npm security reports, with evidence.

The postinstall hook automatically modifies a foreign global AI-agent configuration and enables export of sensitive interaction telemetry. This is concrete install-hook abuse under the st...

The automatic lifecycle hook performs a concrete source rewrite of a foreign AI-agent dependency. Under the install-control-surface policy, that unconsented postinstall mutation is block-...

The package performs unconsented postinstall mutation of a consumer project's AI-agent instruction surfaces and reinforces it by enabling lifecycle execution. Reviewer-directed suppressio...

The automatic postinstall hook mutates both Codex and Claude user skill directories and can recursively delete installed content. This concrete install-time control-surface modification m...

The automatic postinstall hook performs broad, foreign AI-agent configuration mutation and installs an additional global package. This is a concrete install-time control-surface attack un...

The automatic lifecycle chain installs and enables package-controlled AI-agent hooks outside the package and consumer project configuration. This meets the blocking policy for unconsented...

This is an automatic lifecycle remote-payload chain with no integrity verification and disabled TLS validation. It creates a concrete route for arbitrary code execution through a remotely...

The automatic install hook broadly mutates a consumer project's AI-agent configuration and package-manager behavior, then installs reviewer-suppressing rules. This is concrete install-hoo...

The package uses postinstall to persist unrestricted Claude Code command hooks in a user-level configuration file. This meets the install-hook abuse blocking boundary.

The automatic postinstall hook performs broad global AI-agent skill installation and deletion rather than limiting setup to an explicit user action. This meets the install-control-surface...

The automatic global postinstall chain delegates broad AI-client configuration writes to a remotely downloaded binary. This is concrete install-time AI-agent control-surface mutation and...

The package contains a preinstall hook that executes a shell-based external request. This is concrete install-hook abuse with no package functionality supporting it.

The sole npm entrypoint is an obfuscated remote challenge-and-redirect page, not package functionality. Its concealed request and automatic redirect constitute concrete malicious behavior.

The package's only entrypoint is a heavily obfuscated challenge-and-redirect page with a concealed remote host resolution flow. Its random package identity and lack of legitimate library...

The package contains a concealed browser-redirect mechanism behind a challenge-themed page, which is concrete deceptive behavior. Although it has no install hook, the active entrypoint is...

This is concrete, automatic install-time credential and system-data exfiltration to an unrelated Telegram endpoint. The package should be blocked.

The package implements an automatic install-time reconnaissance and external callback chain. Its research labels do not establish consent for this behavior.

The package performs automatic install-time host and environment reconnaissance followed by external transmission. Its research labeling does not establish consent for this behavior.

Automatic lifecycle execution performs unsolicited host and project reconnaissance and transmits it externally. The package's research claim does not remove this concrete install-time exf...

Automatic install-time host reconnaissance and covert transmission to hard-coded external endpoints are concrete malicious behavior. Error suppression and a DNS fallback reinforce the int...

This is concrete, unconsented install-time data exfiltration with redundant network channels. The package's research label does not neutralize the automatic behavior.