In short
Check Point AI Agent Security protects the AI agents and AI applications an organisation builds or deploys. It discovers agents on supported cloud and low-code platforms, assesses each agent's tools, skills and MCP servers, controls which tools they may use, checks actions at runtime and defends against prompt attacks and sensitive data exposure. It builds on Lakera's technology.
Get a quote for Check Point AI Agent Security
Tell us roughly how many users or sites you have and we will reply by email with options and a quote.
What Check Point AI Agent Security is
Check Point AI Agent Security is Check Point's product for securing AI agents and AI applications that organisations build, deploy or run across business workflows: agents that talk to customers or colleagues, process company data, connect to tools and systems, and take actions on the organisation's behalf. In Check Point's AI Defense Plane it appears as the AI Application & Agent Security module, alongside Workforce AI Security for employee use and AI Red Teaming for testing.
Check Point's headline is "Secure AI Agents from Discovery to Runtime". The product page sets out three jobs: discover agents being built and deployed across the business, assess each agent's risk in context, and control agent actions at runtime through policy.
It is part of the wider Check Point AI Security portfolio. If your concern is staff using third-party AI tools rather than agents you build, Workforce AI Security is the better starting point.
What AI Agent Security helps you control
Check Point lists six capabilities:
- Agent discovery and inventory: identify agents running across supported cloud and low-code platforms, for a clearer view of your agent estate.
- Per-agent risk assessment: assess risk across each agent's metadata, usage, tools, skills, MCP servers and connected components.
- Tool and MCP access control: govern which tools and MCP servers agents can use, with allow and deny controls enforced at runtime.
- Agent action control: evaluate tool calls and actions in context to help block unsafe or unauthorised behaviour.
- Prompt attack and data protection: help detect prompt attacks, indirect injection and sensitive data exposure across prompts, responses and tool use.
- Unauthorised action detection: detect agent actions that fall outside user intent, policy or approved workflows.
Check Point says the product supports different AI models and providers, so policies can stay consistent as your architecture changes.
How it works at runtime
Check Point describes AI Agent Security as an inline enforcement layer for agent interactions. It can inspect, in real time:
- prompts sent to the model,
- the model's responses,
- external content the agent pulls in, such as web pages and files, and
- the tool calls the agent makes.
It then applies your security and governance policies before sensitive data is exposed, unsafe output is returned or a risky action runs. Depending on policy, sensitive information can be blocked, redacted or otherwise controlled before it reaches the model, goes back to a user or triggers an action.
This matters because, as the NCSC has pointed out, language models do not keep a hard boundary between instructions and data, so prompt injection is managed rather than cured. Inspecting tool calls and restricting what agents can do limits the damage if a model is misled.
Is Check Point AI Agent Security the right fit for you?
We will check it against your systems and requirements and recommend the right option.
Threats it is designed to address
Check Point groups the threats as:
- Unknown or ungoverned agents deployed across cloud and low-code environments without central oversight.
- Risky tools, permissions and connections: MCP servers, connected systems and unsafe agent actions.
- Prompt attacks and content risks: prompt injection, indirect prompt injection, jailbreak attempts, data exposure, harmful outputs and policy violations.
Indirect attacks deserve emphasis. An agent that reads emails, tickets, documents or web pages can meet instructions planted by anyone who can write to those sources. Our MCP security guide covers the tool-access side in more depth.
AI Agent Security vs Workforce AI Security
The two products are often confused because both mention agents. The difference is whose AI it is.
| Workforce AI Security | AI Agent Security | |
|---|---|---|
| Protects | Employees using AI tools: chat apps, SaaS AI features, desktop apps, coding assistants, agents and MCP on their devices | AI agents and applications your organisation builds and deploys |
| Typical risk | Staff sending sensitive data to AI, shadow AI, risky agent actions on endpoints | Customers or content manipulating your agent; agents misusing tools or exposing data |
| How it sees activity | Browser extension, plus desktop agent in the Enterprise tier | Inline enforcement for agent interactions, plus discovery across supported cloud and low-code platforms |
| Typical owner | IT and security teams | Product, engineering and application security teams |
| Tiers | Essentials and Enterprise | Confirmed at quote |
Plenty of organisations need Workforce AI Security only. If you are building agents as well, the two complement each other. See Workforce AI Security Enterprise for the employee side.
Not sure which option fits? Ask us
Tell us what you need and we will recommend the right option for your organisation, with a quote.
From Lakera Guard to Check Point AI Agent Security
AI Agent Security grew out of Lakera. Check Point announced an agreement to acquire Lakera, which has AI research and development centres in Zurich and San Francisco, on 16 September 2025. Lakera was known for Lakera Guard, its real-time runtime protection for LLM applications and agents, and Lakera Red, for red teaming. Lakera's website now points those products to Check Point AI Agent Security and Check Point AI Red Teaming, which Check Point names as Lakera's flagship solutions.
If you have been searching for "Lakera Guard", this is the product line to look at today. Check Point's AI Defense Plane announcement in March 2026 also cited technology from its Cyata acquisition. We use Check Point's current product names when we scope and quote.
Who AI Agent Security fits
It is likely to be relevant if you can answer yes to any of these:
- Do you run, or plan to run, customer-facing AI assistants or chatbots?
- Are teams building agents on cloud or low-code platforms that act on company systems?
- Do your agents connect to tools or MCP servers that can change data, send messages or spend money?
- Do agents process content from outside the organisation, such as emails, uploads or web pages?
- Do you need evidence for regulators, auditors or customers that your AI applications are governed?
If the answer to all of them is no, your AI risk is probably about employee use, and how Workforce AI Security works is the better read.
Pairing it with red teaming and your firewall
AI Agent Security protects at runtime; it works best alongside testing and network controls:
- AI Red Teaming tests your models, applications and agent endpoints against prompt injection, jailbreaks, data leakage and unsafe tool use before release and after changes. See the red teaming section of our Check Point AI Security guide.
- AI Network Firewall: Check Point's firewall software release R82.20 brings AI Agent Security for LLM applications to Check Point firewalls through AI Guardrails, including prompt injection protection, content moderation and protection for MCP tool responses. See Check Point AI Network Firewall.
- Web application protection: for the web apps that host your AI features, our sister site covers Check Point CloudGuard WAF.
How Forever Systems helps
We help you work out whether AI Agent Security belongs in your plans and, if it does, how to approach it:
- We talk through the agents and AI applications you run or plan, the platforms they use and the tools they can reach.
- We help you decide between Workforce AI Security, AI Agent Security or both, and whether red teaming before release makes sense.
- We work with Check Point to confirm what is supported for your platforms and quote accordingly; ask us for a quote.
To discuss Check Point AI Agent Security, contact us.
How Forever Systems helps you use AI safely
We are an independent UK IT and cyber security provider. We supply Check Point Workforce AI Security and help organisations put sensible, workable rules around AI at work.
- See what is really in use: we can start with a discovery pilot, so you know which AI tools, browser extensions, coding assistants and agents your people use before anyone writes rules
- Policy you can enforce: we help you decide what to approve, restrict or block, and turn your AI acceptable use policy into Workforce AI Security policies
- The right tier: Essentials for browser-based AI, Enterprise when desktop apps, developer code assistants and agents are in the picture, confirmed at quote
- Staged roll-out: from monitoring to warnings, redaction and blocking at a pace your people can absorb, with reviews along the way
- Around the tool: Microsoft 365 permissions ahead of Copilot, device management for the extension and agent, and safe use of AI training for staff
We work with businesses, schools, charities and public sector organisations. Our guides are general information, not legal advice.
Frequently asked questions
What is Check Point AI Agent Security?
It is Check Point's product for securing AI agents and AI applications that organisations build or deploy. It discovers agents across supported cloud and low-code platforms, assesses each agent's risk, controls which tools and MCP servers they can use, and inspects prompts, responses, external content and tool calls at runtime.
Is Check Point AI Agent Security the same as Lakera Guard?
It grew out of Lakera Guard. Check Point announced its agreement to acquire Lakera in September 2025, and Lakera's runtime protection and red teaming products are now offered as Check Point AI Agent Security and Check Point AI Red Teaming.
What is the difference between AI Agent Security and Workforce AI Security?
Workforce AI Security protects how employees use AI tools, including agents and MCP on their own devices. AI Agent Security protects the AI agents and applications your organisation builds and deploys, enforcing policy inline on their prompts, responses and tool calls.
Does AI Agent Security protect against prompt injection?
Check Point says it helps detect prompt attacks, including indirect prompt injection hidden in external content, and can block risky actions before they run. Because prompt injection cannot be fully eliminated, it is best combined with least-privilege tool access and red teaming.
Does it work with different AI models?
Check Point says AI Agent Security is designed to support different AI models and providers, so consistent policies can be applied across your agents. We confirm support for your specific platforms when we scope and quote.
Still have a question about Check Point AI Agent Security?
Ask us and a member of our team will reply by email.