k
keynews.ai
keynews.ai
AgentsThe story, in brief

Run multiple isolated agents in a single Sandbox

Vercel's Sandbox now isolates multiple agents in a single container — file permissions and user isolation for multi-agent systems are now built in.

Reporting by Vercel Blog 1 min brief
Illustration of independent geometric mechanisms passing paper tasks along branching amber tracks.
AI agents and the coordination of work.AI illustration by KeyNews
The KeyNews take

Why it matters

Multi-agent orchestration just got simpler and safer. Practitioners building agent systems now have infrastructure-level isolation baked into a major platform, reducing the work of sandboxing and access control.

The key facts

11 to know

  1. Vercel Sandbox SDK now supports multiple Linux users and groups

  2. Each agent runs as its own user with private home directory

  3. Users cannot read, write, or list each other's files

  4. Shared directories available via group creation for collaboration

  5. Reduces complexity of multi-agent system architecture

  6. Vercel Sandbox SDK adds support for multiple Linux users and groups

  7. Each agent runs as isolated user with private home directory

  8. Shared workspace available via group creation for collaboration

  9. User isolation prevents cross-agent file access (read/write/list)

  10. Available in @vercel/sandbox via createUser() and createGroup() APIs

  11. Shipped July 2026

Go to the source

Vercel Blogvercel.com

Publisher excerpt: The SDK now supports multiple Linux users and groups, so you can run agents side by side in a single Sandbox.@vercel/sandbox Each agent runs as its own user with a private home directory. A group opens a shared workspace when they need to collaborate. This makes multi-agent systems easier to build.…
Read original report
Back to today's editionMore agents news

The wider picture

View all
Illustration of independent geometric mechanisms passing paper tasks along branching amber tracks.
AI illustration by KeyNews
Agents01

Presentation: APIs for Agents: Rethinking API Programs in the MCP Era

Enterprise agents aren't a prototype problem anymore—they're a platform problem. This is how a major financial institution engineered governance, safety, and scale for multi-agent workflows in production.

InfoQ AI/ML
Illustration of independent geometric mechanisms passing paper tasks along branching amber tracks.
AI illustration by KeyNews
Agents02

Can Agentic AI Bridge the Gap with Trusted Enterprise Data?

As agentic AI moves from pilots to production, enterprises face a hard constraint: agents need access to data to be useful, but that access must be verifiable and trustworthy. This is an operational and security problem that will shape how agents are deployed at scale.

SAP News
Illustration of independent geometric mechanisms passing paper tasks along branching amber tracks.
AI illustration by KeyNews
Agents03

How Dr Martens is working with Salesforce to create ‘agentic experiences’ for customers

A major consumer brand is moving beyond chatbots to agentic customer service at scale. This is a real deployment case study showing how agents are reshaping retail operations and customer experience — exactly the kind of industry transformation practitioners need to watch.

ITPro