWS API Documentation
Request Syntax
client.create_harness(
harnessName='string',
clientToken='string',
executionRoleArn='string',
environment={
'agentCoreRuntimeEnvironment': {
'lifecycleConfiguration': {
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'filesystemConfigurations': [
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
]
}
},
environmentArtifact={
'containerConfiguration': {
'containerUri': 'string'
}
},
environmentVariables={
'string': 'string'
},
authorizerConfiguration={
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
model={
'bedrockModelConfig': {
'modelId': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'openAiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'geminiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...,
'topK': 123
}
},
systemPrompt=[
{
'text': 'string'
},
],
tools=[
{
'type': 'remote_mcp'|'agentcore_browser'|'agentcore_gateway'|'inline_function'|'agentcore_code_interpreter',
'name': 'string',
'config': {
'remoteMcp': {
'url': 'string',
'headers': {
'string': 'string'
}
},
'agentCoreBrowser': {
'browserArn': 'string'
},
'agentCoreGateway': {
'gatewayArn': 'string',
'outboundAuth': {
'awsIam': {}
,
'none': {}
,
'oauth': {
'providerArn': 'string',
'scopes': [
'string',
],
'customParameters': {
'string': 'string'
},
'grantType': 'CLIENT_CREDENTIALS'|'AUTHORIZATION_CODE'|'TOKEN_EXCHANGE',
'defaultReturnUrl': 'string'
}
}
},
'inlineFunction': {
'description': 'string',
'inputSchema': {...}|[...]|123|123.4|'string'|True|None
},
'agentCoreCodeInterpreter': {
'codeInterpreterArn': 'string'
}
}
},
],
skills=[
{
'path': 'string'
},
],
allowedTools=[
'string',
],
memory={
'agentCoreMemoryConfiguration': {
'arn': 'string',
'actorId': 'string',
'messagesCount': 123,
'retrievalConfig': {
'string': {
'topK': 123,
'relevanceScore': ...,
'strategyId': 'string'
}
}
}
},
truncation={
'strategy': 'sliding_window'|'summarization'|'none',
'config': {
'slidingWindow': {
'messagesCount': 123
},
'summarization': {
'summaryRatio': ...,
'preserveRecentMessages': 123,
'summarizationSystemPrompt': 'string'
}
}
},
maxIterations=123,
maxTokens=123,
timeoutSeconds=123,
tags={
'string': 'string'
}
)
string
[REQUIRED]
The name of the harness. Must start with a letter and contain only alphanumeric characters and underscores.
string
A unique, case-sensitive identifier to ensure idempotency of the request.
This field is autopopulated if not provided.
string
[REQUIRED]
The ARN of the IAM role that the harness assumes when running. This role must have permissions for the services the agent needs to access, such as Amazon Bedrock for model invocation.
dict
The compute environment configuration for the harness, including network and lifecycle settings.
agentCoreRuntimeEnvironment (dict) --
The AgentCore Runtime environment configuration.
lifecycleConfiguration (dict) --
LifecycleConfiguration lets you manage the lifecycle of runtime sessions and resources in AgentCore Runtime. This configuration helps optimize resource utilization by automatically cleaning up idle sessions and preventing long-running instances from consuming resources indefinitely.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
networkConfiguration (dict) --
SecurityConfig for the Agent.
networkMode (string) -- [REQUIRED]
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) -- [REQUIRED]
The security groups associated with the VPC configuration.
(string) --
subnets (list) -- [REQUIRED]
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
filesystemConfigurations (list) --
The filesystem configurations for the runtime environment.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) -- [REQUIRED]
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
dict
The environment artifact for the harness, such as a custom container image containing additional dependencies.
containerConfiguration (dict) --
Representation of a container configuration.
containerUri (string) -- [REQUIRED]
The ECR URI of the container.
dict
Environment variables to set in the harness runtime environment.
(string) --
(string) --
dict
Represents inbound authorization configuration options used to authenticate incoming requests.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) -- [REQUIRED]
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) -- [REQUIRED]
The name of the custom claim field to check.
inboundTokenClaimValueType (string) -- [REQUIRED]
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) -- [REQUIRED]
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) -- [REQUIRED]
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) -- [REQUIRED]
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) -- [REQUIRED]
The domain to override with a private endpoint.
privateEndpoint (dict) -- [REQUIRED]
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
dict
The model configuration for the harness. Supports Amazon Bedrock, OpenAI, and Google Gemini model providers.
bedrockModelConfig (dict) --
Configuration for an Amazon Bedrock model.
modelId (string) -- [REQUIRED]
The Bedrock model ID.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
openAiModelConfig (dict) --
Configuration for an OpenAI model.
modelId (string) -- [REQUIRED]
The OpenAI model ID.
apiKeyArn (string) -- [REQUIRED]
The ARN of your OpenAI API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
geminiModelConfig (dict) --
Configuration for a Google Gemini model.
modelId (string) -- [REQUIRED]
The Gemini model ID.
apiKeyArn (string) -- [REQUIRED]
The ARN of your Gemini API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
topK (integer) --
The topK set when calling the model.
list
The system prompt that defines the agent's behavior and instructions.
(dict) --
A content block in the system prompt.
text (string) --
The text content of the system prompt block.
list
The tools available to the agent, such as remote MCP servers, AgentCore Gateway, AgentCore Browser, Code Interpreter, or inline functions.
(dict) --
A tool available to the agent loop.
type (string) -- [REQUIRED]
The type of tool.
name (string) --
Unique name for the tool. If not provided, a name will be inferred or generated.
config (dict) --
Tool-specific configuration.
remoteMcp (dict) --
Configuration for remote MCP server.
url (string) -- [REQUIRED]
URL of the MCP endpoint.
headers (dict) --
Custom headers to include when connecting to the remote MCP server.
(string) --
The key of an HTTP header.
(string) --
The value of an HTTP header.
agentCoreBrowser (dict) --
Configuration for AgentCore Browser.
browserArn (string) --
If not populated, the built-in Browser ARN is used.
agentCoreGateway (dict) --
Configuration for AgentCore Gateway.
gatewayArn (string) -- [REQUIRED]
The ARN of the desired AgentCore Gateway.
outboundAuth (dict) --
How harness authenticates to this Gateway. Defaults to AWS_IAM (SigV4) if omitted.
awsIam (dict) --
SigV4-sign requests using the agent's execution role.
none (dict) --
No authentication.
oauth (dict) --
An OAuth credential provider for gateway authentication. This structure contains the configuration for authenticating with the target endpoint using OAuth.
providerArn (string) -- [REQUIRED]
The Amazon Resource Name (ARN) of the OAuth credential provider. This ARN identifies the provider in Amazon Web Services.
scopes (list) -- [REQUIRED]
The OAuth scopes for the credential provider. These scopes define the level of access requested from the OAuth provider.
(string) --
customParameters (dict) --
The custom parameters for the OAuth credential provider. These parameters provide additional configuration for the OAuth authentication process.
(string) --
(string) --
grantType (string) --
Specifies the kind of credentials to use for authorization:
CLIENT_CREDENTIALS - Authorization with a client ID and secret.
AUTHORIZATION_CODE - Authorization with a token that is specific to an individual end user.
TOKEN_EXCHANGE - Authorization using on-behalf-of token exchange. An inbound user token is exchanged for a downstream access token scoped to the target audience.
defaultReturnUrl (string) --
The URL where the end user's browser is redirected after obtaining the authorization code. Generally points to the customer's application.
inlineFunction (dict) --
Configuration for an inline function tool.
description (string) -- [REQUIRED]
Description of what the tool does, provided to the model.
inputSchema (:ref:`document<document>`) -- [REQUIRED]
JSON Schema describing the tool's input parameters.
agentCoreCodeInterpreter (dict) --
Configuration for AgentCore Code Interpreter.
codeInterpreterArn (string) --
If not populated, the built-in Code Interpreter ARN is used.
list
The skills available to the agent. Skills are bundles of files that the agent can pull into its context on demand.
(dict) --
A skill available to the agent.
path (string) --
The filesystem path to the skill definition.
list
The tools that the agent is allowed to use. Supports glob patterns such as * for all tools, @builtin for all built-in tools, or @serverName/toolName for specific MCP server tools.
(string) --
dict
The AgentCore Memory configuration for persisting conversation context across sessions.
agentCoreMemoryConfiguration (dict) --
The AgentCore Memory configuration.
arn (string) -- [REQUIRED]
The ARN of the AgentCore Memory resource.
actorId (string) --
The actor ID for memory operations.
messagesCount (integer) --
The number of messages to retrieve from memory.
retrievalConfig (dict) --
The retrieval configuration for long-term memory, mapping namespace path templates to retrieval settings.
(string) --
(dict) --
Configuration for memory retrieval within a namespace.
topK (integer) --
The maximum number of memory entries to retrieve.
relevanceScore (float) --
The minimum relevance score for retrieved memories.
strategyId (string) --
The ID of the retrieval strategy to use.
dict
The truncation configuration for managing conversation context when it exceeds model limits.
strategy (string) -- [REQUIRED]
The truncation strategy to use.
config (dict) --
The strategy-specific configuration.
slidingWindow (dict) --
Configuration for sliding window truncation.
messagesCount (integer) --
The number of recent messages to retain in the context window.
summarization (dict) --
Configuration for summarization-based truncation.
summaryRatio (float) --
The ratio of content to summarize.
preserveRecentMessages (integer) --
The number of recent messages to preserve without summarization.
summarizationSystemPrompt (string) --
The system prompt used for generating summaries.
integer
The maximum number of iterations the agent loop can execute per invocation.
integer
The maximum total number of output tokens the agent can generate across all model calls within a single invocation.
integer
The maximum duration in seconds for the agent loop execution per invocation.
dict
Tags to apply to the harness resource.
(string) --
(string) --
dict
Response Syntax
{
'harness': {
'harnessId': 'string',
'harnessName': 'string',
'arn': 'string',
'status': 'CREATING'|'CREATE_FAILED'|'UPDATING'|'UPDATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED',
'executionRoleArn': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'model': {
'bedrockModelConfig': {
'modelId': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'openAiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'geminiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...,
'topK': 123
}
},
'systemPrompt': [
{
'text': 'string'
},
],
'tools': [
{
'type': 'remote_mcp'|'agentcore_browser'|'agentcore_gateway'|'inline_function'|'agentcore_code_interpreter',
'name': 'string',
'config': {
'remoteMcp': {
'url': 'string',
'headers': {
'string': 'string'
}
},
'agentCoreBrowser': {
'browserArn': 'string'
},
'agentCoreGateway': {
'gatewayArn': 'string',
'outboundAuth': {
'awsIam': {},
'none': {},
'oauth': {
'providerArn': 'string',
'scopes': [
'string',
],
'customParameters': {
'string': 'string'
},
'grantType': 'CLIENT_CREDENTIALS'|'AUTHORIZATION_CODE'|'TOKEN_EXCHANGE',
'defaultReturnUrl': 'string'
}
}
},
'inlineFunction': {
'description': 'string',
'inputSchema': {...}|[...]|123|123.4|'string'|True|None
},
'agentCoreCodeInterpreter': {
'codeInterpreterArn': 'string'
}
}
},
],
'skills': [
{
'path': 'string'
},
],
'allowedTools': [
'string',
],
'truncation': {
'strategy': 'sliding_window'|'summarization'|'none',
'config': {
'slidingWindow': {
'messagesCount': 123
},
'summarization': {
'summaryRatio': ...,
'preserveRecentMessages': 123,
'summarizationSystemPrompt': 'string'
}
}
},
'environment': {
'agentCoreRuntimeEnvironment': {
'agentRuntimeArn': 'string',
'agentRuntimeName': 'string',
'agentRuntimeId': 'string',
'lifecycleConfiguration': {
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'filesystemConfigurations': [
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
]
}
},
'environmentArtifact': {
'containerConfiguration': {
'containerUri': 'string'
}
},
'environmentVariables': {
'string': 'string'
},
'authorizerConfiguration': {
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
'memory': {
'agentCoreMemoryConfiguration': {
'arn': 'string',
'actorId': 'string',
'messagesCount': 123,
'retrievalConfig': {
'string': {
'topK': 123,
'relevanceScore': ...,
'strategyId': 'string'
}
}
}
},
'maxIterations': 123,
'maxTokens': 123,
'timeoutSeconds': 123,
'failureReason': 'string'
}
}
Response Structure
(dict) --
harness (dict) --
The harness that was created.
harnessId (string) --
The ID of the Harness.
harnessName (string) --
The name of the Harness.
arn (string) --
The ARN of the Harness.
status (string) --
The status of the Harness.
executionRoleArn (string) --
IAM role the Harness assumes when running.
createdAt (datetime) --
The createdAt time of the Harness.
updatedAt (datetime) --
The updatedAt time of the Harness.
model (dict) --
The configuration of the default model used by the Harness.
bedrockModelConfig (dict) --
Configuration for an Amazon Bedrock model.
modelId (string) --
The Bedrock model ID.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
openAiModelConfig (dict) --
Configuration for an OpenAI model.
modelId (string) --
The OpenAI model ID.
apiKeyArn (string) --
The ARN of your OpenAI API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
geminiModelConfig (dict) --
Configuration for a Google Gemini model.
modelId (string) --
The Gemini model ID.
apiKeyArn (string) --
The ARN of your Gemini API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
topK (integer) --
The topK set when calling the model.
systemPrompt (list) --
The system prompt of the Harness.
(dict) --
A content block in the system prompt.
text (string) --
The text content of the system prompt block.
tools (list) --
The tools of the Harness.
(dict) --
A tool available to the agent loop.
type (string) --
The type of tool.
name (string) --
Unique name for the tool. If not provided, a name will be inferred or generated.
config (dict) --
Tool-specific configuration.
remoteMcp (dict) --
Configuration for remote MCP server.
url (string) --
URL of the MCP endpoint.
headers (dict) --
Custom headers to include when connecting to the remote MCP server.
(string) --
The key of an HTTP header.
(string) --
The value of an HTTP header.
agentCoreBrowser (dict) --
Configuration for AgentCore Browser.
browserArn (string) --
If not populated, the built-in Browser ARN is used.
agentCoreGateway (dict) --
Configuration for AgentCore Gateway.
gatewayArn (string) --
The ARN of the desired AgentCore Gateway.
outboundAuth (dict) --
How harness authenticates to this Gateway. Defaults to AWS_IAM (SigV4) if omitted.
awsIam (dict) --
SigV4-sign requests using the agent's execution role.
none (dict) --
No authentication.
oauth (dict) --
An OAuth credential provider for gateway authentication. This structure contains the configuration for authenticating with the target endpoint using OAuth.
providerArn (string) --
The Amazon Resource Name (ARN) of the OAuth credential provider. This ARN identifies the provider in Amazon Web Services.
scopes (list) --
The OAuth scopes for the credential provider. These scopes define the level of access requested from the OAuth provider.
(string) --
customParameters (dict) --
The custom parameters for the OAuth credential provider. These parameters provide additional configuration for the OAuth authentication process.
(string) --
(string) --
grantType (string) --
Specifies the kind of credentials to use for authorization:
CLIENT_CREDENTIALS - Authorization with a client ID and secret.
AUTHORIZATION_CODE - Authorization with a token that is specific to an individual end user.
TOKEN_EXCHANGE - Authorization using on-behalf-of token exchange. An inbound user token is exchanged for a downstream access token scoped to the target audience.
defaultReturnUrl (string) --
The URL where the end user's browser is redirected after obtaining the authorization code. Generally points to the customer's application.
inlineFunction (dict) --
Configuration for an inline function tool.
description (string) --
Description of what the tool does, provided to the model.
inputSchema (:ref:`document<document>`) --
JSON Schema describing the tool's input parameters.
agentCoreCodeInterpreter (dict) --
Configuration for AgentCore Code Interpreter.
codeInterpreterArn (string) --
If not populated, the built-in Code Interpreter ARN is used.
skills (list) --
The skills of the Harness.
(dict) --
A skill available to the agent.
path (string) --
The filesystem path to the skill definition.
allowedTools (list) --
The allowed tools of the Harness. All tools are allowed by default.
(string) --
truncation (dict) --
Configuration for truncating model context.
strategy (string) --
The truncation strategy to use.
config (dict) --
The strategy-specific configuration.
slidingWindow (dict) --
Configuration for sliding window truncation.
messagesCount (integer) --
The number of recent messages to retain in the context window.
summarization (dict) --
Configuration for summarization-based truncation.
summaryRatio (float) --
The ratio of content to summarize.
preserveRecentMessages (integer) --
The number of recent messages to preserve without summarization.
summarizationSystemPrompt (string) --
The system prompt used for generating summaries.
environment (dict) --
The compute environment on which the Harness runs.
agentCoreRuntimeEnvironment (dict) --
The AgentCore Runtime environment configuration.
agentRuntimeArn (string) --
The ARN of the underlying AgentCore Runtime.
agentRuntimeName (string) --
The name of the underlying AgentCore Runtime.
agentRuntimeId (string) --
The ID of the underlying AgentCore Runtime.
lifecycleConfiguration (dict) --
LifecycleConfiguration lets you manage the lifecycle of runtime sessions and resources in AgentCore Runtime. This configuration helps optimize resource utilization by automatically cleaning up idle sessions and preventing long-running instances from consuming resources indefinitely.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
networkConfiguration (dict) --
SecurityConfig for the Agent.
networkMode (string) --
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) --
The security groups associated with the VPC configuration.
(string) --
subnets (list) --
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
filesystemConfigurations (list) --
The filesystem configurations for the runtime environment.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) --
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
environmentArtifact (dict) --
The environment artifact (e.g., container) in which the Harness operates.
containerConfiguration (dict) --
Representation of a container configuration.
containerUri (string) --
The ECR URI of the container.
environmentVariables (dict) --
Environment variables exposed in the environment in which the Harness operates.
(string) --
(string) --
authorizerConfiguration (dict) --
Represents inbound authorization configuration options used to authenticate incoming requests.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) --
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) --
The name of the custom claim field to check.
inboundTokenClaimValueType (string) --
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) --
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) --
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) --
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) --
The domain to override with a private endpoint.
privateEndpoint (dict) --
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
memory (dict) --
AgentCore Memory instance configuration for short and long term memory.
agentCoreMemoryConfiguration (dict) --
The AgentCore Memory configuration.
arn (string) --
The ARN of the AgentCore Memory resource.
actorId (string) --
The actor ID for memory operations.
messagesCount (integer) --
The number of messages to retrieve from memory.
retrievalConfig (dict) --
The retrieval configuration for long-term memory, mapping namespace path templates to retrieval settings.
(string) --
(dict) --
Configuration for memory retrieval within a namespace.
topK (integer) --
The maximum number of memory entries to retrieve.
relevanceScore (float) --
The minimum relevance score for retrieved memories.
strategyId (string) --
The ID of the retrieval strategy to use.
maxIterations (integer) --
The maximum number of iterations in the agent loop allowed before exiting per invocation.
maxTokens (integer) --
The maximum total number of output tokens the agent can generate across all model calls within a single invocation.
timeoutSeconds (integer) --
The maximum duration per invocation.
failureReason (string) --
Reason why create or update operations fail.
{'harness': {'environment': {'agentCoreRuntimeEnvironment': {'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}}}}
Operation to delete a Harness.
See also: AWS API Documentation
Request Syntax
client.delete_harness(
harnessId='string',
clientToken='string'
)
string
[REQUIRED]
The ID of the harness to delete.
string
A unique, case-sensitive identifier to ensure idempotency of the request.
This field is autopopulated if not provided.
dict
Response Syntax
{
'harness': {
'harnessId': 'string',
'harnessName': 'string',
'arn': 'string',
'status': 'CREATING'|'CREATE_FAILED'|'UPDATING'|'UPDATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED',
'executionRoleArn': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'model': {
'bedrockModelConfig': {
'modelId': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'openAiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'geminiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...,
'topK': 123
}
},
'systemPrompt': [
{
'text': 'string'
},
],
'tools': [
{
'type': 'remote_mcp'|'agentcore_browser'|'agentcore_gateway'|'inline_function'|'agentcore_code_interpreter',
'name': 'string',
'config': {
'remoteMcp': {
'url': 'string',
'headers': {
'string': 'string'
}
},
'agentCoreBrowser': {
'browserArn': 'string'
},
'agentCoreGateway': {
'gatewayArn': 'string',
'outboundAuth': {
'awsIam': {},
'none': {},
'oauth': {
'providerArn': 'string',
'scopes': [
'string',
],
'customParameters': {
'string': 'string'
},
'grantType': 'CLIENT_CREDENTIALS'|'AUTHORIZATION_CODE'|'TOKEN_EXCHANGE',
'defaultReturnUrl': 'string'
}
}
},
'inlineFunction': {
'description': 'string',
'inputSchema': {...}|[...]|123|123.4|'string'|True|None
},
'agentCoreCodeInterpreter': {
'codeInterpreterArn': 'string'
}
}
},
],
'skills': [
{
'path': 'string'
},
],
'allowedTools': [
'string',
],
'truncation': {
'strategy': 'sliding_window'|'summarization'|'none',
'config': {
'slidingWindow': {
'messagesCount': 123
},
'summarization': {
'summaryRatio': ...,
'preserveRecentMessages': 123,
'summarizationSystemPrompt': 'string'
}
}
},
'environment': {
'agentCoreRuntimeEnvironment': {
'agentRuntimeArn': 'string',
'agentRuntimeName': 'string',
'agentRuntimeId': 'string',
'lifecycleConfiguration': {
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'filesystemConfigurations': [
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
]
}
},
'environmentArtifact': {
'containerConfiguration': {
'containerUri': 'string'
}
},
'environmentVariables': {
'string': 'string'
},
'authorizerConfiguration': {
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
'memory': {
'agentCoreMemoryConfiguration': {
'arn': 'string',
'actorId': 'string',
'messagesCount': 123,
'retrievalConfig': {
'string': {
'topK': 123,
'relevanceScore': ...,
'strategyId': 'string'
}
}
}
},
'maxIterations': 123,
'maxTokens': 123,
'timeoutSeconds': 123,
'failureReason': 'string'
}
}
Response Structure
(dict) --
harness (dict) --
The harness that was deleted.
harnessId (string) --
The ID of the Harness.
harnessName (string) --
The name of the Harness.
arn (string) --
The ARN of the Harness.
status (string) --
The status of the Harness.
executionRoleArn (string) --
IAM role the Harness assumes when running.
createdAt (datetime) --
The createdAt time of the Harness.
updatedAt (datetime) --
The updatedAt time of the Harness.
model (dict) --
The configuration of the default model used by the Harness.
bedrockModelConfig (dict) --
Configuration for an Amazon Bedrock model.
modelId (string) --
The Bedrock model ID.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
openAiModelConfig (dict) --
Configuration for an OpenAI model.
modelId (string) --
The OpenAI model ID.
apiKeyArn (string) --
The ARN of your OpenAI API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
geminiModelConfig (dict) --
Configuration for a Google Gemini model.
modelId (string) --
The Gemini model ID.
apiKeyArn (string) --
The ARN of your Gemini API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
topK (integer) --
The topK set when calling the model.
systemPrompt (list) --
The system prompt of the Harness.
(dict) --
A content block in the system prompt.
text (string) --
The text content of the system prompt block.
tools (list) --
The tools of the Harness.
(dict) --
A tool available to the agent loop.
type (string) --
The type of tool.
name (string) --
Unique name for the tool. If not provided, a name will be inferred or generated.
config (dict) --
Tool-specific configuration.
remoteMcp (dict) --
Configuration for remote MCP server.
url (string) --
URL of the MCP endpoint.
headers (dict) --
Custom headers to include when connecting to the remote MCP server.
(string) --
The key of an HTTP header.
(string) --
The value of an HTTP header.
agentCoreBrowser (dict) --
Configuration for AgentCore Browser.
browserArn (string) --
If not populated, the built-in Browser ARN is used.
agentCoreGateway (dict) --
Configuration for AgentCore Gateway.
gatewayArn (string) --
The ARN of the desired AgentCore Gateway.
outboundAuth (dict) --
How harness authenticates to this Gateway. Defaults to AWS_IAM (SigV4) if omitted.
awsIam (dict) --
SigV4-sign requests using the agent's execution role.
none (dict) --
No authentication.
oauth (dict) --
An OAuth credential provider for gateway authentication. This structure contains the configuration for authenticating with the target endpoint using OAuth.
providerArn (string) --
The Amazon Resource Name (ARN) of the OAuth credential provider. This ARN identifies the provider in Amazon Web Services.
scopes (list) --
The OAuth scopes for the credential provider. These scopes define the level of access requested from the OAuth provider.
(string) --
customParameters (dict) --
The custom parameters for the OAuth credential provider. These parameters provide additional configuration for the OAuth authentication process.
(string) --
(string) --
grantType (string) --
Specifies the kind of credentials to use for authorization:
CLIENT_CREDENTIALS - Authorization with a client ID and secret.
AUTHORIZATION_CODE - Authorization with a token that is specific to an individual end user.
TOKEN_EXCHANGE - Authorization using on-behalf-of token exchange. An inbound user token is exchanged for a downstream access token scoped to the target audience.
defaultReturnUrl (string) --
The URL where the end user's browser is redirected after obtaining the authorization code. Generally points to the customer's application.
inlineFunction (dict) --
Configuration for an inline function tool.
description (string) --
Description of what the tool does, provided to the model.
inputSchema (:ref:`document<document>`) --
JSON Schema describing the tool's input parameters.
agentCoreCodeInterpreter (dict) --
Configuration for AgentCore Code Interpreter.
codeInterpreterArn (string) --
If not populated, the built-in Code Interpreter ARN is used.
skills (list) --
The skills of the Harness.
(dict) --
A skill available to the agent.
path (string) --
The filesystem path to the skill definition.
allowedTools (list) --
The allowed tools of the Harness. All tools are allowed by default.
(string) --
truncation (dict) --
Configuration for truncating model context.
strategy (string) --
The truncation strategy to use.
config (dict) --
The strategy-specific configuration.
slidingWindow (dict) --
Configuration for sliding window truncation.
messagesCount (integer) --
The number of recent messages to retain in the context window.
summarization (dict) --
Configuration for summarization-based truncation.
summaryRatio (float) --
The ratio of content to summarize.
preserveRecentMessages (integer) --
The number of recent messages to preserve without summarization.
summarizationSystemPrompt (string) --
The system prompt used for generating summaries.
environment (dict) --
The compute environment on which the Harness runs.
agentCoreRuntimeEnvironment (dict) --
The AgentCore Runtime environment configuration.
agentRuntimeArn (string) --
The ARN of the underlying AgentCore Runtime.
agentRuntimeName (string) --
The name of the underlying AgentCore Runtime.
agentRuntimeId (string) --
The ID of the underlying AgentCore Runtime.
lifecycleConfiguration (dict) --
LifecycleConfiguration lets you manage the lifecycle of runtime sessions and resources in AgentCore Runtime. This configuration helps optimize resource utilization by automatically cleaning up idle sessions and preventing long-running instances from consuming resources indefinitely.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
networkConfiguration (dict) --
SecurityConfig for the Agent.
networkMode (string) --
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) --
The security groups associated with the VPC configuration.
(string) --
subnets (list) --
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
filesystemConfigurations (list) --
The filesystem configurations for the runtime environment.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) --
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
environmentArtifact (dict) --
The environment artifact (e.g., container) in which the Harness operates.
containerConfiguration (dict) --
Representation of a container configuration.
containerUri (string) --
The ECR URI of the container.
environmentVariables (dict) --
Environment variables exposed in the environment in which the Harness operates.
(string) --
(string) --
authorizerConfiguration (dict) --
Represents inbound authorization configuration options used to authenticate incoming requests.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) --
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) --
The name of the custom claim field to check.
inboundTokenClaimValueType (string) --
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) --
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) --
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) --
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) --
The domain to override with a private endpoint.
privateEndpoint (dict) --
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
memory (dict) --
AgentCore Memory instance configuration for short and long term memory.
agentCoreMemoryConfiguration (dict) --
The AgentCore Memory configuration.
arn (string) --
The ARN of the AgentCore Memory resource.
actorId (string) --
The actor ID for memory operations.
messagesCount (integer) --
The number of messages to retrieve from memory.
retrievalConfig (dict) --
The retrieval configuration for long-term memory, mapping namespace path templates to retrieval settings.
(string) --
(dict) --
Configuration for memory retrieval within a namespace.
topK (integer) --
The maximum number of memory entries to retrieve.
relevanceScore (float) --
The minimum relevance score for retrieved memories.
strategyId (string) --
The ID of the retrieval strategy to use.
maxIterations (integer) --
The maximum number of iterations in the agent loop allowed before exiting per invocation.
maxTokens (integer) --
The maximum total number of output tokens the agent can generate across all model calls within a single invocation.
timeoutSeconds (integer) --
The maximum duration per invocation.
failureReason (string) --
Reason why create or update operations fail.
{'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}
Gets an Amazon Bedrock AgentCore Runtime.
See also: AWS API Documentation
Request Syntax
client.get_agent_runtime(
agentRuntimeId='string',
agentRuntimeVersion='string'
)
string
[REQUIRED]
The unique identifier of the AgentCore Runtime to retrieve.
string
The version of the AgentCore Runtime to retrieve.
dict
Response Syntax
{
'agentRuntimeArn': 'string',
'agentRuntimeName': 'string',
'agentRuntimeId': 'string',
'agentRuntimeVersion': 'string',
'createdAt': datetime(2015, 1, 1),
'lastUpdatedAt': datetime(2015, 1, 1),
'roleArn': 'string',
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'status': 'CREATING'|'CREATE_FAILED'|'UPDATING'|'UPDATE_FAILED'|'READY'|'DELETING',
'lifecycleConfiguration': {
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
'failureReason': 'string',
'description': 'string',
'workloadIdentityDetails': {
'workloadIdentityArn': 'string'
},
'agentRuntimeArtifact': {
'containerConfiguration': {
'containerUri': 'string'
},
'codeConfiguration': {
'code': {
's3': {
'bucket': 'string',
'prefix': 'string',
'versionId': 'string'
}
},
'runtime': 'PYTHON_3_10'|'PYTHON_3_11'|'PYTHON_3_12'|'PYTHON_3_13'|'PYTHON_3_14'|'NODE_22',
'entryPoint': [
'string',
]
}
},
'protocolConfiguration': {
'serverProtocol': 'MCP'|'HTTP'|'A2A'|'AGUI'
},
'environmentVariables': {
'string': 'string'
},
'authorizerConfiguration': {
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
'requestHeaderConfiguration': {
'requestHeaderAllowlist': [
'string',
]
},
'metadataConfiguration': {
'requireMMDSV2': True|False
},
'filesystemConfigurations': [
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
]
}
Response Structure
(dict) --
agentRuntimeArn (string) --
The Amazon Resource Name (ARN) of the AgentCore Runtime.
agentRuntimeName (string) --
The name of the AgentCore Runtime.
agentRuntimeId (string) --
The unique identifier of the AgentCore Runtime.
agentRuntimeVersion (string) --
The version of the AgentCore Runtime.
createdAt (datetime) --
The timestamp when the AgentCore Runtime was created.
lastUpdatedAt (datetime) --
The timestamp when the AgentCore Runtime was last updated.
roleArn (string) --
The IAM role ARN that provides permissions for the AgentCore Runtime.
networkConfiguration (dict) --
The network configuration for the AgentCore Runtime.
networkMode (string) --
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) --
The security groups associated with the VPC configuration.
(string) --
subnets (list) --
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
status (string) --
The current status of the AgentCore Runtime.
lifecycleConfiguration (dict) --
The life cycle configuration for the AgentCore Runtime.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
failureReason (string) --
The reason for failure if the AgentCore Runtime is in a failed state.
description (string) --
The description of the AgentCore Runtime.
workloadIdentityDetails (dict) --
The workload identity details for the AgentCore Runtime.
workloadIdentityArn (string) --
The ARN associated with the workload identity.
agentRuntimeArtifact (dict) --
The artifact of the AgentCore Runtime.
containerConfiguration (dict) --
The container configuration for the agent artifact.
containerUri (string) --
The ECR URI of the container.
codeConfiguration (dict) --
The code configuration for the agent runtime artifact, including the source code location and execution settings.
code (dict) --
The source code location and configuration details.
s3 (dict) --
The Amazon Amazon S3 object that contains the source code for the agent runtime.
bucket (string) --
The name of the Amazon S3 bucket. This bucket contains the stored data.
prefix (string) --
The prefix for objects in the Amazon S3 bucket. This prefix is added to the object keys to organize the data.
versionId (string) --
The version ID of the Amazon Amazon S3 object. If not specified, the latest version of the object is used.
runtime (string) --
The runtime environment for executing the agent code. Specify the programming language and version to use for the agent runtime. For valid values, see the list of supported runtimes.
entryPoint (list) --
The entry point for the code execution, specifying the function or method that should be invoked when the code runs.
(string) --
protocolConfiguration (dict) --
The protocol configuration for an agent runtime. This structure defines how the agent runtime communicates with clients.
serverProtocol (string) --
The server protocol for the agent runtime. This field specifies which protocol the agent runtime uses to communicate with clients.
environmentVariables (dict) --
Environment variables set in the AgentCore Runtime environment.
(string) --
(string) --
authorizerConfiguration (dict) --
The authorizer configuration for the AgentCore Runtime.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) --
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) --
The name of the custom claim field to check.
inboundTokenClaimValueType (string) --
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) --
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) --
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) --
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) --
The domain to override with a private endpoint.
privateEndpoint (dict) --
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
requestHeaderConfiguration (dict) --
Configuration for HTTP request headers that will be passed through to the runtime.
requestHeaderAllowlist (list) --
A list of HTTP request headers that are allowed to be passed through to the runtime.
(string) --
metadataConfiguration (dict) --
Configuration for microVM Metadata Service (MMDS) settings for the AgentCore Runtime.
requireMMDSV2 (boolean) --
Enables MMDSv2 (microVM Metadata Service Version 2) requirement for the agent runtime. When set to true, the runtime microVM will only accept MMDSv2 requests.
filesystemConfigurations (list) --
The filesystem configurations mounted into the AgentCore Runtime.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) --
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
{'networkConfiguration': {'vpcConfig': {'requireServiceS3Endpoint': 'boolean'}}}
Gets information about a custom browser.
See also: AWS API Documentation
Request Syntax
client.get_browser(
browserId='string'
)
string
[REQUIRED]
The unique identifier of the browser to retrieve.
dict
Response Syntax
{
'browserId': 'string',
'browserArn': 'string',
'name': 'string',
'description': 'string',
'executionRoleArn': 'string',
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'vpcConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'recording': {
'enabled': True|False,
's3Location': {
'bucket': 'string',
'prefix': 'string',
'versionId': 'string'
}
},
'browserSigning': {
'enabled': True|False
},
'enterprisePolicies': [
{
'location': {
's3': {
'bucket': 'string',
'prefix': 'string',
'versionId': 'string'
}
},
'type': 'MANAGED'|'RECOMMENDED'
},
],
'certificates': [
{
'location': {
'secretsManager': {
'secretArn': 'string'
}
}
},
],
'status': 'CREATING'|'CREATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED'|'DELETED',
'failureReason': 'string',
'createdAt': datetime(2015, 1, 1),
'lastUpdatedAt': datetime(2015, 1, 1)
}
Response Structure
(dict) --
browserId (string) --
The unique identifier of the browser.
browserArn (string) --
The Amazon Resource Name (ARN) of the browser.
name (string) --
The name of the browser.
description (string) --
The description of the browser.
executionRoleArn (string) --
The IAM role ARN that provides permissions for the browser.
networkConfiguration (dict) --
The network configuration for a browser. This structure defines how the browser connects to the network.
networkMode (string) --
The network mode for the browser. This field specifies how the browser connects to the network.
vpcConfig (dict) --
The VPC configuration for the browser. This configuration is required when the network mode is set to VPC.
securityGroups (list) --
The security groups associated with the VPC configuration.
(string) --
subnets (list) --
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
recording (dict) --
The recording configuration for a browser. This structure defines how browser sessions are recorded.
enabled (boolean) --
Indicates whether recording is enabled for the browser. When set to true, browser sessions are recorded.
s3Location (dict) --
The Amazon S3 location where browser recordings are stored. This location contains the recorded browser sessions.
bucket (string) --
The name of the Amazon S3 bucket. This bucket contains the stored data.
prefix (string) --
The prefix for objects in the Amazon S3 bucket. This prefix is added to the object keys to organize the data.
versionId (string) --
The version ID of the Amazon Amazon S3 object. If not specified, the latest version of the object is used.
browserSigning (dict) --
The browser signing configuration that shows whether cryptographic agent identification is enabled for web bot authentication.
enabled (boolean) --
Indicates whether browser signing is currently enabled for cryptographic agent identification using HTTP message signatures.
enterprisePolicies (list) --
The list of enterprise policy files configured for the browser.
(dict) --
Browser enterprise policy configuration.
location (dict) --
The location of the enterprise policy file.
s3 (dict) --
The Amazon S3 location for storing data. This structure defines where in Amazon S3 data is stored.
bucket (string) --
The name of the Amazon S3 bucket. This bucket contains the stored data.
prefix (string) --
The prefix for objects in the Amazon S3 bucket. This prefix is added to the object keys to organize the data.
versionId (string) --
The version ID of the Amazon Amazon S3 object. If not specified, the latest version of the object is used.
type (string) --
The type of browser enterprise policy. Available values are MANAGED and RECOMMENDED.
certificates (list) --
The list of certificates configured for the browser.
(dict) --
A certificate to install in the browser or code interpreter.
location (dict) --
The location of the certificate.
secretsManager (dict) --
The Amazon Web Services Secrets Manager location of the certificate.
secretArn (string) --
The ARN of the Amazon Web Services Secrets Manager secret containing the certificate.
status (string) --
The current status of the browser.
failureReason (string) --
The reason for failure if the browser is in a failed state.
createdAt (datetime) --
The timestamp when the browser was created.
lastUpdatedAt (datetime) --
The timestamp when the browser was last updated.
{'networkConfiguration': {'vpcConfig': {'requireServiceS3Endpoint': 'boolean'}}}
Gets information about a custom code interpreter.
See also: AWS API Documentation
Request Syntax
client.get_code_interpreter(
codeInterpreterId='string'
)
string
[REQUIRED]
The unique identifier of the code interpreter to retrieve.
dict
Response Syntax
{
'codeInterpreterId': 'string',
'codeInterpreterArn': 'string',
'name': 'string',
'description': 'string',
'executionRoleArn': 'string',
'networkConfiguration': {
'networkMode': 'PUBLIC'|'SANDBOX'|'VPC',
'vpcConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'status': 'CREATING'|'CREATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED'|'DELETED',
'certificates': [
{
'location': {
'secretsManager': {
'secretArn': 'string'
}
}
},
],
'failureReason': 'string',
'createdAt': datetime(2015, 1, 1),
'lastUpdatedAt': datetime(2015, 1, 1)
}
Response Structure
(dict) --
codeInterpreterId (string) --
The unique identifier of the code interpreter.
codeInterpreterArn (string) --
The Amazon Resource Name (ARN) of the code interpreter.
name (string) --
The name of the code interpreter.
description (string) --
The description of the code interpreter.
executionRoleArn (string) --
The IAM role ARN that provides permissions for the code interpreter.
networkConfiguration (dict) --
The network configuration for a code interpreter. This structure defines how the code interpreter connects to the network.
networkMode (string) --
The network mode for the code interpreter. This field specifies how the code interpreter connects to the network.
vpcConfig (dict) --
The VPC configuration for the code interpreter. This configuration is required when the network mode is set to VPC.
securityGroups (list) --
The security groups associated with the VPC configuration.
(string) --
subnets (list) --
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
status (string) --
The current status of the code interpreter.
certificates (list) --
The list of certificates configured for the code interpreter.
(dict) --
A certificate to install in the browser or code interpreter.
location (dict) --
The location of the certificate.
secretsManager (dict) --
The Amazon Web Services Secrets Manager location of the certificate.
secretArn (string) --
The ARN of the Amazon Web Services Secrets Manager secret containing the certificate.
failureReason (string) --
The reason for failure if the code interpreter is in a failed state.
createdAt (datetime) --
The timestamp when the code interpreter was created.
lastUpdatedAt (datetime) --
The timestamp when the code interpreter was last updated.
{'harness': {'environment': {'agentCoreRuntimeEnvironment': {'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}}}}
Operation to get a single Harness.
See also: AWS API Documentation
Request Syntax
client.get_harness(
harnessId='string'
)
string
[REQUIRED]
The ID of the harness to retrieve.
dict
Response Syntax
{
'harness': {
'harnessId': 'string',
'harnessName': 'string',
'arn': 'string',
'status': 'CREATING'|'CREATE_FAILED'|'UPDATING'|'UPDATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED',
'executionRoleArn': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'model': {
'bedrockModelConfig': {
'modelId': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'openAiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'geminiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...,
'topK': 123
}
},
'systemPrompt': [
{
'text': 'string'
},
],
'tools': [
{
'type': 'remote_mcp'|'agentcore_browser'|'agentcore_gateway'|'inline_function'|'agentcore_code_interpreter',
'name': 'string',
'config': {
'remoteMcp': {
'url': 'string',
'headers': {
'string': 'string'
}
},
'agentCoreBrowser': {
'browserArn': 'string'
},
'agentCoreGateway': {
'gatewayArn': 'string',
'outboundAuth': {
'awsIam': {},
'none': {},
'oauth': {
'providerArn': 'string',
'scopes': [
'string',
],
'customParameters': {
'string': 'string'
},
'grantType': 'CLIENT_CREDENTIALS'|'AUTHORIZATION_CODE'|'TOKEN_EXCHANGE',
'defaultReturnUrl': 'string'
}
}
},
'inlineFunction': {
'description': 'string',
'inputSchema': {...}|[...]|123|123.4|'string'|True|None
},
'agentCoreCodeInterpreter': {
'codeInterpreterArn': 'string'
}
}
},
],
'skills': [
{
'path': 'string'
},
],
'allowedTools': [
'string',
],
'truncation': {
'strategy': 'sliding_window'|'summarization'|'none',
'config': {
'slidingWindow': {
'messagesCount': 123
},
'summarization': {
'summaryRatio': ...,
'preserveRecentMessages': 123,
'summarizationSystemPrompt': 'string'
}
}
},
'environment': {
'agentCoreRuntimeEnvironment': {
'agentRuntimeArn': 'string',
'agentRuntimeName': 'string',
'agentRuntimeId': 'string',
'lifecycleConfiguration': {
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'filesystemConfigurations': [
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
]
}
},
'environmentArtifact': {
'containerConfiguration': {
'containerUri': 'string'
}
},
'environmentVariables': {
'string': 'string'
},
'authorizerConfiguration': {
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
'memory': {
'agentCoreMemoryConfiguration': {
'arn': 'string',
'actorId': 'string',
'messagesCount': 123,
'retrievalConfig': {
'string': {
'topK': 123,
'relevanceScore': ...,
'strategyId': 'string'
}
}
}
},
'maxIterations': 123,
'maxTokens': 123,
'timeoutSeconds': 123,
'failureReason': 'string'
}
}
Response Structure
(dict) --
harness (dict) --
The harness resource.
harnessId (string) --
The ID of the Harness.
harnessName (string) --
The name of the Harness.
arn (string) --
The ARN of the Harness.
status (string) --
The status of the Harness.
executionRoleArn (string) --
IAM role the Harness assumes when running.
createdAt (datetime) --
The createdAt time of the Harness.
updatedAt (datetime) --
The updatedAt time of the Harness.
model (dict) --
The configuration of the default model used by the Harness.
bedrockModelConfig (dict) --
Configuration for an Amazon Bedrock model.
modelId (string) --
The Bedrock model ID.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
openAiModelConfig (dict) --
Configuration for an OpenAI model.
modelId (string) --
The OpenAI model ID.
apiKeyArn (string) --
The ARN of your OpenAI API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
geminiModelConfig (dict) --
Configuration for a Google Gemini model.
modelId (string) --
The Gemini model ID.
apiKeyArn (string) --
The ARN of your Gemini API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
topK (integer) --
The topK set when calling the model.
systemPrompt (list) --
The system prompt of the Harness.
(dict) --
A content block in the system prompt.
text (string) --
The text content of the system prompt block.
tools (list) --
The tools of the Harness.
(dict) --
A tool available to the agent loop.
type (string) --
The type of tool.
name (string) --
Unique name for the tool. If not provided, a name will be inferred or generated.
config (dict) --
Tool-specific configuration.
remoteMcp (dict) --
Configuration for remote MCP server.
url (string) --
URL of the MCP endpoint.
headers (dict) --
Custom headers to include when connecting to the remote MCP server.
(string) --
The key of an HTTP header.
(string) --
The value of an HTTP header.
agentCoreBrowser (dict) --
Configuration for AgentCore Browser.
browserArn (string) --
If not populated, the built-in Browser ARN is used.
agentCoreGateway (dict) --
Configuration for AgentCore Gateway.
gatewayArn (string) --
The ARN of the desired AgentCore Gateway.
outboundAuth (dict) --
How harness authenticates to this Gateway. Defaults to AWS_IAM (SigV4) if omitted.
awsIam (dict) --
SigV4-sign requests using the agent's execution role.
none (dict) --
No authentication.
oauth (dict) --
An OAuth credential provider for gateway authentication. This structure contains the configuration for authenticating with the target endpoint using OAuth.
providerArn (string) --
The Amazon Resource Name (ARN) of the OAuth credential provider. This ARN identifies the provider in Amazon Web Services.
scopes (list) --
The OAuth scopes for the credential provider. These scopes define the level of access requested from the OAuth provider.
(string) --
customParameters (dict) --
The custom parameters for the OAuth credential provider. These parameters provide additional configuration for the OAuth authentication process.
(string) --
(string) --
grantType (string) --
Specifies the kind of credentials to use for authorization:
CLIENT_CREDENTIALS - Authorization with a client ID and secret.
AUTHORIZATION_CODE - Authorization with a token that is specific to an individual end user.
TOKEN_EXCHANGE - Authorization using on-behalf-of token exchange. An inbound user token is exchanged for a downstream access token scoped to the target audience.
defaultReturnUrl (string) --
The URL where the end user's browser is redirected after obtaining the authorization code. Generally points to the customer's application.
inlineFunction (dict) --
Configuration for an inline function tool.
description (string) --
Description of what the tool does, provided to the model.
inputSchema (:ref:`document<document>`) --
JSON Schema describing the tool's input parameters.
agentCoreCodeInterpreter (dict) --
Configuration for AgentCore Code Interpreter.
codeInterpreterArn (string) --
If not populated, the built-in Code Interpreter ARN is used.
skills (list) --
The skills of the Harness.
(dict) --
A skill available to the agent.
path (string) --
The filesystem path to the skill definition.
allowedTools (list) --
The allowed tools of the Harness. All tools are allowed by default.
(string) --
truncation (dict) --
Configuration for truncating model context.
strategy (string) --
The truncation strategy to use.
config (dict) --
The strategy-specific configuration.
slidingWindow (dict) --
Configuration for sliding window truncation.
messagesCount (integer) --
The number of recent messages to retain in the context window.
summarization (dict) --
Configuration for summarization-based truncation.
summaryRatio (float) --
The ratio of content to summarize.
preserveRecentMessages (integer) --
The number of recent messages to preserve without summarization.
summarizationSystemPrompt (string) --
The system prompt used for generating summaries.
environment (dict) --
The compute environment on which the Harness runs.
agentCoreRuntimeEnvironment (dict) --
The AgentCore Runtime environment configuration.
agentRuntimeArn (string) --
The ARN of the underlying AgentCore Runtime.
agentRuntimeName (string) --
The name of the underlying AgentCore Runtime.
agentRuntimeId (string) --
The ID of the underlying AgentCore Runtime.
lifecycleConfiguration (dict) --
LifecycleConfiguration lets you manage the lifecycle of runtime sessions and resources in AgentCore Runtime. This configuration helps optimize resource utilization by automatically cleaning up idle sessions and preventing long-running instances from consuming resources indefinitely.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
networkConfiguration (dict) --
SecurityConfig for the Agent.
networkMode (string) --
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) --
The security groups associated with the VPC configuration.
(string) --
subnets (list) --
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
filesystemConfigurations (list) --
The filesystem configurations for the runtime environment.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) --
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
environmentArtifact (dict) --
The environment artifact (e.g., container) in which the Harness operates.
containerConfiguration (dict) --
Representation of a container configuration.
containerUri (string) --
The ECR URI of the container.
environmentVariables (dict) --
Environment variables exposed in the environment in which the Harness operates.
(string) --
(string) --
authorizerConfiguration (dict) --
Represents inbound authorization configuration options used to authenticate incoming requests.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) --
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) --
The name of the custom claim field to check.
inboundTokenClaimValueType (string) --
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) --
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) --
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) --
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) --
The domain to override with a private endpoint.
privateEndpoint (dict) --
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
memory (dict) --
AgentCore Memory instance configuration for short and long term memory.
agentCoreMemoryConfiguration (dict) --
The AgentCore Memory configuration.
arn (string) --
The ARN of the AgentCore Memory resource.
actorId (string) --
The actor ID for memory operations.
messagesCount (integer) --
The number of messages to retrieve from memory.
retrievalConfig (dict) --
The retrieval configuration for long-term memory, mapping namespace path templates to retrieval settings.
(string) --
(dict) --
Configuration for memory retrieval within a namespace.
topK (integer) --
The maximum number of memory entries to retrieve.
relevanceScore (float) --
The minimum relevance score for retrieved memories.
strategyId (string) --
The ID of the retrieval strategy to use.
maxIterations (integer) --
The maximum number of iterations in the agent loop allowed before exiting per invocation.
maxTokens (integer) --
The maximum total number of output tokens the agent can generate across all model calls within a single invocation.
timeoutSeconds (integer) --
The maximum duration per invocation.
failureReason (string) --
Reason why create or update operations fail.
{'authorizerType': 'CUSTOM_JWT | AWS_IAM'}
Lists all registries in the account. You can optionally filter results by status using the status parameter, or by authorizer type using the authorizerType parameter.
See also: AWS API Documentation
Request Syntax
client.list_registries(
maxResults=123,
nextToken='string',
status='CREATING'|'READY'|'UPDATING'|'CREATE_FAILED'|'UPDATE_FAILED'|'DELETING'|'DELETE_FAILED',
authorizerType='CUSTOM_JWT'|'AWS_IAM'
)
integer
The maximum number of results to return in the response. If the total number of results is greater than this value, use the token returned in the response in the nextToken field when making another request to return the next batch of results.
string
If the total number of results is greater than the maxResults value provided in the request, enter the token returned in the nextToken field in the response in this field to return the next batch of results.
string
Filter registries by their current status. Possible values include CREATING, READY, UPDATING, CREATE_FAILED, UPDATE_FAILED, DELETING, and DELETE_FAILED.
string
Filter registries by their authorizer type. Possible values are CUSTOM_JWT and AWS_IAM. For more information about authorizer types, see the RegistryAuthorizerType enum.
dict
Response Syntax
{
'registries': [
{
'name': 'string',
'description': 'string',
'registryId': 'string',
'registryArn': 'string',
'authorizerType': 'CUSTOM_JWT'|'AWS_IAM',
'status': 'CREATING'|'READY'|'UPDATING'|'CREATE_FAILED'|'UPDATE_FAILED'|'DELETING'|'DELETE_FAILED',
'statusReason': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1)
},
],
'nextToken': 'string'
}
Response Structure
(dict) --
registries (list) --
The list of registry summaries. For details about the fields in each summary, see the RegistrySummary data type.
(dict) --
Contains summary information about a registry.
name (string) --
The name of the registry.
description (string) --
The description of the registry.
registryId (string) --
The unique identifier of the registry.
registryArn (string) --
The Amazon Resource Name (ARN) of the registry.
authorizerType (string) --
The type of authorizer used by the registry. This controls the authorization method for the Search and Invoke APIs used by consumers.
CUSTOM_JWT - Authorize with a bearer token.
AWS_IAM - Authorize with your Amazon Web Services IAM credentials.
status (string) --
The current status of the registry. Possible values include CREATING, READY, UPDATING, CREATE_FAILED, UPDATE_FAILED, DELETING, and DELETE_FAILED.
statusReason (string) --
The reason for the current status, typically set when the status is a failure state.
createdAt (datetime) --
The timestamp when the registry was created.
updatedAt (datetime) --
The timestamp when the registry was last updated.
nextToken (string) --
If the total number of results is greater than the maxResults value provided in the request, use this token when making another request in the nextToken field to return the next batch of results.
{'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}
Updates an existing Amazon Secure Agent.
See also: AWS API Documentation
Request Syntax
client.update_agent_runtime(
agentRuntimeId='string',
agentRuntimeArtifact={
'containerConfiguration': {
'containerUri': 'string'
},
'codeConfiguration': {
'code': {
's3': {
'bucket': 'string',
'prefix': 'string',
'versionId': 'string'
}
},
'runtime': 'PYTHON_3_10'|'PYTHON_3_11'|'PYTHON_3_12'|'PYTHON_3_13'|'PYTHON_3_14'|'NODE_22',
'entryPoint': [
'string',
]
}
},
roleArn='string',
networkConfiguration={
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
description='string',
authorizerConfiguration={
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
requestHeaderConfiguration={
'requestHeaderAllowlist': [
'string',
]
},
protocolConfiguration={
'serverProtocol': 'MCP'|'HTTP'|'A2A'|'AGUI'
},
lifecycleConfiguration={
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
metadataConfiguration={
'requireMMDSV2': True|False
},
environmentVariables={
'string': 'string'
},
filesystemConfigurations=[
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
],
clientToken='string'
)
string
[REQUIRED]
The unique identifier of the AgentCore Runtime to update.
dict
[REQUIRED]
The updated artifact of the AgentCore Runtime.
containerConfiguration (dict) --
The container configuration for the agent artifact.
containerUri (string) -- [REQUIRED]
The ECR URI of the container.
codeConfiguration (dict) --
The code configuration for the agent runtime artifact, including the source code location and execution settings.
code (dict) -- [REQUIRED]
The source code location and configuration details.
s3 (dict) --
The Amazon Amazon S3 object that contains the source code for the agent runtime.
bucket (string) -- [REQUIRED]
The name of the Amazon S3 bucket. This bucket contains the stored data.
prefix (string) -- [REQUIRED]
The prefix for objects in the Amazon S3 bucket. This prefix is added to the object keys to organize the data.
versionId (string) --
The version ID of the Amazon Amazon S3 object. If not specified, the latest version of the object is used.
runtime (string) -- [REQUIRED]
The runtime environment for executing the agent code. Specify the programming language and version to use for the agent runtime. For valid values, see the list of supported runtimes.
entryPoint (list) -- [REQUIRED]
The entry point for the code execution, specifying the function or method that should be invoked when the code runs.
(string) --
string
[REQUIRED]
The updated IAM role ARN that provides permissions for the AgentCore Runtime.
dict
[REQUIRED]
The updated network configuration for the AgentCore Runtime.
networkMode (string) -- [REQUIRED]
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) -- [REQUIRED]
The security groups associated with the VPC configuration.
(string) --
subnets (list) -- [REQUIRED]
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
string
The updated description of the AgentCore Runtime.
dict
The updated authorizer configuration for the AgentCore Runtime.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) -- [REQUIRED]
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) -- [REQUIRED]
The name of the custom claim field to check.
inboundTokenClaimValueType (string) -- [REQUIRED]
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) -- [REQUIRED]
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) -- [REQUIRED]
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) -- [REQUIRED]
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) -- [REQUIRED]
The domain to override with a private endpoint.
privateEndpoint (dict) -- [REQUIRED]
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
dict
The updated configuration for HTTP request headers that will be passed through to the runtime.
requestHeaderAllowlist (list) --
A list of HTTP request headers that are allowed to be passed through to the runtime.
(string) --
dict
The protocol configuration for an agent runtime. This structure defines how the agent runtime communicates with clients.
serverProtocol (string) -- [REQUIRED]
The server protocol for the agent runtime. This field specifies which protocol the agent runtime uses to communicate with clients.
dict
The updated life cycle configuration for the AgentCore Runtime.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
dict
The updated configuration for microVM Metadata Service (MMDS) settings for the AgentCore Runtime.
requireMMDSV2 (boolean) -- [REQUIRED]
Enables MMDSv2 (microVM Metadata Service Version 2) requirement for the agent runtime. When set to true, the runtime microVM will only accept MMDSv2 requests.
dict
Updated environment variables to set in the AgentCore Runtime environment.
(string) --
(string) --
list
The updated filesystem configurations to mount into the AgentCore Runtime.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) -- [REQUIRED]
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
string
A unique, case-sensitive identifier to ensure idempotency of the request.
This field is autopopulated if not provided.
dict
Response Syntax
{
'agentRuntimeArn': 'string',
'agentRuntimeId': 'string',
'workloadIdentityDetails': {
'workloadIdentityArn': 'string'
},
'agentRuntimeVersion': 'string',
'createdAt': datetime(2015, 1, 1),
'lastUpdatedAt': datetime(2015, 1, 1),
'status': 'CREATING'|'CREATE_FAILED'|'UPDATING'|'UPDATE_FAILED'|'READY'|'DELETING'
}
Response Structure
(dict) --
agentRuntimeArn (string) --
The Amazon Resource Name (ARN) of the updated AgentCore Runtime.
agentRuntimeId (string) --
The unique identifier of the updated AgentCore Runtime.
workloadIdentityDetails (dict) --
The workload identity details for the updated AgentCore Runtime.
workloadIdentityArn (string) --
The ARN associated with the workload identity.
agentRuntimeVersion (string) --
The version of the updated AgentCore Runtime.
createdAt (datetime) --
The timestamp when the AgentCore Runtime was created.
lastUpdatedAt (datetime) --
The timestamp when the AgentCore Runtime was last updated.
status (string) --
The current status of the updated AgentCore Runtime.
{'environment': {'agentCoreRuntimeEnvironment': {'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}}}
Response {'harness': {'environment': {'agentCoreRuntimeEnvironment': {'networkConfiguration': {'networkModeConfig': {'requireServiceS3Endpoint': 'boolean'}}}}}}
Operation to update a Harness.
See also: AWS API Documentation
Request Syntax
client.update_harness(
harnessId='string',
clientToken='string',
executionRoleArn='string',
environment={
'agentCoreRuntimeEnvironment': {
'lifecycleConfiguration': {
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'filesystemConfigurations': [
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
]
}
},
environmentArtifact={
'optionalValue': {
'containerConfiguration': {
'containerUri': 'string'
}
}
},
environmentVariables={
'string': 'string'
},
authorizerConfiguration={
'optionalValue': {
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
}
},
model={
'bedrockModelConfig': {
'modelId': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'openAiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'geminiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...,
'topK': 123
}
},
systemPrompt=[
{
'text': 'string'
},
],
tools=[
{
'type': 'remote_mcp'|'agentcore_browser'|'agentcore_gateway'|'inline_function'|'agentcore_code_interpreter',
'name': 'string',
'config': {
'remoteMcp': {
'url': 'string',
'headers': {
'string': 'string'
}
},
'agentCoreBrowser': {
'browserArn': 'string'
},
'agentCoreGateway': {
'gatewayArn': 'string',
'outboundAuth': {
'awsIam': {}
,
'none': {}
,
'oauth': {
'providerArn': 'string',
'scopes': [
'string',
],
'customParameters': {
'string': 'string'
},
'grantType': 'CLIENT_CREDENTIALS'|'AUTHORIZATION_CODE'|'TOKEN_EXCHANGE',
'defaultReturnUrl': 'string'
}
}
},
'inlineFunction': {
'description': 'string',
'inputSchema': {...}|[...]|123|123.4|'string'|True|None
},
'agentCoreCodeInterpreter': {
'codeInterpreterArn': 'string'
}
}
},
],
skills=[
{
'path': 'string'
},
],
allowedTools=[
'string',
],
memory={
'optionalValue': {
'agentCoreMemoryConfiguration': {
'arn': 'string',
'actorId': 'string',
'messagesCount': 123,
'retrievalConfig': {
'string': {
'topK': 123,
'relevanceScore': ...,
'strategyId': 'string'
}
}
}
}
},
truncation={
'strategy': 'sliding_window'|'summarization'|'none',
'config': {
'slidingWindow': {
'messagesCount': 123
},
'summarization': {
'summaryRatio': ...,
'preserveRecentMessages': 123,
'summarizationSystemPrompt': 'string'
}
}
},
maxIterations=123,
maxTokens=123,
timeoutSeconds=123
)
string
[REQUIRED]
The ID of the harness to update.
string
A unique, case-sensitive identifier to ensure idempotency of the request.
This field is autopopulated if not provided.
string
The ARN of the IAM role that the harness assumes when running. If not specified, the existing value is retained.
dict
The compute environment configuration for the harness. If not specified, the existing value is retained.
agentCoreRuntimeEnvironment (dict) --
The AgentCore Runtime environment configuration.
lifecycleConfiguration (dict) --
LifecycleConfiguration lets you manage the lifecycle of runtime sessions and resources in AgentCore Runtime. This configuration helps optimize resource utilization by automatically cleaning up idle sessions and preventing long-running instances from consuming resources indefinitely.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
networkConfiguration (dict) --
SecurityConfig for the Agent.
networkMode (string) -- [REQUIRED]
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) -- [REQUIRED]
The security groups associated with the VPC configuration.
(string) --
subnets (list) -- [REQUIRED]
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
filesystemConfigurations (list) --
The filesystem configurations for the runtime environment.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) -- [REQUIRED]
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) -- [REQUIRED]
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) -- [REQUIRED]
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
dict
The environment artifact for the harness. Use the optionalValue wrapper to set a new value, or set it to null to clear the existing configuration.
optionalValue (dict) --
The updated environment artifact value, or null to clear the existing configuration.
containerConfiguration (dict) --
Representation of a container configuration.
containerUri (string) -- [REQUIRED]
The ECR URI of the container.
dict
Environment variables to set in the harness runtime environment. If specified, this replaces all existing environment variables. If not specified, the existing value is retained.
(string) --
(string) --
dict
Wrapper for updating an optional AuthorizerConfiguration field with PATCH semantics. When present in an update request, the authorizer configuration is replaced with optionalValue. When absent, the authorizer configuration is left unchanged. To unset, include the wrapper with optionalValue not specified.
optionalValue (dict) --
The updated authorizer configuration value. If not specified, it will clear the current authorizer configuration of the resource.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) -- [REQUIRED]
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) -- [REQUIRED]
The name of the custom claim field to check.
inboundTokenClaimValueType (string) -- [REQUIRED]
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) -- [REQUIRED]
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) -- [REQUIRED]
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) -- [REQUIRED]
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) -- [REQUIRED]
The domain to override with a private endpoint.
privateEndpoint (dict) -- [REQUIRED]
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) -- [REQUIRED]
The ID of the VPC that contains your private resource.
subnetIds (list) -- [REQUIRED]
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) -- [REQUIRED]
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
dict
The model configuration for the harness. If not specified, the existing value is retained.
bedrockModelConfig (dict) --
Configuration for an Amazon Bedrock model.
modelId (string) -- [REQUIRED]
The Bedrock model ID.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
openAiModelConfig (dict) --
Configuration for an OpenAI model.
modelId (string) -- [REQUIRED]
The OpenAI model ID.
apiKeyArn (string) -- [REQUIRED]
The ARN of your OpenAI API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
geminiModelConfig (dict) --
Configuration for a Google Gemini model.
modelId (string) -- [REQUIRED]
The Gemini model ID.
apiKeyArn (string) -- [REQUIRED]
The ARN of your Gemini API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
topK (integer) --
The topK set when calling the model.
list
The system prompt that defines the agent's behavior. If not specified, the existing value is retained.
(dict) --
A content block in the system prompt.
text (string) --
The text content of the system prompt block.
list
The tools available to the agent. If specified, this replaces all existing tools. If not specified, the existing value is retained.
(dict) --
A tool available to the agent loop.
type (string) -- [REQUIRED]
The type of tool.
name (string) --
Unique name for the tool. If not provided, a name will be inferred or generated.
config (dict) --
Tool-specific configuration.
remoteMcp (dict) --
Configuration for remote MCP server.
url (string) -- [REQUIRED]
URL of the MCP endpoint.
headers (dict) --
Custom headers to include when connecting to the remote MCP server.
(string) --
The key of an HTTP header.
(string) --
The value of an HTTP header.
agentCoreBrowser (dict) --
Configuration for AgentCore Browser.
browserArn (string) --
If not populated, the built-in Browser ARN is used.
agentCoreGateway (dict) --
Configuration for AgentCore Gateway.
gatewayArn (string) -- [REQUIRED]
The ARN of the desired AgentCore Gateway.
outboundAuth (dict) --
How harness authenticates to this Gateway. Defaults to AWS_IAM (SigV4) if omitted.
awsIam (dict) --
SigV4-sign requests using the agent's execution role.
none (dict) --
No authentication.
oauth (dict) --
An OAuth credential provider for gateway authentication. This structure contains the configuration for authenticating with the target endpoint using OAuth.
providerArn (string) -- [REQUIRED]
The Amazon Resource Name (ARN) of the OAuth credential provider. This ARN identifies the provider in Amazon Web Services.
scopes (list) -- [REQUIRED]
The OAuth scopes for the credential provider. These scopes define the level of access requested from the OAuth provider.
(string) --
customParameters (dict) --
The custom parameters for the OAuth credential provider. These parameters provide additional configuration for the OAuth authentication process.
(string) --
(string) --
grantType (string) --
Specifies the kind of credentials to use for authorization:
CLIENT_CREDENTIALS - Authorization with a client ID and secret.
AUTHORIZATION_CODE - Authorization with a token that is specific to an individual end user.
TOKEN_EXCHANGE - Authorization using on-behalf-of token exchange. An inbound user token is exchanged for a downstream access token scoped to the target audience.
defaultReturnUrl (string) --
The URL where the end user's browser is redirected after obtaining the authorization code. Generally points to the customer's application.
inlineFunction (dict) --
Configuration for an inline function tool.
description (string) -- [REQUIRED]
Description of what the tool does, provided to the model.
inputSchema (:ref:`document<document>`) -- [REQUIRED]
JSON Schema describing the tool's input parameters.
agentCoreCodeInterpreter (dict) --
Configuration for AgentCore Code Interpreter.
codeInterpreterArn (string) --
If not populated, the built-in Code Interpreter ARN is used.
list
The skills available to the agent. If specified, this replaces all existing skills. If not specified, the existing value is retained.
(dict) --
A skill available to the agent.
path (string) --
The filesystem path to the skill definition.
list
The tools that the agent is allowed to use. If specified, this replaces all existing allowed tools. If not specified, the existing value is retained.
(string) --
dict
The AgentCore Memory configuration. Use the optionalValue wrapper to set a new value, or set it to null to clear the existing configuration.
optionalValue (dict) --
The updated memory configuration value, or null to clear the existing configuration.
agentCoreMemoryConfiguration (dict) --
The AgentCore Memory configuration.
arn (string) -- [REQUIRED]
The ARN of the AgentCore Memory resource.
actorId (string) --
The actor ID for memory operations.
messagesCount (integer) --
The number of messages to retrieve from memory.
retrievalConfig (dict) --
The retrieval configuration for long-term memory, mapping namespace path templates to retrieval settings.
(string) --
(dict) --
Configuration for memory retrieval within a namespace.
topK (integer) --
The maximum number of memory entries to retrieve.
relevanceScore (float) --
The minimum relevance score for retrieved memories.
strategyId (string) --
The ID of the retrieval strategy to use.
dict
The truncation configuration for managing conversation context. If not specified, the existing value is retained.
strategy (string) -- [REQUIRED]
The truncation strategy to use.
config (dict) --
The strategy-specific configuration.
slidingWindow (dict) --
Configuration for sliding window truncation.
messagesCount (integer) --
The number of recent messages to retain in the context window.
summarization (dict) --
Configuration for summarization-based truncation.
summaryRatio (float) --
The ratio of content to summarize.
preserveRecentMessages (integer) --
The number of recent messages to preserve without summarization.
summarizationSystemPrompt (string) --
The system prompt used for generating summaries.
integer
The maximum number of iterations the agent loop can execute per invocation. If not specified, the existing value is retained.
integer
The maximum total number of output tokens the agent can generate across all model calls within a single invocation. If not specified, the existing value is retained.
integer
The maximum duration in seconds for the agent loop execution per invocation. If not specified, the existing value is retained.
dict
Response Syntax
{
'harness': {
'harnessId': 'string',
'harnessName': 'string',
'arn': 'string',
'status': 'CREATING'|'CREATE_FAILED'|'UPDATING'|'UPDATE_FAILED'|'READY'|'DELETING'|'DELETE_FAILED',
'executionRoleArn': 'string',
'createdAt': datetime(2015, 1, 1),
'updatedAt': datetime(2015, 1, 1),
'model': {
'bedrockModelConfig': {
'modelId': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'openAiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...
},
'geminiModelConfig': {
'modelId': 'string',
'apiKeyArn': 'string',
'maxTokens': 123,
'temperature': ...,
'topP': ...,
'topK': 123
}
},
'systemPrompt': [
{
'text': 'string'
},
],
'tools': [
{
'type': 'remote_mcp'|'agentcore_browser'|'agentcore_gateway'|'inline_function'|'agentcore_code_interpreter',
'name': 'string',
'config': {
'remoteMcp': {
'url': 'string',
'headers': {
'string': 'string'
}
},
'agentCoreBrowser': {
'browserArn': 'string'
},
'agentCoreGateway': {
'gatewayArn': 'string',
'outboundAuth': {
'awsIam': {},
'none': {},
'oauth': {
'providerArn': 'string',
'scopes': [
'string',
],
'customParameters': {
'string': 'string'
},
'grantType': 'CLIENT_CREDENTIALS'|'AUTHORIZATION_CODE'|'TOKEN_EXCHANGE',
'defaultReturnUrl': 'string'
}
}
},
'inlineFunction': {
'description': 'string',
'inputSchema': {...}|[...]|123|123.4|'string'|True|None
},
'agentCoreCodeInterpreter': {
'codeInterpreterArn': 'string'
}
}
},
],
'skills': [
{
'path': 'string'
},
],
'allowedTools': [
'string',
],
'truncation': {
'strategy': 'sliding_window'|'summarization'|'none',
'config': {
'slidingWindow': {
'messagesCount': 123
},
'summarization': {
'summaryRatio': ...,
'preserveRecentMessages': 123,
'summarizationSystemPrompt': 'string'
}
}
},
'environment': {
'agentCoreRuntimeEnvironment': {
'agentRuntimeArn': 'string',
'agentRuntimeName': 'string',
'agentRuntimeId': 'string',
'lifecycleConfiguration': {
'idleRuntimeSessionTimeout': 123,
'maxLifetime': 123
},
'networkConfiguration': {
'networkMode': 'PUBLIC'|'VPC',
'networkModeConfig': {
'securityGroups': [
'string',
],
'subnets': [
'string',
],
'requireServiceS3Endpoint': True|False
}
},
'filesystemConfigurations': [
{
'sessionStorage': {
'mountPath': 'string'
},
's3FilesAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
},
'efsAccessPoint': {
'accessPointArn': 'string',
'mountPath': 'string'
}
},
]
}
},
'environmentArtifact': {
'containerConfiguration': {
'containerUri': 'string'
}
},
'environmentVariables': {
'string': 'string'
},
'authorizerConfiguration': {
'customJWTAuthorizer': {
'discoveryUrl': 'string',
'allowedAudience': [
'string',
],
'allowedClients': [
'string',
],
'allowedScopes': [
'string',
],
'customClaims': [
{
'inboundTokenClaimName': 'string',
'inboundTokenClaimValueType': 'STRING'|'STRING_ARRAY',
'authorizingClaimMatchValue': {
'claimMatchValue': {
'matchValueString': 'string',
'matchValueStringList': [
'string',
]
},
'claimMatchOperator': 'EQUALS'|'CONTAINS'|'CONTAINS_ANY'
}
},
],
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
},
'privateEndpointOverrides': [
{
'domain': 'string',
'privateEndpoint': {
'selfManagedLatticeResource': {
'resourceConfigurationIdentifier': 'string'
},
'managedVpcResource': {
'vpcIdentifier': 'string',
'subnetIds': [
'string',
],
'endpointIpAddressType': 'IPV4'|'IPV6',
'securityGroupIds': [
'string',
],
'tags': {
'string': 'string'
},
'routingDomain': 'string'
}
}
},
]
}
},
'memory': {
'agentCoreMemoryConfiguration': {
'arn': 'string',
'actorId': 'string',
'messagesCount': 123,
'retrievalConfig': {
'string': {
'topK': 123,
'relevanceScore': ...,
'strategyId': 'string'
}
}
}
},
'maxIterations': 123,
'maxTokens': 123,
'timeoutSeconds': 123,
'failureReason': 'string'
}
}
Response Structure
(dict) --
harness (dict) --
The updated harness.
harnessId (string) --
The ID of the Harness.
harnessName (string) --
The name of the Harness.
arn (string) --
The ARN of the Harness.
status (string) --
The status of the Harness.
executionRoleArn (string) --
IAM role the Harness assumes when running.
createdAt (datetime) --
The createdAt time of the Harness.
updatedAt (datetime) --
The updatedAt time of the Harness.
model (dict) --
The configuration of the default model used by the Harness.
bedrockModelConfig (dict) --
Configuration for an Amazon Bedrock model.
modelId (string) --
The Bedrock model ID.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
openAiModelConfig (dict) --
Configuration for an OpenAI model.
modelId (string) --
The OpenAI model ID.
apiKeyArn (string) --
The ARN of your OpenAI API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
geminiModelConfig (dict) --
Configuration for a Google Gemini model.
modelId (string) --
The Gemini model ID.
apiKeyArn (string) --
The ARN of your Gemini API key on AgentCore Identity.
maxTokens (integer) --
The maximum number of tokens to allow in the generated response per model call.
temperature (float) --
The temperature to set when calling the model.
topP (float) --
The topP set when calling the model.
topK (integer) --
The topK set when calling the model.
systemPrompt (list) --
The system prompt of the Harness.
(dict) --
A content block in the system prompt.
text (string) --
The text content of the system prompt block.
tools (list) --
The tools of the Harness.
(dict) --
A tool available to the agent loop.
type (string) --
The type of tool.
name (string) --
Unique name for the tool. If not provided, a name will be inferred or generated.
config (dict) --
Tool-specific configuration.
remoteMcp (dict) --
Configuration for remote MCP server.
url (string) --
URL of the MCP endpoint.
headers (dict) --
Custom headers to include when connecting to the remote MCP server.
(string) --
The key of an HTTP header.
(string) --
The value of an HTTP header.
agentCoreBrowser (dict) --
Configuration for AgentCore Browser.
browserArn (string) --
If not populated, the built-in Browser ARN is used.
agentCoreGateway (dict) --
Configuration for AgentCore Gateway.
gatewayArn (string) --
The ARN of the desired AgentCore Gateway.
outboundAuth (dict) --
How harness authenticates to this Gateway. Defaults to AWS_IAM (SigV4) if omitted.
awsIam (dict) --
SigV4-sign requests using the agent's execution role.
none (dict) --
No authentication.
oauth (dict) --
An OAuth credential provider for gateway authentication. This structure contains the configuration for authenticating with the target endpoint using OAuth.
providerArn (string) --
The Amazon Resource Name (ARN) of the OAuth credential provider. This ARN identifies the provider in Amazon Web Services.
scopes (list) --
The OAuth scopes for the credential provider. These scopes define the level of access requested from the OAuth provider.
(string) --
customParameters (dict) --
The custom parameters for the OAuth credential provider. These parameters provide additional configuration for the OAuth authentication process.
(string) --
(string) --
grantType (string) --
Specifies the kind of credentials to use for authorization:
CLIENT_CREDENTIALS - Authorization with a client ID and secret.
AUTHORIZATION_CODE - Authorization with a token that is specific to an individual end user.
TOKEN_EXCHANGE - Authorization using on-behalf-of token exchange. An inbound user token is exchanged for a downstream access token scoped to the target audience.
defaultReturnUrl (string) --
The URL where the end user's browser is redirected after obtaining the authorization code. Generally points to the customer's application.
inlineFunction (dict) --
Configuration for an inline function tool.
description (string) --
Description of what the tool does, provided to the model.
inputSchema (:ref:`document<document>`) --
JSON Schema describing the tool's input parameters.
agentCoreCodeInterpreter (dict) --
Configuration for AgentCore Code Interpreter.
codeInterpreterArn (string) --
If not populated, the built-in Code Interpreter ARN is used.
skills (list) --
The skills of the Harness.
(dict) --
A skill available to the agent.
path (string) --
The filesystem path to the skill definition.
allowedTools (list) --
The allowed tools of the Harness. All tools are allowed by default.
(string) --
truncation (dict) --
Configuration for truncating model context.
strategy (string) --
The truncation strategy to use.
config (dict) --
The strategy-specific configuration.
slidingWindow (dict) --
Configuration for sliding window truncation.
messagesCount (integer) --
The number of recent messages to retain in the context window.
summarization (dict) --
Configuration for summarization-based truncation.
summaryRatio (float) --
The ratio of content to summarize.
preserveRecentMessages (integer) --
The number of recent messages to preserve without summarization.
summarizationSystemPrompt (string) --
The system prompt used for generating summaries.
environment (dict) --
The compute environment on which the Harness runs.
agentCoreRuntimeEnvironment (dict) --
The AgentCore Runtime environment configuration.
agentRuntimeArn (string) --
The ARN of the underlying AgentCore Runtime.
agentRuntimeName (string) --
The name of the underlying AgentCore Runtime.
agentRuntimeId (string) --
The ID of the underlying AgentCore Runtime.
lifecycleConfiguration (dict) --
LifecycleConfiguration lets you manage the lifecycle of runtime sessions and resources in AgentCore Runtime. This configuration helps optimize resource utilization by automatically cleaning up idle sessions and preventing long-running instances from consuming resources indefinitely.
idleRuntimeSessionTimeout (integer) --
Timeout in seconds for idle runtime sessions. When a session remains idle for this duration, it will be automatically terminated. Default: 900 seconds (15 minutes).
maxLifetime (integer) --
Maximum lifetime for the instance in seconds. Once reached, instances will be automatically terminated and replaced. Default: 28800 seconds (8 hours).
networkConfiguration (dict) --
SecurityConfig for the Agent.
networkMode (string) --
The network mode for the AgentCore Runtime.
networkModeConfig (dict) --
The network mode configuration for the AgentCore Runtime.
securityGroups (list) --
The security groups associated with the VPC configuration.
(string) --
subnets (list) --
The subnets associated with the VPC configuration.
(string) --
requireServiceS3Endpoint (boolean) --
Controls whether a service-managed Amazon S3 gateway endpoint is provisioned in the VPC network topology for the agent runtime. This gateway is used by Amazon Bedrock AgentCore Runtime to download code and container images during agent startup.
Starting May 5, 2026, Amazon Bedrock AgentCore Runtime is gradually rolling out a change to how network isolation is configured for VPC mode agents. Agent runtimes created on or after this rollout will no longer include the service-managed Amazon S3 gateway. Instead, all network access, including to Amazon S3, is governed exclusively by your VPC configuration. This field cannot be set on agent runtimes created after the rollout. Passing this field in an UpdateAgentRuntime request for these agent runtimes returns a ValidationException.
Agent runtimes created before the rollout are not affected and continue to operate with the service-managed Amazon S3 gateway. To enforce full VPC network isolation on these existing agent runtimes, set this field to false via the UpdateAgentRuntime API. Before opting out, ensure your VPC provides the Amazon S3 access required for agent startup. If this field is not specified or is set to true, the service-managed Amazon S3 gateway remains provisioned.
This field is only supported in the UpdateAgentRuntime API for pre-rollout agent runtimes. Passing this field in a CreateAgentRuntime request returns a ValidationException.
filesystemConfigurations (list) --
The filesystem configurations for the runtime environment.
(dict) --
Configuration for a filesystem that can be mounted into the AgentCore Runtime.
sessionStorage (dict) --
Configuration for session storage. Session storage provides persistent storage that is preserved across AgentCore Runtime session invocations.
mountPath (string) --
The mount path for the session storage filesystem inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
s3FilesAccessPoint (dict) --
Configuration for an Amazon S3 Files access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the S3 Files access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the S3 Files access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
efsAccessPoint (dict) --
Configuration for an Amazon EFS access point to mount into the AgentCore Runtime.
accessPointArn (string) --
The ARN of the EFS access point to mount into the AgentCore Runtime.
mountPath (string) --
The mount path for the EFS access point inside the AgentCore Runtime. The path must be under /mnt with exactly one subdirectory level (for example, /mnt/data).
environmentArtifact (dict) --
The environment artifact (e.g., container) in which the Harness operates.
containerConfiguration (dict) --
Representation of a container configuration.
containerUri (string) --
The ECR URI of the container.
environmentVariables (dict) --
Environment variables exposed in the environment in which the Harness operates.
(string) --
(string) --
authorizerConfiguration (dict) --
Represents inbound authorization configuration options used to authenticate incoming requests.
customJWTAuthorizer (dict) --
The inbound JWT-based authorization, specifying how incoming requests should be authenticated.
discoveryUrl (string) --
This URL is used to fetch OpenID Connect configuration or authorization server metadata for validating incoming tokens.
allowedAudience (list) --
Represents individual audience values that are validated in the incoming JWT token validation process.
(string) --
allowedClients (list) --
Represents individual client IDs that are validated in the incoming JWT token validation process.
(string) --
allowedScopes (list) --
An array of scopes that are allowed to access the token.
(string) --
customClaims (list) --
An array of objects that define a custom claim validation name, value, and operation
(dict) --
Defines the name of a custom claim field and rules for finding matches to authenticate its value.
inboundTokenClaimName (string) --
The name of the custom claim field to check.
inboundTokenClaimValueType (string) --
The data type of the claim value to check for.
Use STRING if you want to find an exact match to a string you define.
Use STRING_ARRAY if you want to fnd a match to at least one value in an array you define.
authorizingClaimMatchValue (dict) --
Defines the value or values to match for and the relationship of the match.
claimMatchValue (dict) --
The value or values to match for.
matchValueString (string) --
The string value to match for.
matchValueStringList (list) --
An array of strings to check for a match.
(string) --
claimMatchOperator (string) --
Defines the relationship between the claim field value and the value or values you're matching for.
privateEndpoint (dict) --
The private endpoint configuration for a gateway target. Defines how the gateway connects to private resources in your VPC.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
privateEndpointOverrides (list) --
The private endpoint overrides for the custom JWT authorizer configuration.
(dict) --
A mapping of a specific domain to a private endpoint for secure connectivity through a VPC Lattice resource configuration.
domain (string) --
The domain to override with a private endpoint.
privateEndpoint (dict) --
The private endpoint configuration for the specified domain.
selfManagedLatticeResource (dict) --
Configuration for connecting to a private resource using a self-managed VPC Lattice resource configuration.
resourceConfigurationIdentifier (string) --
The ARN or ID of the VPC Lattice resource configuration.
managedVpcResource (dict) --
Configuration for connecting to a private resource using a managed VPC Lattice resource. The gateway creates and manages the VPC Lattice resources on your behalf.
vpcIdentifier (string) --
The ID of the VPC that contains your private resource.
subnetIds (list) --
The subnet IDs within the VPC where the VPC Lattice resource gateway is placed.
(string) --
endpointIpAddressType (string) --
The IP address type for the resource configuration endpoint.
securityGroupIds (list) --
The security group IDs to associate with the VPC Lattice resource gateway. If not specified, the default security group for the VPC is used.
(string) --
tags (dict) --
Tags to apply to the managed VPC Lattice resource gateway.
(string) --
(string) --
routingDomain (string) --
An intermediate domain to use as the resource configuration endpoint instead of the actual target domain. Use this when you want to route traffic through an intermediate component such as a VPC endpoint or internal load balancer. For more information, see xref:lattice-vpc-egress-routing-domain[Route traffic through an intermediate domain].
memory (dict) --
AgentCore Memory instance configuration for short and long term memory.
agentCoreMemoryConfiguration (dict) --
The AgentCore Memory configuration.
arn (string) --
The ARN of the AgentCore Memory resource.
actorId (string) --
The actor ID for memory operations.
messagesCount (integer) --
The number of messages to retrieve from memory.
retrievalConfig (dict) --
The retrieval configuration for long-term memory, mapping namespace path templates to retrieval settings.
(string) --
(dict) --
Configuration for memory retrieval within a namespace.
topK (integer) --
The maximum number of memory entries to retrieve.
relevanceScore (float) --
The minimum relevance score for retrieved memories.
strategyId (string) --
The ID of the retrieval strategy to use.
maxIterations (integer) --
The maximum number of iterations in the agent loop allowed before exiting per invocation.
maxTokens (integer) --
The maximum total number of output tokens the agent can generate across all model calls within a single invocation.
timeoutSeconds (integer) --
The maximum duration per invocation.
failureReason (string) --
Reason why create or update operations fail.