Request a credit limit increase, governed by Aegis.

This form calls a real Strands agent deployed on Amazon Bedrock AgentCore Runtime, in the product owner's own AWS account. Every request is routed through Aegis's policy engine before anything executes, and every verdict is written to a signed, hash-chained Decision Record in a real DynamoDB ledger — independently verifiable below, not just reported back by the agent.

What is real and what is not: the AgentCore Runtime invocation, the policy decision and the Decision Record write are all real, live AWS operations against account 785959780490 (region eu-west-1). The underlying language-model call is currently mocked (the runtime's MIDASBRAIN_MOCK_LLM environment variable is unset, and the agent code defaults to mock mode when that variable is absent) — so the proposal shape the model would produce is generated deterministically from your input, not by a live Claude call. This page does not imply a live LLM call is happening.

Request credit limit increase

Defaults match the worked example from the MidasBrain blueprint.

Decision ID: —

Autonomy tier: —

Decision Record (from DynamoDB, not from the agent's own response):

Raw agent response

MidasBrain — conceived and produced by Antonio D. Esposito

Solution Architect & Product Lead · Delivered by Accenture