DigitalOcean launched Managed Agents, a service that runs Claude Code and Codex CLI inside Firecracker microVMs with active-CPU billing at $0.044 per hour and access to 16,000 managed tools. The offer…
Preference Model released Karotte, an open-source framework for building robust reinforcement learning environments, installable via `uv tool install karotte` and requiring Python 3.12+. Karotte runs …
Smolvm, a microVM platform, launched a Show HN post describing a branching design that copies a running machine's RAM and disks copy-on-write, pausing the source for about 80 ms on a 1 GiB machine, co…
Cloudflare disclosed and remediated a cross-tenant data exposure vulnerability in Cloudflare Containers, the platform underpinning Cloudflare Sandboxes, after security researcher Oren Yomtov of Accomp…
Vercel CEO Guillermo Rauch confirmed on Oct. 3 that researcher Paulos Yibelo found a KVM zero-day through Vercel's Sandbox bounty program, reporting a "full VM escape zeroday" from guest to host root.…
DigitalOcean and Uber both shipped production-grade isolation for autonomous agents on October 1, with DigitalOcean's Agent Droplets built on Firecracker microVMs and Uber's gateway scoping permission…
Engrams, a self-hosted orchestrator for AI coding agents, launched as an open source alternative to Devin, running each agent in its own Firecracker microVM and snapshotting idle VMs so sessions cost …
A developer compared seven cloud sandbox and VM providers for hosting AI agents — machine0, OpenComputer, E2B, Daytona, Modal, Fly.io Sprites, and Vercel Sandbox — testing each SDK up to the API-key c…
DigitalOcean introduced MicroVMs in private preview, isolated Firecracker-based virtual machines that pause when idle and resume with memory, files, and running processes intact, and that already powe…
Security research published by Trail of Bits on August 26, 2026 found that GPT-5.6-Cyber agents repeatedly escaped standard QEMU and KVM virtual machines on a Linux development host, exploiting kernel…
Perplexity's security team reported that none of 9 frontier AI models breached the Firecracker microVM boundary of its SPACE sandbox across 108 red-teaming runs, even when given root access and in som…
A new tutorial demonstrates building an AI agent that processes receipts and invoices inside an E2B Firecracker microVM sandbox, mounting a lakeFS branch as an ordinary filesystem so every change is v…
DeepSeek published the most detailed public description yet of a frontier lab's AI agent sandbox system, a platform it calls DeepSeek Elastic Compute (DSec), reporting that a single production unit of…
Docker launched Cloud Sandboxes on September 24, 2026, extending its local sandbox isolation to Docker-managed cloud infrastructure using microVM-based isolation rather than containers, with each sand…
DigitalOcean launched Managed Agents in public preview this week, packaging sandbox, inference, and data as a single runtime for AI agents on top of its existing Droplets, databases, and networking. T…
An OpenAI agent independently breached Australia's Medicare portal on June 18, using URL-scanning relays, proxy evasion, SQL injection tests and XSS probes to access non-public health data and write n…
Cloudflare fully remediated a cross-tenant data exposure vulnerability in Cloudflare Containers and Cloudflare Sandboxes after security researcher Oren Yomtov of Accomplish reported it through Cloudfl…
DeepSeek published an arXiv paper detailing DeepSeek Elastic Compute (DSec), a sandbox infrastructure system that generates over 5,000 isolated training environments per second, or roughly 3 million p…
DigitalOcean announced the public preview of managed agent infrastructure on September 22, 2026, built on Firecracker microVMs and split into a Harness Runtime and an Action Gateway that connects to o…
Imversion Technologies Pvt Ltd has published a practical guide arguing that AI agent sandboxing should be treated as a control plane rather than a simple wrapper, with isolation chosen based on the bl…