Cursor launches environments for cloud agents with multi-repo support and versioning
Cursor has launched tools for cloud agents: cloud-based developers can now work in full environments with their own code, dependencies and APIs. Multi-repo support lets agents see the entire project context. Docker-based builds have been improved, and security controls with change auditing have been added. The core problem is simple: an agent that can write code but cannot run tests, call an API or run a build is useless.
AI-processed from Cursor Blog; edited by Hamidun News
Cursor has introduced new tools for configuring the development environments of cloud developer agents. The core problem: an agent that can write code but cannot run tests, call an API or run a build is useless. Cloud agents need environments that resemble your laptop, with cloned repos, installed dependencies, and access to APIs and build systems.
Multi-repo environments
In large companies, code is split across dozens of repositories. A microservice architecture requires changing code in several places at once: a new API has to be integrated into three microservices, added to the database and reflected in the frontend. Cursor has added support for multi-repo environments, so a single agent can now see all the repositories it needs at the same time and understand how a change in one part of the project affects the others.
This is especially useful for large organizations. Amplitude uses Cursor Automations in its Slack channels. According to Stephen Chen, Senior Engineering Manager at Amplitude, an agent can work out a problem, determine on its own which repos are involved, and open a PR with a fix in the right places, with full context.
What multi-repo support provides:
- The agent sees the links between different repos and microservices
- It can open PRs with changes in several places at once
- It reuses one environment across sessions instead of recreating it from scratch
- It gets the full project context needed for complex tasks
- It works with tools that require access to several repos at once
Docker configuration and automatic builds
Cursor has improved support for Dockerfile-based configuration of cloud environments. The new feature is build secrets: a secure way to pass private keys for package registries directly into the Dockerfile. These secrets are used only while building the image and do not end up in the agent's working environment.
Layer caching has sped up builds by 70%, since only changed layers are rebuilt. This matters when environments are updated frequently.
If you don't want to write a Dockerfile from scratch, Cursor can generate one automatically. The system scans your repos, identifies the tools and dependencies you need, and produces a ready-made configuration for versioning. The feature is still in private beta and is rolling out to Enterprise customers.
Versioning, auditing and security
Cursor now has a better understanding of the state of the dev environment while it works. During setup, the system asks questions, highlights missing credentials and validates the configuration. The agent can see which version of the environment it is running in.
Reliability has also been improved: if the configuration fails, the agent does not crash but falls back to a base image with warnings.
Every environment now has a version history, so you can review changes and roll back to an earlier version. Admins can restrict rollback to themselves. The audit log records every action. Ingress and secrets are now scoped to the individual dev environment. A team can grant different access levels to different environments; for example, staging may have access to production APIs, while dev gets only mock services.
What this means
Cloud agents are moving from experimental tool to production development. Full environments with multi-repo support, versioning and security controls make them capable of serious tasks, from finding and fixing bugs to deploying entire features. For engineering teams, this means a significant share of routine development may move to the cloud.
- Nov 1, 2026OpenAI Rewrites GPT Behavior Rules: Five Values and a Chain of Command
- Oct 31, 2026Author puts his script before an AI jury: the verdict cost 850 thousand tokens
- Oct 31, 2026Anthropic: Fold Subagents Back Into a Single Loop — One Customer's Tasks Ran Twice as Fast
- Oct 30, 2026Mistral opens Large 4 weights: strong at code, but Chinese models score 32 to its 26
Key Questions Answered
What are the key details on Multi-repo environments?
In large companies, code is split across dozens of repositories. A microservice architecture requires changing code in several places at once: a new API has to be integrated into three microservices, added to the database and reflected in the frontend. Cursor has added support for multi-repo environments, so a single agent can now see all the repositories it needs at the same time and understand how a change in one part of the project affects the others. This is especially useful for large organizations. Amplitude uses Cursor Automations in its Slack channels. According to
What are the key details on Docker configuration and automatic builds?
Cursor has improved support for Dockerfile-based configuration of cloud environments. The new feature is build secrets: a secure way to pass private keys for package registries directly into the Dockerfile. These secrets are used only while building the image and do not end up in the agent's working environment. Layer caching has sped up builds by 70%, since only changed layers are rebuilt. This matters when environments are updated frequently. If you don't want to write a Dockerfile from scratch, Cursor can generate one automatically. The system scans your repos, identifie
What are the key details on Versioning, auditing and security?
Cursor now has a better understanding of the state of the dev environment while it works. During setup, the system asks questions, highlights missing credentials and validates the configuration. The agent can see which version of the environment it is running in. Reliability has also been improved: if the configuration fails, the agent does not crash but falls back to a base image with warnings. Every environment now has a version history, so you can review changes and roll back to an earlier version. Admins can restrict rollback to themselves. The audit log records every a
Build & Ship Production AI Agents Without Boilerplate Code
Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.