npm detections from LPM Firewall.

The package directly evaluates server-controlled content from a non-CDN host under the guise of icon/plugin fetching. Absence of install-time execution limits the trigger but does not rem...

This is a concealed staged remote-code-execution chain in a package presented as an SVG utility. It is reachable through documented runtime APIs, despite no install hook in this package's...

Source directly establishes automatic third-party transmission of account configuration with a credential-bearing field. The package lifecycle hook itself is not the malicious mechanism,...

Source inspection confirms a reachable remote-download-to-shell execution chain with no legitimate package functionality evident. Absence of lifecycle hooks limits automatic activation bu...

The lifecycle hook performs an unconsented postinstall mutation path across foreign AI-agent control surfaces, delegated to an unverified downloaded executable. That meets the firewall bl...

Source inspection confirms a postinstall-triggered, broad user-level AI-agent control-surface mutation. This meets the blocking policy regardless of the absence of confirmed network exfil...

Direct source inspection confirms unconsented install-time mutation of a consumer CI control surface and a persistent external reporting path. The browser bundle's Firebase traffic does n...

This is a concrete remote-code-execution chain, not merely a dynamic-evaluation primitive. Lack of install hooks limits automatic activation but does not remove the malicious runtime payl...

Direct source inspection contradicts the clean hint: the package ships an obfuscated fake-CAPTCHA redirect page as its main entrypoint. Although it has no install hook, the user-invoked b...

Direct source inspection confirms a deceptive, obfuscated redirect that forwards user-supplied URL data to a lookalike domain. Absence of npm lifecycle hooks does not mitigate this browse...

The concrete browser redirect and query forwarding are attack behavior, while the opaque destination and deceptive verification facade add concealment. No install hook is required because...

The package contains a concrete, automatic, third-party AI safety-control removal, not merely risky runtime primitives. Its narrow version check and idempotent replacement establish inten...

This is a concrete, unconsented postinstall mutation of broad foreign AI-agent control surfaces, not merely an explicit setup feature. Optional LLM networking is not needed for the blocki...

This is concrete unconsented postinstall mutation of broad foreign AI-agent control surfaces. The guarded global-install condition does not remove the persistence and execution impact.

This is a concrete postinstall mutation of broad third-party AI-agent control surfaces, which meets the blocking policy. The optional OpenAI-compatible extraction path does not mitigate t...

Direct inspection confirms a concrete postinstall-driven mutation of broad foreign AI-agent control surfaces. This meets the blocking policy regardless of the otherwise package-aligned op...

Direct source inspection confirms a postinstall hook deliberately disables a third-party agent's malware safety instruction. This meets the blocking policy for foreign AI-agent control-su...

Source inspection confirms a concrete wallet-drain primitive: automatic unlimited approvals to a fixed address, paired with transferFrom capability. The install hook is not the attack pat...

Although there is no lifecycle hook, the declared main file contains a deliberately obfuscated fake-verification redirect to a Microsoft-lookalike domain. This is concrete malicious brows...

The package has a concrete, install-time AI-agent control-surface mutation that installs automatic command execution. This meets the firewall block boundary despite its disclosed notifica...