SYNOPSIS
About the Workshop
The age of autonomous AI agents is here and so are the security challenges that come with them. As organizations race to deploy AI agents in production, many are building shaky foundations: underdefined permissions, unclear audit trails, and governance gaps that auditors and adversaries alike will eventually find. The AgentCore Security Workshop is your hands-on experience with those gaps and your blueprint for closing them.
In this intensive, practitioner-focused workshop, you'll move beyond theory and get your hands dirty building secure AI agents on Amazon Bedrock AgentCore, AWS's purpose-built platform for production-grade AI agents that works across any framework and any model. You'll implement the security controls that matter most such as robust identity management, governance guardrails, data protection patterns, and the continuous monitoring posture that turns observability into accountability. Every lab is grounded in real-world threat scenarios that production AI teams actually face.
You'll walk out knowing how to build AI agents and how to lock them down properly, covering identity, governance, data protection, and monitoring in a single hands-on session.
Key Learning Objectives
By the end of this workshop, you will be able to:
- Design and implement least-privilege identity models for AI agents using IAM roles, policies, and permission boundaries that constrain what agents can access and what they can't
- Configure and enforce governance guardrails on Amazon Bedrock AgentCore to control agent behavior, prevent prompt injection, and align AI outputs with organizational policy
- Apply data protection controls to sensitive inputs and outputs flowing through agent pipelines, including encryption at rest and in transit
- Deploy serverless agent tooling securely using AWS Lambda, isolating execution logic without exposing infrastructure management overhead
- Instrument end-to-end agent activity logging with AWS CloudTrail to capture API calls, user actions, and cross-environment events in hybrid and multicloud deployments
- Build a continuous monitoring posture that surfaces anomalous agent behavior, supports incident response, and satisfies compliance and audit requirements
- Translate security principles into production-ready code through guided labs that mirror real deployment scenarios on Bedrock AgentCore.
Who Should Attend
Participants should ideally have some familiarity with AWS security services such as IAM, CloudTrail and Lambda. Prior awareness of Amazon Bedrock AgentCore would also be helpful, although not mandatory.
Registration is open to all members of the SG chapter and is subject to approval due to limited capacity. Priority will be given to women members of the SG chapter (active status). Registration confirmation will be sent by 31 July 2026.
ABOUT THE INSTRUCTOR
Kevin Low
Senior Security Solutions Architect at Amazon Web Services (AWS)
Kevin is a Senior Security Solutions Architect in AWS where he helps customers across ASEAN build securely. He specialises in threat detection and incident response and AI Security. Outside of work, he is trying to train his poodle, Noodle, to play football.
************************************
This is a chapter professional development event thus 1.5 CPE hours (Group B) will be available for your CPE submission. To facilitate submission of CPE points on your behalf by the local chapter - please ensure your ISC2 member ID is updated in your profile.
Notes:
1) As spaces are limited, if you cannot attend after registration confirmation, please cancel your registration or email info@isc2chapter.sg at least 2 business days before the event.
2) For ISC2 members residing in Singapore who are not yet members of our local Singapore Chapter, please sign up with us through our ISC2 Singapore Chapter website under "Membership-Join us"
- https://www.isc2chapter.sg/join-us
Membership Rates:
Professional Member: $50/year
Associate Member (Non-credential holders): $30/year
Student Member: $10/year
