Cycode launches workstation protection for developers

Cycode has launched Workstation Protection for developers, adding real-time checks on package installs at developer workstations.

The feature is designed to stop malicious software packages before they are installed on a machine. It sits within Cycode's Agentic Development Lifecycle security offering and extends controls from code and prompts to the developer device itself.

Workstation Protection intercepts package installations in real time and assesses them before the software reaches the workstation. It applies two controls at that point: release-age gating, which blocks versions published too recently, and checks against a threat intelligence feed to stop packages already identified as malicious.

The release-age control is intended to address the period immediately after publication, when compromised software releases are often first detected. Enforcement happens on the device rather than through local package manager settings, with the aim of applying the same policy regardless of package manager version or configuration.

Each installation is also checked against a continuously updated list of known malicious packages. This allows Cycode to block software that has already aged beyond any cooldown period, as well as older versions that remain harmful.

Attack surface

The launch reflects growing concern over software supply chain attacks aimed at developer environments, not just source code repositories or build systems. Such attacks can install malicious packages on workstations to steal credentials, establish persistent access, and create a route into wider software delivery processes before code is even written or scanned.

Cycode linked that trend to the wider use of AI-assisted development tools and coding agents. It said attackers are exploiting both compromised maintainer accounts in open-source ecosystems and attempts to manipulate coding agents into installing harmful packages.

That, Cycode argued, has widened the attack surface from developers and source code management systems to every workstation running an automated coding tool. The company cited several recent attacks as examples.

In one case involving Keyv, Cycode said a hijacked maintainer account spread a preinstall worm across more than 800 packages and over 1,300 versions, affecting software representing more than two billion monthly installs. According to Cycode, the attack harvested cloud and continuous integration credentials.

It also pointed to an incident involving LiteLLM, where malicious packages were published in a library with 95 million monthly downloads. Cycode said the attack led to credential theft, movement across Kubernetes environments, and a persistent backdoor.

Another example was Shai-Hulud 2.0, which Cycode described as a self-replicating npm worm that reached about 350 maintainers and exfiltrated secrets to more than 25,000 attacker-created GitHub repositories.

Earlier control

Cycode said the product is deployed through mobile device management and does not require a separate console or additional infrastructure. Security teams can set cooldown policies centrally, with the intention of extending the same protections quickly across developer machines running coding agents.

The company positions the workstation feature as an extension of a broader system spanning AI tools, prompt activity, and generated code. Signals from those areas are brought together in what Cycode calls a Context Intelligence Graph to give security teams a single view of risk across the development lifecycle.

"Risk enters the agentic development lifecycle (ADLC) for our customers before a single line of code is generated. Developers connect AI tools nobody approved, agents install packages nobody reviewed, and secrets leak into prompts and file reads nobody sees," said Seth Robbins, President of Cycode.

"This launch extends Cycode's platform to the workstation, the earliest control point in the software supply chain where malicious packages are installed and must be stopped," Robbins said.

The launch comes as security suppliers and software teams adapt their controls to the spread of generative AI tools in engineering workflows. As more developers rely on assistants and autonomous coding agents to fetch dependencies, write code, and interact with repositories, security concerns are shifting to earlier stages of the software creation process.

Cycode said the new workstation feature is available in early access. "Security has to start where risk enters the development lifecycle, not after," said Dor Atias, Co-Founder and Chief Product and Engineering Officer at Cycode.

"Blocking malicious packages is the earliest control point in a complete agentic development and software supply chain security solution. Protect every workstation. Control every input. Secure every output. That is what unlocks secure AI development," Atias said.