AI Agent Runtime Security (OWASP Agent Control Standard) Fundamentals Training
A one-day fundamentals programme on OWASP Agent Control Standard. Learn how runtime controls can restrict what AI agents access and do within business systems. Participants work through practical examples and a guided case before drafting actions for their own organisation.
Who should attend: AI developers, architects, platform owners and security or technical governance teams. No prior framework knowledge is required.
HRD Corp claim support available
Programme Agenda
9:00 AM to 10:00 AM
Why Runtime Control Matters
Distinguish a policy written before deployment from a check applied when an agent proposes an action. Introduce the Agent Control Standard and its relationship to existing agent platforms.
10:00 AM to 10:15 AM
Break
10:15 AM to 11:30 AM
Agents, Guardians and Control Hooks
Explain an observed agent, a separate policy decision component and lifecycle hooks at a conceptual level. Trace where a proposed tool call can be inspected before reaching a business system.
11:30 AM to 12:30 PM
Policies and Decisions
Translate business authority into allow, deny or approval conditions using a payment-adjacent example without initiating transactions. Discuss identity, permission scope, user mandate and safe behaviour when a control cannot respond.
12:30 PM to 1:30 PM
Lunch
1:30 PM to 2:45 PM
Traceability and Component Visibility
Introduce action traces and agent component inventories. Identify the records needed to explain what ran, what was proposed, which policy applied and who approved an exception.
2:45 PM to 3:00 PM
Break
3:00 PM to 4:00 PM
Integration and Limitations
Discuss coverage gaps, unsupported hooks, failures, latency and version changes. Identify why a policy promise is unreliable if an execution path can bypass enforcement or logs cannot be trusted.
4:00 PM to 5:00 PM
Practical Exercise: An Agent Action Gate
Design a conceptual control gate for a fictional service agent. Define permitted actions, escalation decisions, evidence and failure handling, then review illustrative action requests.
Key Outcomes
- Explain the purpose of runtime agent controls
- Map control hooks and enforcement boundaries
- Draft action policies and approval conditions
- Identify traceability and integration gaps
Training Mode Physical / Online / Hybrid
HRD Corp Subject to programme eligibility and HRD Corp approval
Level Fundamentals / Beginner. No prior framework knowledge required.
Duration 1 Day | 9:00 AM to 5:00 PM (6.5 learning hours, excluding breaks)
Venue In-house at the client's premises, or delivered via the client's preferred platform (Microsoft Teams, Zoom, or equivalent)
Assessment Group case exercise and a 10-question knowledge check
Certificate Orbix Certificate of Completion upon full attendance; not a professional certification
Malaysia Relevance
High relevance for Malaysian AI developers and security teams implementing agent oversight. ACS is an evolving OWASP project; the course checks the current specification and distinguishes architectural concepts from verified interoperability.
Practical Takeaways
A runtime control decision table and agent action trace checklist.
Download a sample practice worksheet (TXT). This fictional exercise is a preview of the learning approach. About Orbix; contact us to confirm the facilitator and relevant experience for your delivery.
Course Completion and Next Steps
Participants receive an Orbix Certificate of Completion upon full attendance. This introductory course does not confer professional certification or establish an organisation's compliance. Further implementation and assessment depend on the organisation's scope and obligations.
Frequently Asked Questions
Related Fundamentals Training
Generative AI Security (OWASP LLM Top 10) Fundamentals AI Security Controls (CSA AICM) Fundamentals AI Agent Security (OWASP Agentic Top 10) Fundamentals AI Security Verification (OWASP AISVS) Fundamentals View all fundamentals programmes
Reference material: OWASP: Agent Control Standard. OWASP ACS: project documentation.