AI coding agent security

Let engineers use AI coding agents on real systems, with governed access.

Find the coding agents on your workstations, see what they do, decide the MCP calls they make on the laptop, and give them governed access to repositories and infrastructure through Vekri Gateway.

Contact Sales Request a Demo

What you can answer

The questions, and how Vekri answers them.

Which agents run where?
Inventory of coding agents, their versions, hooks and MCP servers on each enrolled macOS or Linux device.
- AI coding agent discovery
- MCP client and server inventory
What are they doing?
Claude Code's hooks report sessions, prompts and tool use, including built-in shell and file tools; local rules flag risky sequences.
- Agent lifecycle hooks for Claude Code
- Local behavioral detection
Which MCP calls are allowed?
Approved servers run behind the local MCP gateway, which asks Vekri to decide every tool call and inspects every result.
- Local MCP gateway
- Managed local MCP profiles
- Prompt-injection detection on the local MCP gateway
How do agents reach production?
Through Vekri Gateway, signed in as the employee, inside an assignment, with approval for privileged writes and a broker holding the credentials.
- Enterprise SSO for employees' agents
- Group-scoped access assignments
- Exact-operation approvals
- Credential broker

See it work

Step by step.

Local MCP gateway

Illustrative data

Claude Code

On Erin's laptop

vekri-endpoint mcp proxy

Local MCP gateway · profile github-eng

list_issues · allow

create_pull_request · approved, ran once

delete_repository · deny (profile)

Vekri unreachable · failed closed

Denied: delete_repository

GitHub MCP server

Started by the proxy

Received: list_issues

Received: create_pull_request

Never received: delete_repository

Decisions · user:erin · device ws-dev-07

Allowedlist_issues
Approvedcreate_pull_requestsingle use
Denieddelete_repositorynot executed

What happens

Applies to MCP servers configured through a Vekri profile. Built-in tools and servers configured directly are outside it, and inventory shows them as not mediated.

Capabilities

Go deeper.

AI application discovery
Know which AI applications run on your workstations.
MCP security
Decide every MCP tool call that runs through Vekri, on the laptop and in the cluster.
Prompt injection protection
Catch instructions hidden in tool results before the agent reads them.

Products

The parts of Vekri involved.

Contact Sales

See Vekri with your own AI applications, agents and systems.

Talk to us about the AI your organization uses, the systems agents reach and how you deploy: in Vekri's cloud, or inside your own Kubernetes cluster.

Contact Sales Request a Demo

Let engineers use AI coding agents on real systems, with governed access.

The questions, and how Vekri answers them.

Which agents run where?

What are they doing?

Which MCP calls are allowed?

How do agents reach production?

Step by step.

Go deeper.

AI application discovery

MCP security

Prompt injection protection

What it does not claim.

The parts of Vekri involved.

Vekri Endpoint

Vekri Protect

Vekri Gateway

Vekri Observe

See Vekri with your own AI applications, agents and systems.