npm detections from LPM Firewall.

The package contains a concealed import-time dropper and a decoded credential-stealing payload with persistence and network exfiltration. This is concrete malicious behavior unrelated to...

This is a concrete, concealed import-time credential theft and persistence chain, not ABI functionality or a scanner-only signal.

This is a concrete hidden import-time malware chain with persistence and credential exfiltration, unrelated to the package's advertised Ethereum helper functions.

Source inspection confirms a concrete automatic exfiltration path for sensitive account metadata to a custom endpoint. The harmless workspace-only postinstall hook does not mitigate that...

Source inspection confirms a nonfunctional package whose only substantive content is an opaque archive of AI-agent configuration and logs. This is a concrete sensitive-data distribution p...

This is concrete import-time data exfiltration, not a monitoring utility. The absence of lifecycle hooks does not mitigate execution through the package main entrypoint.

This is concrete import-time data theft and public npm exfiltration, not a benign dynamic-code pattern. The absence of lifecycle hooks does not mitigate the malicious main-entry behavior.

Source inspection confirms a concrete remote-code-execution chain, not merely suspicious primitives. The lack of install hooks limits the trigger to package use but does not reduce the im...

The source confirms unconsented install-time mutation of the consuming project's `.claude` agent configuration. Although no exfiltration or remote execution was found, this meets the fire...

Direct source inspection confirms unconsented install-time host-data collection and external exfiltration. This is concrete malicious behavior, not a benign package capability.

Direct source inspection confirms a hard-coded external Telegram control/output channel combined with permission-bypassing prompt mode and powerful local tools. The absence of install-tim...

Confirmed import-time remote-code execution is concrete malicious behavior, not a path-module feature. No install hook is needed because normal package use activates the payload.

This is concrete, silent runtime credential/configuration exfiltration through a hard-coded third-party endpoint. The benign package-local postinstall hook does not mitigate that behavior.

Source inspection confirms automatic remote logging of sensitive mail-account configuration, including a full serialized IMAP object, to an unrelated logging endpoint. The benign workspac...

This is a concrete remote payload delivery mechanism disguised as a package entrypoint, not a benign library behavior. No install hook is needed for the attack to activate when consumers...

Source establishes a concrete remote payload execution chain, not merely suspicious APIs. The absence of install hooks does not mitigate execution once the exported middleware is used.

The package contains an executable payload specifically designed to archive agent data and covertly transmit it through npm publishing. Absence of a lifecycle hook limits automatic activa...

The package contains a concrete, immediately executing credential/workspace archive exfiltration script using an obfuscated publishing token. The lack of a manifest entrypoint limits auto...

Although no lifecycle hook or import path triggers it automatically, the package contains a concrete credential/configuration theft and npm-upload mechanism. The source establishes malici...

The source contains a complete, self-executing archive-and-publication chain targeting AI-agent data, with an obfuscated registry credential. Lack of lifecycle wiring reduces automatic ex...