@cdklabs/generative-ai-cdk-constructs
@cdklabs/generative-ai-cdk-constructs / bedrock / AgentProps
Properties for creating a CDK managed Bedrock Agent.
readonly
optional
actionGroups:AgentActionGroup
[]
The Action Groups associated with the agent.
readonly
optional
codeInterpreterEnabled:boolean
Select whether the agent can generate, run, and troubleshoot code when trying to complete a task
- false
readonly
optional
description:string
A description of the agent.
- No description is provided.
readonly
optional
existingRole:IRole
The existing IAM Role for the agent to use. Ensure the role has a trust policy that allows the Bedrock service to assume the role.
- A new role is created for you.
readonly
optional
forceDelete:boolean
Whether to delete the resource even if it’s in use.
- true
readonly
foundationModel:IInvokable
The foundation model used for orchestration by the agent.
readonly
optional
guardrail:IGuardrail
The guardrail that will be associated with the agent.
readonly
optional
idleSessionTTL:Duration
How long sessions should be kept open for the agent. If no conversation occurs during this time, the session expires and Amazon Bedrock deletes any data provided before the timeout.
- 1 hour
readonly
instruction:string
The instruction used by the agent. This determines how the agent will perform his task. This instruction must have a minimum of 40 characters.
readonly
optional
kmsKey:IKey
The KMS key of the agent if custom encryption is configured.
- An AWS managed key is used.
readonly
optional
knowledgeBases:IKnowledgeBase
[]
The KnowledgeBases associated with the agent.
readonly
optional
name:string
The name of the agent.
- A name is generated by CDK.
readonly
optional
promptOverrideConfiguration:PromptOverrideConfiguration
Overrides some prompt templates in different parts of an agent sequence configuration.
- No overrides are provided.
readonly
optional
shouldPrepareAgent:boolean
Specifies whether to automatically update the DRAFT
version of the agent after
making changes to the agent. The DRAFT
version can be continually iterated
upon during internal development.
- false
readonly
optional
userInputEnabled:boolean
Select whether the agent can prompt additional information from the user when it does not have enough information to respond to an utterance
- false