Investigate malicious npm packages and AI-agent policy findings. Each report identifies the package version, supporting evidence, advisory sources, and current Firewall decision. Malware findings and policy violations are labelled separately.
Updated 12 Oct 2026, 14:29 UTC. Ordered by latest scan.
This is concrete persistent collection and exfiltration of sensitive local AI-agent transcript content to a package-controlled endpoint, activated by normal CLI use without a separate opt...
The package transmits authentication credentials and optional 2FA material to a hard-coded non-Facebook endpoint during normal recovery behavior. The absence of lifecycle hooks does not m...
Direct source shows a default external receiver for supplied account credentials in normal runtime login flows. No install hook is needed for this concrete credential-exfiltration behavior.
Source establishes a default-enabled recovery path that forwards sensitive Facebook credentials and an optional 2FA secret to an unrelated hardcoded host. This is concrete credential exfi...