npm malware reports, with evidence.

This is unconsented postinstall mutation of broad, foreign AI-agent control surfaces combined with persistent background execution. The package's stated purpose does not remove the instal...

The package has a concrete automatic install path that broadly modifies consumer AI-agent configuration and explicitly suppresses review of its generated files. This meets the install-hoo...

This is a lifecycle bootstrapper for a remote unsigned native payload that deliberately bypasses operating-system download protections and persists it in user launch surfaces. The behavio...

The automatic postinstall globally changes the consumer environment by installing a separate AI agent, rather than merely preparing this package. Combined with default auto-approved agent...

The automatic global installation of a separate AI-agent CLI and persistent user configuration mutation are concrete install-hook abuse. No secret exfiltration was found, but the unconsen...

The package automatically harvests environment data during installation and uses an obfuscated network command. Its runtime self-dependency selects a different release, making this a mali...

The package has a concrete postinstall hook that rewrites a foreign AI SDK OAuth callback page. This meets the install-control-surface block boundary despite the absence of shown secret t...

This is a concrete install-hook supply-chain mechanism, not a normal transparent setup step. It remotely replaces executable package code and package metadata during installation.

This is a covert proctoring-evasion and AI-assisted capture tool with persistence and a bundled credential harvester. Its harmless postinstall hook does not mitigate the malicious runtime...

OpenSSF Malicious Packages via OSV confirms acme-diagnostics@1.0.1 as malicious (MAL-2025-6396): Malicious code in acme-diagnostics (npm)

OpenSSF Malicious Packages via OSV confirms acme-diagnostics@1.0.0 as malicious (MAL-2025-6396): Malicious code in acme-diagnostics (npm)

This is an unconsented install-time chain that mutates broad third-party AI-agent control surfaces and host configuration while executing a remotely obtained binary. The combination meets...

The automatic postinstall performs non-interactive consumer-project mutation and persists itself by rewriting the consumer lifecycle script. The same apply path alters multiple Codex cont...

This package uses an automatic lifecycle hook to install external code globally and stage opaque remote executables for later execution without integrity verification. That is concrete in...

The package performs unconsented postinstall mutation of several external AI-agent control surfaces and bootstraps an unchecked executable. This is a concrete install-hook abuse chain, no...

The automatic lifecycle hook performs broad, persistent user-environment mutation and installs a separate AI-agent CLI without an explicit user command. This is concrete install-hook abus...

The package automatically installs an opaque native payload from a fixed IP with TLS verification disabled and makes it executable. Its postinstall also wires hidden agent-related assets...

This package contains an install-time path that can mutate a remote database's authorization state and consumer project files. The opt-in environment gate reduces accidental activation bu...

The automatic lifecycle chain downloads and executes a remote script by default, which is unconsented install-time remote code execution. This is concrete install-hook abuse, not merely a...

This is an unconsented postinstall mutation of a foreign, globally installed AI-agent control surface. The conditional check and opt-out do not make the automatic lifecycle mutation user-...

The package performs automatic postinstall mutation of foreign AI-agent configuration and starts third-party MCP tooling, including a path that forwards credentials. This meets the instal...