npm security reports, with evidence.

This is a concrete default-enabled exfiltration path to package-configured remote receivers, not merely a generic HTTP client. It is runtime-only rather than install-time, but it can disc...

The package performs unconsented, install-time transmission of host and installation-context data to an unrelated webhook endpoint. Its normal runtime entrypoint is benign, but the lifecy...

The package’s preinstall hook performs a concrete, automatic outbound collection of host and installation metadata to a third-party webhook. The normal runtime entrypoint does not mitigat...

This is concrete install-time mutation of a foreign AI-agent configuration, meeting the firewall block boundary despite its limited intended UI effect.

Direct source inspection confirms an automatically triggered reverse shell with a hard-coded endpoint and no package-aligned purpose. This is concrete install-time remote code execution.

This is an unconsented postinstall mutation of the foreign Cursor AI-agent control surface. The global-install guard narrows exposure but does not provide user consent.

Source inspection confirms configured external logging of rich application records to embedded webhook receivers. The absence of install hooks reduces scope but does not remove the concre...

Direct source inspection confirms an automatically triggered reverse shell with multiple fallbacks, unrelated to the advertised date formatter.

This is an unconsented postinstall mutation of a broad AI-agent control surface, not merely a package-local setup. The guarded detection does not obtain consent, and the same path also en...

Source inspection confirms an active, hard-coded external logging path that serializes runtime records and is enabled by default for ScanOrder. No install-time behavior was found, but the...

This is concrete, unconsented install-time remote command execution, not functionality required by a date formatter. The benign exported helper does not mitigate the malicious lifecycle p...

The package’s postinstall path performs unconsented cross-vendor AI-agent configuration and persistence, which meets the blocking policy. The absence of a fixed exfiltration host does not...

The package has a concrete install-time AI-agent control-surface write into the consuming project. No separate network or credential-theft behavior was found, but the unconsented postinst...

Direct source inspection confirms an install-time HTTPS exfiltration chain. The package has no apparent functional implementation beyond collecting and transmitting host data.

Source confirms a hardcoded non-HTTPS IP default receiving bearer-authenticated activity uploads. The absence of lifecycle scripts limits propagation but does not remove the concrete cred...

This is a concrete, unconsented postinstall mutation of broad foreign AI-agent control surfaces, which meets the blocking policy. No network exfiltration is needed for the control-hijack...

This package contains an automatic install-time telemetry and encrypted remote-payload execution chain. Its benign runtime API does not justify that behavior.

Source confirms a default-enabled remote reporting path for data derived from Claude activity, rather than merely local dashboard telemetry. This is concrete unconsented data exfiltration...

The package contains concrete obfuscated, runtime network-to-eval-and-spawn behavior unrelated to its declared PostCSS purpose. No lifecycle hook is needed because the payload executes wh...

The package executes a remotely fetched binary during postinstall, with release selection and checksum material both controlled by the same external source. This establishes a concrete in...