Rogue Cursor and Claude Code add-ons tied to SSH access and persistence

Key points

• Harmful add-ons were spread to developers using Cursor and Claude Code.
• The add-ons enabled SSH access and maintained remote presence.
• Persistence files were placed on infected systems.
• Device information was gathered from Windows and macOS developer machines.
• No CVE was assigned.

Researchers have tied a set of harmful add-ons for Cursor and Anthropic’s Claude Code to unwanted SSH access, long-term remote presence, and profiling of developer hardware. The activity centers on people who publish extensions with deceptive intent, rather than on a flaw tracked by a CVE; no CVE identifier has been assigned.

Developers on both Windows and macOS who use these tools are in scope. According to the facts, the add-ons were distributed to those users, persistence files were placed on affected devices, and details about the machines were gathered. That combination matters because developer workstations often hold source code, credentials, and access to production systems.

The SSH component is especially concerning: it can give an intruder a way back into a machine after the initial extension is removed, while the collected device details may help an attacker decide what to target next. Because the extensions live inside familiar development tools, they can blend into normal workflows and escape casual review.

Context: extension marketplaces and plugin ecosystems are attractive because installation is easy and trust is often implicit. The absence of a CVE does not mean the risk is negligible; it means the issue is supply-chain and user-trust related rather than a single patched software bug. Watch for new reports naming specific extensions, updates from Cursor and Anthropic, and signs of unexpected SSH access or persistence artifacts on developer endpoints.

What to do now

1. Inventory Cursor and Claude Code extensions on developer endpoints and remove anything not explicitly approved, prioritizing unknown or untrusted publishers.
2. Inspect Windows and macOS developer machines for unauthorized SSH keys, startup items, launch agents or daemons, scheduled tasks, and other persistence artifacts.
3. Rotate SSH keys and credentials from affected systems, then review access logs for unexpected sessions or remote connections.
4. Restrict extension installation through policy and require security review before new add-ons are deployed.
5. Monitor outbound SSH and unusual network activity from developer workstations for signs of remote access.
6. Preserve evidence and escalate to the security team if persistence files or unauthorized access are found.
7. Review the device inventory for Windows and macOS developer systems and check for signs of device information collection.

Original source

Hackread