Invokes an AWS Bedrock AgentCore runtime asynchronously from a DevRev workflow step, returning a session ID immediately so the workflow completes within its time budget while the agent runs.

Features

• Asynchronous fire-and-forget invocation The action dispatches the AgentCore request without waiting for the agent to finish, keeping the workflow step within its execution time budget and returning a unique session ID that can be used by downstream steps.
• Secure credential injection AWS credentials are supplied via a dedicated AWS IAM Role keyring, and the DevRev PAT is injected as a custom request header through middleware — ensuring sensitive tokens never appear in the request body or in CloudWatch logs.
• Optional dashboard context An optional Dashboard ID input lets the action append the target dashboard's identifier to the prompt automatically, enabling agent workflows that populate or modify an existing DevRev dashboard created by a prior workflow step.
• Configurable AWS region The AWS region used to reach the AgentCore endpoint is selectable at the operation level, making it straightforward to target runtimes deployed in different regions.

Installation

1. AWS IAM Role keyring — Add an AWS session-token credential (access key ID, secret access key, and session token) that has permission to invoke the target AgentCore runtime.
2. DevRev PAT keyring — Add a long-lived DevRev personal access token. A service account PAT is strongly recommended because tokens available inside a running workflow are short-lived (under 30 minutes) and may expire before the agent finishes.
3. Agent Runtime ARN — Provide the full ARN of the AgentCore runtime to invoke (e.g. arn:aws:bedrock-agentcore:us-east-1:ACCOUNT:runtime/NAME-ID).
4. Prompt — Enter the natural-language instruction the agent should act on (e.g. "build me a ticket volume dashboard").
5. Dashboard ID (optional) — If the agent should work against an existing DevRev dashboard, provide its DON ID here. It is automatically appended to the prompt.
6. AWS Region — Select the AWS region where the runtime is deployed.