Skip to content
Back to Magazine
automation-aiops Updated 3 min read

Sandboxed Work: the new execution perimeter for production agents

Key takeaways
  • → Filesystem: what it can read, create, or modify.
  • → Network: which endpoints it may touch and from where.
  • → Secrets: which credentials are injected and for how long.
  • → Tools: which commands, APIs, or MCP servers it may invoke.
Contents

Problem

An agent that only drafts text can fail with little damage. An agent that executes code, queries private systems, installs dependencies, writes files, or touches infrastructure fails on another level.

The error is no longer a bad response. It is a side effect.

Many companies try to solve it with prompts: “don’t do anything dangerous”. That is not a control. It is a hope written in natural language.

Thesis

Sandboxed Work will be the new perimeter for agents that do real work.

The point is not to cage the model. The model does not execute. The point is to cage the action: filesystem, network, secrets, tools, processes, time, cost, and permissions.

Mature architecture separates brain, orchestrator, sandbox, and target systems.

Framework

An agentic sandbox must define five limits:

  • Filesystem: what it can read, create, or modify.
  • Network: which endpoints it may touch and from where.
  • Secrets: which credentials are injected and for how long.
  • Tools: which commands, APIs, or MCP servers it may invoke.
  • Timebox: how long the work runs before the environment is killed.

Mini-case: a support agent reproduces a client bug. In a sandbox it can clone the repo, install dependencies, run tests, read sanitized logs, and propose a patch. On a shared machine it could contaminate the environment, leak secrets, or leave processes alive.

Measurable signal: percentage of agent actions with ephemeral environment, logs, and declared limits.

Stance: if the agent can execute, it must also be containable.

Why it matters now

Cloudflare has pushed sandboxes for agents as isolated and scalable environments. Anthropic has introduced Managed Agents with external sandboxes and MCP tunnels. AWS has taken MCP to cloud operations with IAM, CloudWatch, CloudTrail, and bounded execution.

The trend is not just “more agents”. It is agents with hands.

And when a system has hands, it needs gloves, a cleanroom, and movement logs.

Anti-example

“We have a shared runner for all agents.”

That mixes contexts, permissions, and execution residues. An exploration agent should not live in the same perimeter as an agent that touches production.

Protocol (3 steps)

  1. Classify actions by risk. Read, write, execute, private network, secrets, and production.
  2. Create sandboxes by class. Do not use the same environment for investigation, build, and sensitive systems.
  3. Destroy by default. The environment must expire, log, and clean up.
Layer Minimum control Operational question
filesystem isolated directory what it can read
network allowlist where it can call
secrets ephemeral tokens how long they last
tools permissions per tool what it can do
time timeout when it dies

Sources consulted

Next step

Draw an execution diagram of your most dangerous agent: model, orchestrator, sandbox, secrets, network, tools, logs, and target system. If you don’t know where the limit is, you still don’t have a perimeter.


Translated from the Spanish original with AI assistance and reviewed for accuracy. Read the original in Spanish.

Prepare a meeting with this article

Decision

Separate reliable automation from fragile demo before granting it autonomy.

Meeting

Operations review, architecture, security or platform.

Risk

Adding speed with no observability, rollback, ownership or stop criterion.

Agent prompt: identify guardrails, control points, likely failures and autonomy criteria

sandboxes agents aiops cloudflare
Cite this article

Berthelius, V. (2026). “Sandboxed Work: the new execution perimeter for production agents”. BRTHLS Magazine. https://www.brthls.com/magazine/sandboxed-work-execution-perimeter-en

Read next

  1. Editorial connection

    Claude Managed Agents + Cloudflare: the perimeter becomes a product

    Enterprise agents shift focus from model quality to controlled execution environments, with Claude Managed Agents and Cloudflare leading the perimeter-as-product trend.

    Updated 4 min read

  2. Editorial connection

    AI Traces: the layer that turns agents into auditable systems

    Explains why AI tracing is essential for enterprise agents, outlining a five-level framework and steps to implement auditable runs.

    Updated 4 min read

  3. Editorial connection

    Microsoft Foundry Local + Scout: When Agent Work Moves to the Edge

    Microsoft unites cloud and edge with Foundry Local and Scout, redefining agent architecture for enterprise operations.

    Updated 5 min read

Fractional CAIO · Free diagnostic

Is your company ready to operate with AI?

30 minutes. No pitch. An honest read on where you are and what to move first.

Book free diagnostic