npm detections from LPM Firewall.

The source establishes a concrete automatic lifecycle path that modifies foreign AI-agent settings and agent definitions, then persists it through automatic global self-updates. This meet...

This package implements concrete remote-triggered credential exfiltration and silent AI-agent trust mutation. Its benign postinstall banner does not mitigate the runtime collection path.

The package performs remote payload download and execution automatically from postinstall, then installs global and agent-related tooling. The safe-mode and skip variables do not establis...

The package has a concrete automatic lifecycle path that writes third-party AI-agent configuration and skills. The global-install guard and loopback endpoint reduce scope but do not estab...

The automatic lifecycle hook establishes persistent background execution through the operating system service manager. This is an unconsented install-time persistence mechanism affecting...

The package gives a remote relay direct control over shell commands and a Bash script executed on the user's machine, while uploading recorded session data. This is concrete remote code e...

The package has an automatic postinstall chain that downloads package-controlled content and mutates a consumer AI-agent control surface. This is concrete install-hook abuse rather than a...

The package's postinstall hook performs automatic remote-code execution, global package installation, and remote skill acquisition. Those install-time actions exceed package setup and pre...

This is unconsented postinstall mutation of broad AI-agent control surfaces in the consumer project. No separate network or credential-theft behavior is required to establish the install-...

This package has an automatic install hook that performs destructive dependency mutation and attempts privileged system software installation. That is concrete install-hook abuse even tho...

The package automatically creates root-owned persistence during installation and enables a self-updating execution path. The local-install and cloud-link guards reduce scope but do not ne...

The package's only implemented behavior is collecting local identity data and posting it to a hard-coded external callback. The absence of an install hook limits the trigger but does not...

The source establishes automatic postinstall mutation, opaque remote binary delivery without cryptographic verification, global AI-package installation, and subsequent execution. These be...

The malicious behavior is runtime-triggered through the default API path and performs unconsented authenticated actions selected by a remote operator. The benign Node version check does n...

This is concealed, remotely directed mutation of a user's WhatsApp account during ordinary library use. The benign Node-version preinstall check does not mitigate the runtime account abuse.

The package's implemented behavior is an external transfer of local identity and host data to an unrelated hard-coded endpoint. Explicit CLI invocation does not make this covert fingerpri...

The automatic lifecycle hook changes foreign AI-agent instruction locations and installs an extension, while the injected content directs AI behavior. This is concrete install-hook abuse...

The automatic global-install hook changes foreign, broad Codex and Claude Code control surfaces, including a default agent definition. That meets the install-hook abuse policy despite the...

The automatic postinstall chain broadly modifies third-party AI-agent skill directories. This meets the install-control-surface blocking policy despite the absence of demonstrated network...

The source establishes automatic installation of a foreign Claude Code CLI plus persistent shell, registry, and launcher mutations. This meets the install-time foreign AI-agent control-su...

The package automatically downloads and executes an unverified native executable that installs hooks for external AI-agent tools. This satisfies the install-time AI-agent control-surface...