AI Security workshop
Securing AI Agents
Control what your agents can do before they do it.
- 1 day
- Hands-on in AI Governance Platform
- On your own system
The problem
Agents with tool access turn a bad model output into a real action. This workshop designs permissions, pre-execution checks, and audit trails for your agents.
Who it's for
Security architects and engineering teams deploying agents with access to systems or data.
What to bring
Your agents’ tools, the systems they touch, and who can trigger them.
The course side
What your team learns
- 1
Agent inventory
What agents run, what they can access, and who owns them.
- 2
Excessive agency
Threats specific to LLM agents and how they get exploited.
- 3
Pre-execution policy checks
Stop unsafe actions before they run, with human approval where needed.
- 4
Evidence and audit
Logging every decision so incidents can be reconstructed.
Builds on these course lectures
- The Sources of Threats for LLM Agents
- Agents vs Responsible AI Adoption
- GenAI Emerging Defenses
The platform side
What your team builds in AI Governance Platform
- Inventory your agents, their tools, and the systems they touch
- Risk score each agent by access and autonomy
- Record permission and pre-execution policies as controls
Also used: Agent permission matrix, Runtime policy patterns
About AI Governance Platform →Your team leaves with
- • An agent permission matrix
- • Runtime policy rules and an audit-logging design
Book this workshop
Request Securing AI Agents
It's pre-selected in the form. Add others to combine them. Contact us for pricing: fill in your details and we'll reply within one business day.
Often combined with
- Hands-On LLM Red-Teaming · 2 days
- Prompt Injection & Data Leakage Defense · 1 day