npm AI-agent policy reports, with evidence.

This is concrete unconsented postinstall mutation of a foreign AI-agent control surface. Safeguards reduce accidental corruption but do not remove the automatic host-code modification.

The automatic global extension installation and subsequent prompt augmentation form a concrete AI-agent control-surface attack. No data exfiltration was required to establish the maliciou...

The package performs unconsented postinstall mutation of consumer AI-agent configuration, including an enabled remote MCP channel and broad permissions. This meets the install-control-sur...

The automatic lifecycle chain downloads or executes an opaque native payload and installs AI-agent hooks without user action. The payload's embedded bypass command makes the resulting age...

The package automatically changes generic consumer AI-agent skill directories during npm installation and can populate them with remotely supplied content. Its safeguards do not establish...

This is a concrete automatic postinstall mutation of a foreign AI-agent control surface. Guarded anchors reduce accidental breakage but do not provide user consent for the host rewrite.

This is a concrete automatic postinstall mutation of a foreign, global AI-agent control surface. The behavior meets the install-control-surface blocking rule regardless of the absence of...

This is an unconsented install-time mutation of a foreign AI-agent control surface. The automatic hook registration meets the blocking policy even though the payload binaries are not incl...

The package has a concrete automatic install-time chain that alters two foreign global agent configurations and installs instructions controlling later agent actions. This meets the insta...

The automatic postinstall chain mutates broad consumer AI-agent control surfaces and package-manager safeguards. Reviewer-directed suppression of secret reporting further establishes mali...

The lifecycle hook performs persistent, cross-agent configuration changes and enables a command-executing MCP plugin without an explicit user command. This meets the install-hook abuse bl...

The automatic postinstall hook rewrites a foreign AI runtime in the consumer dependency tree. This is a concrete install-hook abuse even though the inspected code does not exfiltrate data.

The package has a concrete automatic lifecycle chain that writes and prunes skills in two unrelated agent configuration roots, then modifies the user Python environment. This meets the po...

This is concrete unconsented postinstall mutation of a consumer project’s broad AI-agent control surfaces, coupled with review evasion and instructions to expose credentials. No automatic...

The package performs unconsented postinstall installation of executable plugins into a separate Herdr control surface and activates a background clipboard bridge. This meets the install-h...

This is unconsented postinstall mutation of a broad foreign AI-agent control surface. The absence of observed exfiltration does not neutralize the automatic cross-agent instruction deploy...

This is malicious under the install-control-surface policy because a postinstall hook automatically rewrites existing global Cursor and Codex agent configuration. The additional consumer-...

This is an automatic postinstall invocation of an opaque native binary to install AI-agent hooks, combined with unauthenticated executable download. It meets the install-control-surface b...

This is an unconsented postinstall mutation of foreign, broad AI-agent control surfaces. It meets the install-control-surface blocking policy despite using a local stdio server.

This is an unconsented postinstall mutation of multiple foreign AI-agent control surfaces, which meets the blocking policy. The additional install-time authenticated request increases the...

The automatic install-time network fetch and forced replacement of shared executables are concrete, while the runtime embeds directives that control other AI agents' authentication and co...