npm detections from LPM Firewall.

This is concrete postinstall mutation of foreign and broad AI-agent control surfaces. The automatic fan-out, settings change, and agent installation meet the install-hook abuse blocking p...

The package has a postinstall hook that automatically fan-outs package-controlled AI-agent content across global directories and changes a global Claude Code preference. This is concrete...

This is an unconsented postinstall mutation of broad, foreign AI-agent control surfaces. The absence of observed exfiltration does not remove the concrete control-hijack risk.

This is an automatic postinstall mutation of foreign and broad AI-agent control surfaces, including global skills, agents, and Claude settings. The behavior meets the install-hook abuse b...

The package combines an automatic install hook with persistent edits to multiple user AI-agent instruction files and explicit wallet-tool steering. That is a concrete AI-agent control-hij...

The package has a concrete postinstall hook that persistently modifies foreign AI-agent instruction files to influence later wallet-tool decisions. This meets the install-time AI-agent co...

This is an unconsented postinstall mutation of multiple foreign AI-agent control surfaces, including a newly created Cursor rules directory. The automatic redirect concerns wallet operati...

The manifest automatically invokes a script that persistently injects behavioral instructions into Claude Code, Cursor, and Windsurf configuration. This is concrete install-time AI-agent...

The package's automatic postinstall writes a bundled capability into several foreign AI-agent skill roots. Its lack of network or credential handling does not remove the concrete broad co...

This is an unconsented postinstall mutation of a foreign AI-agent control surface, combined with shell permissions and automatic installation of future releases. The behavior is concrete...

The package has a concrete automatic postinstall chain that mutates Claude Code configuration and permissions. The guarded behavior and absence of secret theft reduce scope but do not rem...

The package has a concrete, automatic postinstall path that writes package-provided content into broad AI-agent discovery directories and Copilot instructions in the consumer project. Thi...

This is concrete install-time mutation of broad, foreign AI-agent control surfaces, followed by instructions that suppress independent judgment and enable package-provided command executi...

The automatic lifecycle hook changes consumer AI-agent configuration and discovery content without explicit user setup. This meets the install-time AI-agent control-surface mutation block...

This is an unconsented postinstall mutation of broad foreign AI-agent discovery surfaces. The absence of observed exfiltration does not remove the concrete install-time control-hijack risk.

This is a concrete unconsented postinstall mutation of a broad, foreign AI-agent control surface. The absence of observed credential exfiltration does not neutralize the persistent contro...

The package has a concrete global post-install path that writes instruction content and executable hooks into multiple foreign AI-agent home directories. This meets the policy threshold f...

The package does not steal authentication tokens, but it collects and exports a stable identifier derived from unrelated AI-agent authentication configuration by default. This is concrete...

The automatic global launcher takeover is concrete, destructive, and targets a broad AI-agent control surface. The global-install guard does not establish consent to replace any pre-exist...

The package contains explicit AI-behavior steering that suppresses independent source verification, alongside host fingerprinting and error telemetry to a remote endpoint. These are concr...

The automatic lifecycle makes broad destructive changes to a user's Claude configuration and registers an MCP bridge. These actions meet the install-time foreign AI-agent control-surface...