npm AI-agent policy reports, with evidence.

The package contains a concrete, default-on postinstall chain that installs third-party AI-agent skills globally. The opt-out does not make this installation consented.

This is concrete unconsented postinstall mutation of a broad AI-agent control surface, meeting the blocking policy. Opt-out support does not establish prior user consent.

The install chain concretely rewrites Claude Code configuration from postinstall. That foreign AI-agent hook installation is sufficient for a block verdict under the stated policy.

The lifecycle chain is source-confirmed and mutates ~/.claude/settings.json to execute the package on multiple Claude events. This meets the firewall block policy for unconsented postinst...

The package performs a concrete unconsented postinstall deletion of existing AI credentials/settings selected by competitor markers, followed by automatic replacement credential provision...

The concrete postinstall path deletes existing agent credentials/settings and installs new remote-provisioned credentials without consent. The launcher and dormant hooks do not mitigate t...

The package's install-time behavior is not limited to its own state: it modifies user-level Claude hooks and establishes persistence without an explicit follow-up command. This meets the...

This is concrete install-time mutation of broad, foreign AI-agent control surfaces. It meets the blocking policy despite no observed credential exfiltration or remote payload download.

The concrete postinstall rewrite of a foreign AI-agent dependency is sufficient for blocking under the install-control-surface policy. The observed gateway credential use does not indepen...

This is concrete, unconsented postinstall mutation of foreign/broad AI-agent control surfaces. The lack of an install-time exfiltration request does not remove the control-hijack risk cre...

The package performs unconsented postinstall mutation of a foreign AI-agent control surface for rebranding. Runtime authentication behavior does not mitigate that install-time tampering.

This is concrete install-time mutation of a foreign AI-agent control surface, coupled with routing users to the package vendor's provider. The authentication network traffic is package-al...

The concrete postinstall mutation of a foreign AI-agent dependency meets the install-control-surface block boundary. Network and credential code is package-aligned, but does not mitigate...

This is a concrete, unconsented postinstall mutation of broad foreign AI-agent control surfaces, meeting the blocking policy. No separate exfiltration chain is required for this verdict.

The observable purpose is a keyboard notification utility, but its npm postinstall automatically injects executable hooks across third-party AI-agent configurations. That meets the firewa...

Direct inspection confirms the critical behavior is the guarded but automatic postinstall mutation of multiple foreign AI-agent control surfaces. The scanner's fixed-gateway credential-ex...

The confirmed postinstall mutation of ~/.claude/skills meets the policy threshold for an unconsented foreign/broad AI-agent control-surface write. The reverse-shell and credential-exfiltr...

The lifecycle behavior is concrete and reaches global Claude Code skill/configuration surfaces without an explicit user command. This meets the firewall block policy irrespective of the a...

The package performs an unconsented postinstall mutation of foreign AI-agent control surfaces and force-removes prior content first. No additional exfiltration was needed to meet the inst...

The package’s global postinstall directly installs persistent command hooks into Claude Code’s user configuration and launches background persistence. This meets the block threshold regar...

This meets the install-control-surface block criterion: unconsented postinstall mutation of broad foreign AI-agent configurations. Runtime provider networking appears feature-aligned and...