npm detections from LPM Firewall.

Source inspection confirms a concrete, automatic install-time telemetry and IP-exfiltration chain, not merely suspicious primitives. The behavior is unnecessary for package installation a...

Source inspection confirms concrete install-time network collection and telemetry exfiltration of the installer host's public IP to a package-controlled Sentry project. No broader payload...

Source inspection confirms an unconsented postinstall chain that writes/stages user-level Codex integration and starts a persistent gateway. This meets the install-time foreign AI-agent c...

The package contains a concrete unconsented `postinstall` mutation of global AI-agent instruction directories. Under the install-control-surface policy, this warrants blocking even though...

The package has a concrete unconsented postinstall mutation of a foreign consumer project’s MCP configuration. No source-confirmed exfiltration, remote payload execution, or destructive b...

Source inspection confirms the critical behavior rather than merely scanner hints: npm postinstall injects a wildcard Claude Code hook into the consuming project. The hook captures tool-u...

Source inspection confirms the lifecycle hook injects a wildcard Claude Code hook into the consumer project, not merely first-party package configuration. The resulting handler processes...

Direct inspection confirms a concrete npm postinstall chain that modifies consumer-project MCP configuration and activates package code. Under the install-control-surface policy, this war...

Source inspection confirms a concrete postinstall AI-agent control-surface hijack and configurable data-upload behavior. The localhost default does not remove the risk because the hook is...

`index.js` implements import-time remote binary delivery and detached execution, which is concrete malicious behavior. The lack of an npm lifecycle hook does not mitigate execution when t...

The package performs concrete install-time writes to the user's Claude skill directory and executes follow-on tooling. That meets the firewall block boundary for unconsented postinstall m...

Source inspection confirms the lifecycle trigger and documents the resulting daemon and agent-plugin control changes. This meets the blocking policy for unconsented npm postinstall AI-age...

Direct source inspection confirms an automatic postinstall chain that writes AI-agent configuration into consumer projects. This meets the provided policy's blocking condition regardless...

Direct source inspection confirms an unconsented `postinstall` write to a consumer-owned Claude Code hook configuration with wildcard interception. Under the stated install-control-surfac...

Source inspection confirms a concrete postinstall chain that writes package and remotely fetched content into ~/.claude/skills. Under the firewall policy, this warrants blocking regardles...

The lifecycle hook concretely writes AI-agent rule files into the consumer project during `npm install`. The non-overwrite guard and lack of network activity do not remove the unconsented...

The package has no install hook, but its ordinary runtime API launches a detached downloader that evaluates server-controlled code. This is concrete malicious behavior, not a benign use o...

Direct source inspection confirms automatic remote code loading and a browser proxy/service-worker stack that handles routed requests and cookies. There is no npm lifecycle execution, but...

Although npm installation itself has no lifecycle hook, the declared `index.html` entrypoint immediately imports remote code and ships a service-worker proxy to an external transport. The...

This is a browser application disguised as an npm package, with automatic remote script execution and proxy service-worker setup. Although no install hook or host-side harvesting was foun...

No npm lifecycle execution exists, but the declared HTML entrypoint contains concrete runtime remote-code loading and a traffic-intercepting proxy tunnel. The package's generic republish...