Compromised MemTensor Packages Deliver sckit Credential Stealer via npm and PyPI

Threat actors have successfully compromised two legitimate MemTensor packages, `@memtensor/memos-cloud-openclaw-plugin` on npm and `MemoryOS` on PyPI, to distribute a sophisticated credential-stealing malware named `sckit`. The malicious versions, specifically npm versions 0.1.21, 0.1.23, 0.1.25 and PyPI version 2.0.34, inject a platform-specific Go-based implant. This supply chain attack targets developers relying on these AI memory integration libraries.

The `sckit` implant is a cross-platform Go binary designed for Windows, Linux, and macOS, capable of harvesting a wide array of sensitive data. In npm, the payload activates upon agent gateway startup or during memory-recall events, passing host environment and user prompt text. For PyPI, the malware executes immediately when the `memos` module is imported into an application. The stolen credentials, including AWS keys, GitHub/GitLab tokens, npm/PyPI tokens, and various API keys, are exfiltrated to `skyleen[.]fr`. Attackers reportedly gained access by exploiting MemTensor's GitHub Actions release pipelines to obtain publish tokens.

This incident directly impacts the OpenClaw ecosystem, particularly developers leveraging `MemTensor` for AI memory integration within agentic systems. The compromise of `@memtensor/memos-cloud-openclaw-plugin`, explicitly tied to OpenClaw, means that agents utilizing these versions could inadvertently expose critical operational credentials. The malware's activation during "agent gateway starts" and "memory-recall events" highlights a direct threat vector into the operational integrity and data handling of AI agents. Furthermore, the broad scope of credential theft, including source-code platforms and cloud services, jeopardizes the entire development and deployment pipeline for OpenClaw agents.

This is a high-strength signal for developers, researchers, and operators within the OpenClaw ecosystem and broader AI agent development space. Developers must immediately audit their dependencies for affected `MemTensor` versions and implement robust software supply chain security practices