Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Duration 21 hours
Course Outline
Foundations: Threat Models for Agentic AI
- Categorizing agentic threats: misuse, privilege escalation, data leakage, and supply-chain vulnerabilities
- Analyzing adversary profiles and attacker capabilities specific to autonomous agents
- Mapping critical assets, trust boundaries, and control points for agent interactions
Governance, Policy, and Risk Management
- Establishing governance frameworks for agentic systems, including roles, responsibilities, and approval gates
- Designing policies covering acceptable use, escalation protocols, data handling, and auditability
- Addressing compliance requirements and strategies for evidence collection during audits
Non-Human Identity & Authentication for Agents
- Defining agent identities using service accounts, JWTs, and short-lived credentials
- Implementing least-privilege access patterns and just-in-time credentialing strategies
- Managing identity lifecycles, including rotation, delegation, and revocation
Access Controls, Secrets, and Data Protection
- Applying fine-grained access control models and capability-based patterns for agents
- Implementing secrets management, encryption in transit and at rest, and data minimization practices
- Safeguarding sensitive knowledge sources and PII from unauthorized agent access
Observability, Auditing, and Incident Response
- Creating telemetry for agent behavior, including intent tracing, command logging, and provenance tracking
- Integrating with SIEM tools, defining alerting thresholds, and ensuring forensic readiness
- Developing runbooks and playbooks for managing agent-related incidents and containment
Red-Teaming Agentic Systems
- Planning red-team exercises, defining scope, rules of engagement, and safe failover mechanisms
- Applying adversarial techniques such as prompt injection, tool misuse, chain-of-thought manipulation, and API abuse
- Executing controlled attacks to measure exposure levels and impact
Hardening and Mitigations
- Deploying engineering controls such as response throttles, capability gating, and sandboxing
- Enforcing policy and orchestration controls via approval flows, human-in-the-loop mechanisms, and governance hooks
- Implementing model and prompt-level defenses through input validation, canonicalization, and output filtering
Operationalizing Safe Agent Deployments
- Adopting deployment patterns such as staging, canary releases, and progressive rollouts for agents
- Managing change control, testing pipelines, and pre-deployment safety checks
- Coordinating cross-functional governance across security, legal, product, and operations teams
Capstone: Red-Team / Blue-Team Exercise
- Executing a simulated red-team attack against a sandboxed agent environment
- Defending, detecting, and remediating incidents as the blue team using established controls and telemetry
- Presenting findings, remediation plans, and recommended policy updates
Summary and Next Steps
Requirements
- Strong expertise in security engineering, system administration, or cloud operations
- Proficiency with AI/ML concepts and an understanding of large language model (LLM) dynamics
- Practical experience in Identity and Access Management (IAM) and secure system architecture
Target Audience
- Security engineers and professional red-teamers
- AI operations and platform engineers
- Compliance officers and risk managers
- Engineering leaders overseeing the deployment of AI agents
Testimonials (1)
inventory and identifying the different risk exposures within AI
