Droids now have a fleet of sandboxes
- Source
- factory.com
- Date
Context
Factory's Environments package a repository's tools and dependencies into a reusable starting point for cloud agents. The company builds portable images ahead of time, so a new Droid can begin from a prepared instead of repeating setup. Its October 12 announcement says that when an automation points at an Environment, each scheduled or event-triggered run starts in a fresh sandbox from the last validated build.
The reusable definition also has to follow changes in the codebase. Factory says it rebuilds Environments every few hours and promotes only successful builds. This separates maintaining the shared setup from running individual tasks, letting several agents use a prepared environment without depending on one long-lived machine.
- Factory's Readiness scores repositories out of five; the company says most enterprise customers start at level two and its Missions work best at level four or above.
- Automations triggered by schedules, Slack, GitHub, or webhooks, when pointed at an Environment, each start on a fresh sandbox from the last validated build and run as the user or a service account under the same policy.
- Admins control which hosts a Droid can reach, which shell commands need approval or are banned, and which Connectors and Skills are allowed; blocked commands are stopped by the before they run.
- sandbox — An isolated environment where AI-generated code or agent actions run without being able to touch anything real.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- multi-agent — Using several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.
- agent harness — The scaffolding around a model that turns it into a working agent — the loop, the tools it can call, and the rules for when to stop.
Checking sign-in…
Loading comments…