npm detections from LPM Firewall.

The package has a concrete automatic postinstall chain that writes skills into two independent agent configuration roots and installs additional dependencies. This meets the policy thresh...

This is unconsented postinstall mutation of a foreign project that plants instructions intended to cause broad AI-agent control-surface writes. The source establishes the complete install...

This package includes a browser credential collector whose default destination is an external host, not its local CLI server. Although npm installation has no automatic install hook and c...

The package performs broad, destructive consumer-project mutation automatically at install time and persists it through an injected build plugin. Its reviewer-directed discussion of prior...

This is an active remote payload path into an AI agent, not merely a documented local integration. The absence of an npm lifecycle hook reduces install-time risk but does not remove the r...

The package contains an automatic, obfuscated remote-code loader in both primary entry points. Although no install hook exists and the URL suggests a formatter dependency, the import-time...

This is unconsented postinstall mutation of broad, foreign AI-agent control surfaces, including activation of network access and remote instruction loading. The behavior meets the install...

The automatic postinstall performs broad, persistent mutation of a consumer project's AI-agent and development control surfaces. Conditional targeting and the absence of observed exfiltra...

Source inspection confirms unconsented collection and external transmission of local identity and host-fingerprint data during normal package use. The absence of an install hook does not...

The hidden runtime import activates an obfuscated network interceptor and fingerprinting self-update daemon, beyond the documented API-client behavior. The absence of an install hook does...

This is an automatic install-time data-exfiltration payload with no package functionality beyond collection and transmission. The hardcoded external receiver and sensitive local-file read...

The automatic archive-and-upload path is concrete and can disclose unexcluded project secrets to a remote service. The lack of an install hook limits the trigger, but does not remove the...

This is concrete automatic postinstall mutation of foreign, user-wide AI-agent configuration, which the policy requires blocking. The opt-out variable and error handling do not establish...

The automatic postinstall path modifies project AI-agent permissions and removes its own npm evidence afterward. This is a concrete install-hook abuse chain even though no direct secret e...

This is unconsented postinstall mutation of a foreign AI-agent control surface, followed by self-healing detached execution and CDP page control. That concrete lifecycle chain meets the p...

The automatic lifecycle hook broadly mutates a consumer project, and the injected build plugin enforces those mutations by discarding later edits. This is concrete install-hook abuse desp...

The package combines active AI-verification suppression with unneeded remote host fingerprinting and error telemetry. These are concrete runtime behaviors rather than a benign static-anal...

The automatic lifecycle hook mutates both Codex and Claude skill directories and performs user-level dependency installation. This meets the install-control-surface policy for a publish b...

This is an unconsented install-time remote-code-execution chain that can broadly alter the host outside the reviewed package contents. The guards reduce accidental execution but do not ma...

This is a concrete postinstall mutation of a foreign project's AGENTS.md and agent instruction files. The guarded and non-destructive implementation does not remove the unconsented contro...

The automatic postinstall patch chain rewrites an external AI-agent package and establishes a per-user hook on future launches. This meets the install-hook abuse policy even though no sec...