Kiwi

Kiwi for AWS

Have the agent on a Kiwi step call AWS APIs under the IAM permissions it signs in with and read the AWS documentation, through the AWS MCP Server.

  • Your agent works in AWS through its own connection. Kiwi does not call AWS and never holds its credentials.
  • A workflow skill can list AWS among the tools it requires. When you supply your agent's available tools to a skill check and AWS is not among them, the check reports SK15-UNSUPPORTED-TOOL.
  • A project Admin can add links to AWS, with notes, under Connections and bind them to the steps that use them. The agent reads them with list_connections.
  • The agent can report the tools and plugins it ran with, including its AWS connection, with report_configuration.
  • The agent submits links to its work in AWS as the step's outputs. People review them, and sign-off stays with the people your workflow names.

What your agent can do in AWS

  • Call any of more than 15,000 AWS API operations, limited by the IAM permissions of the signed-in identity.
  • Search and read the AWS documentation.
  • Run a Python script in a sandbox without network access that uses the same IAM permissions.
  • Review the server's calls afterwards in AWS CloudTrail.

Set up

  1. Connect your agent to Kiwi

    Add Kiwi's hosted MCP server, https://kiwiflow.io/mcp, to your coding agent.

  2. Connect your agent to AWS

    Add the AWS MCP server to your agent with this URL. The credentials stay with your agent.

    https://aws-mcp.us-east-1.api.aws/mcp
  3. Example skill

    Publish instructions like these as a workflow skill that lists AWS among its required tools.

    Read the AWS account, region and stack name from the step's inputs.
    Sign in with a read-only IAM role and describe the stack's resources and
    its recent events in AWS.
    Record the resource ARNs and any failed events in the step's outputs.
    Changes to production resources need the run's Owner.
    If a call is denied, report a blocker rather than asking for wider access.