Blog
Agent safety, isolation, and building with Islo.
· Rotem Tamir
In a controlled Harbor run, the tests failed and a fairness judge caught the cheat, yet submitted code overwrote the final score. We examine the trust boundary inside verification and the Terminal-Bench community's work to secure it.
· Adam Gold
Replaying real engineering work sounds straightforward until the agent needs the internet. The eval has to freeze that internet at the moment the work was done.
· Adam Gold
ARIMLABS moved agent evaluations from local machines to an always-on Islo workflow with isolated computers, parallel trials, and inspectable reruns.
· Adam Gold
AI coding began on developer laptops. Autonomous work needs shared, persistent, observable and governable cloud environments.
· Adam Gold
What is a software factory? Compare Devin, Cursor, Linear, Factory, and Islo across autonomy, infrastructure, verification, governance, and fit.
· Adam Gold
Run Islo coding agents on infrastructure you own—your AWS account or on-prem Linux machines—while Islo keeps operating the product for you.
· Adam Gold
Alien runs AI agents called Sims on Islo to test deployment scenarios across clouds, architectures, and customer-controlled environments.
· Adam Gold
How OpenAI's model escaped its evaluation sandbox through an allowlisted package proxy, and why host-level allow and deny is the wrong granularity for containing AI agents.
· Adam Gold
How Obert uses Islo to turn well-scoped product work into trusted agent workflows and give a small AI-native team more engineering capacity.
· Assaf Ben Josef
We automated implementation first and it backfired. Starting with review and verification instead, and wiring it all together with webhooks and label-driven triggers, gave us an autonomous engineering loop where a Linear label goes in and a merged PR comes out.
· Rotem Tamir
Introducing RewardHackBench, an open-source benchmark for measuring whether sandbox policies prevent reward hacking in agent evaluations.
· Adam Gold
A technical look at what Agent Sandbox solves inside Kubernetes, what it leaves to the platform team, and how to decide if it's the right primitive for your setup.
· Adam Gold
Fast boot times are useful for request-shaped compute. Coding agents need something closer to a secure, persistent, workstation-shaped runtime.
· Adam Gold
A framework for evaluating what you need to run AI coding agents at scale and where most teams get stuck.
· Adam Gold
Claude Code for Web is great for quick coding help in the browser. Islo lets you run Claude Code alongside Cursor, orchestrate multiple agents, and connect to real services - all unattended.
· Adam Gold
Cursor Cloud Agents is great for quick edits in the browser. Islo lets you run Cursor alongside Claude Code, orchestrate multiple agents, and connect to real services - all unattended.
· Adam Gold
Stateful fake servers for GitHub, Slack, and Stripe that work with official SDKs - built for AI agent development.
· Yossi Eliaz
Monty sandboxes runtime-generated Python. Islo gives coding agents a zero-setup environment. Here's what Monty blocks, what can still crash, and how the layers fit.
· Adam Gold
Here are five sandbox options that work with Claude Code (or with AI coding agents in general), and how they compare.
· Adam Gold
Why sandboxing matters for AI agents - isolation, boundaries, and blast radius so you can run agents unattended without risking credentials or production.
· Adam Gold
How I sent my own database credentials to a server I control using Clawdbot - and why prompt injection and full machine access make that possible in about 20 minutes.
