npm detections from LPM Firewall.

This is concrete install-hook abuse rather than ordinary platform setup: it automatically alters another AI-agent package's credential, authentication, and sandbox controls. No exfiltrati...

This is a concrete automatic install-time mutation of upstream AI-agent authentication and control code. The later launcher further weakens user agent permissions, so the package meets th...

The package automatically exfiltrates detected AI coding-agent session metadata during normal CLI use. The lack of an install hook reduces installation risk but does not neutralize the ru...

The package has a concrete automatic postinstall chain that broadly mutates foreign AI-agent control surfaces, including when unconfigured. This meets the install-hook abuse blocking boun...

The automatic lifecycle hook performs concrete, persistent modifications to a consumer repository's AI-agent control surface and development configuration. Its scope checks do not establi...

The automatic lifecycle hook directly mutates existing third-party agent configuration and installs a package-controlled launcher. This is concrete unconsented install-hook abuse under th...

The global postinstall path automatically mutates Claude Code configuration and permissions. This meets the install-control-surface blocking policy despite safeguards against replacing an...

This is an unconsented postinstall mutation of both consumer-project and user-wide AI-agent control surfaces. The source establishes the behavior directly, so it warrants blocking despite...

The automatic postinstall hook mutates both Codex and Claude user skill directories, a broad foreign AI-agent control surface. This meets the install-hook abuse blocking policy even thoug...

This package automatically weakens multiple security controls in a foreign AI-agent dependency tree and deletes its sandbox components. That concrete install-time control-surface mutation...

This is an unconsented install-time mutation of foreign project and AI-agent control surfaces, with durable execution through host postinstall and Git configuration. The lack of observed...

The automatic postinstall chain modifies the consumer project and configures multiple AI-agent platforms without an explicit user command. This meets the install-control-surface block cri...

This is unconsented postinstall execution of an opaque remote binary that automatically modifies a broad AI-agent control surface on global installs. The checksum does not establish trust...

The package uses automatic postinstall execution to change a consumer's package lifecycle and install or remove AI-agent plugins. Those broad, persistent mutations meet the install-hook a...

This package combines concealed third-party credential use, automatic external routing of agent data, broad network exposure, and disabled Codex safety controls. The behavior creates a co...

The postinstall hook automatically weakens authentication and credential protection in foreign AI-agent packages while removing sandbox dependencies. This meets the policy for malicious i...

This is a concrete automatic lifecycle attack path, not an explicit CLI setup action. Although no exfiltration was found in the inspected install path, the broad cross-platform agent-cont...

This is not an automatic lifecycle attack, but the explicit installer creates persistent Claude Code hooks from remotely fetched, obfuscated code. The remote payload execution path warran...

The package's automatic postinstall downloads and executes a foreign installer, then changes the installed tool's configuration. This is concrete install-hook abuse, not necessary local b...

The runtime entrypoint contains concrete, delayed, targeted browser disruption and a remote audio payload unrelated to a component library. This establishes malicious protestware behavior...

The package creates a concrete automatic chain from npm postinstall to an externally fetched executable that installs agent/editor hooks. This meets the install-control-surface blocking p...