npm security reports, with evidence.

The package has a concrete automatic postinstall chain that mutates a broad AI-agent control surface and registers persistent session-start execution. This meets the install-hook abuse bl...

The automatic postinstall hook writes a foreign, user-wide AI-agent control surface and configures future command execution. Although no exfiltration or network behavior was found, this i...

The automatic postinstall execution of an unverified native payload to alter Codex and Claude Code integrations is a concrete install-time AI-agent control-surface attack. The absence of...

The postinstall hook automatically mutates a foreign global AI-agent environment by installing opencode-ai. This is an unconsented install-time AI control-surface change, so it meets the...

This is a concrete automatic install-time persistence chain combined with silent system dependency installation. Its Windows-only scope and local watchdog target do not make the unconsent...

This package performs unconsented postinstall modification of multiple foreign AI-agent control surfaces and installs an unpinned remote MCP command. That is concrete install-hook abuse u...

This is concrete remote code execution activated by a normal import, with detached-process persistence. The absence of an install lifecycle hook does not mitigate the import-time payload...

The package establishes an import-triggered detached process that fetches and executes remotely supplied code. This is concrete remote code execution and persistence-like behavior, not a...

The complete remote-code execution path is present in the published source and activates on ordinary import. Detached execution and PID tracking add persistence behavior.

The package implements an import-triggered detached remote-code loader, not a bounded package feature. The absence of an npm lifecycle hook does not mitigate runtime arbitrary code execut...

This is a concrete postinstall chain that installs a native payload and uses it to alter third-party AI-agent hooks. The behavior meets the install-hook abuse block policy despite package...

This is an unconsented postinstall mutation of a broad AI-agent control surface, with automatic third-party command setup and credential forwarding. The global-install gate does not estab...

The package contains deliberate, top-level, targeted browser disruption in both published entrypoints. Although it has no install hook, its import-time protestware behavior is concrete an...

This is a concrete automatic install-time remote code delivery path with certificate verification disabled and no payload integrity check. The postinstall also mutates hidden asset links,...

This package performs concrete, unconsented postinstall persistence and silent software installation on Windows. The behavior is automatic rather than an explicit user setup command.

Static inspection confirms an automatic lifecycle path that modifies global Codex trust and hook configuration based on the presence of Codex state. This creates persistent package-contro...

The package has a concrete automatic lifecycle chain that deletes files and installs remotely obtained skills across multiple AI-agent directories. This meets the install-hook abuse block...

The automatic postinstall hook performs broad persistent mutations and installs a third-party AI-agent executable without an explicit user command. This meets the install-time foreign AI-...

The package contains an automatic lifecycle hook that reads and transmits a credential from the consumer environment. Although framed as license validation, this is a concrete install-tim...

The package has a concrete automatic credential-transmission path in preinstall. The limited request payload does not remove the risk of exporting an environment secret during dependency...

The package has a concrete automatic install-time credential transmission path. Static icon exports do not justify sending an environment secret during installation.