npm security reports, with evidence.

The package has a concrete automatic postinstall path that mutates multiple consumer-project AI-agent control surfaces. This meets the install-control-surface blocking rule despite the ab...

The automatic postinstall performs unconsented, multi-platform AI-agent control-surface mutation and can start a project migration. This meets the install-hook abuse blocking policy even...

This package contains concrete credential transfer and unsigned remote-script execution paths, not merely inert tooling. The npm lifecycle hook itself is not the primary issue, but the ru...

This package has a concrete automatic install-time chain that modifies a foreign agent's authentication controls. Its launcher additionally defaults the agent to unrestricted permissions,...

This is malicious under the install-control-surface policy: postinstall modifies all detected coding-agent configurations without user consent while installing a remote executable payload...

The automatic lifecycle hook mutates a foreign AI-agent control surface and deploys publisher-controlled remote content into the consumer workspace. This is a concrete unconsented install...

This is an unconsented postinstall mutation of a foreign AI-agent control surface in the consumer workspace. Its ownership and conflict checks do not remove the risk from creating that su...

The package’s postinstall behavior materially differs from its stated validation purpose and automatically launches a persistent service. Its local-only default does not make the unconsen...

The package contains a complete automatic remote-code acquisition and dynamic-execution path on normal CLI use. Integrity validation does not mitigate the unpinned trust in mutable regist...

The only package source defines an automatic preinstall hook that establishes a remote interactive shell and performs external data posting. This is concrete install-time malware behavior.

The package has a concrete automatic postinstall chain that changes user-wide Claude settings to capture and export session content. This meets the install-hook abuse policy for a malicio...

The sole install-time script is an explicit reverse-shell command with external output forwarding. This is concrete malicious behavior activated without user interaction beyond installation.

The install-time hook contains concrete remote-shell behavior, not a package-aligned setup action. This is malicious install-hook abuse.

The sole lifecycle script implements an install-time reverse shell and sends its output externally. This is concrete malicious behavior, not package setup.

This is unconsented postinstall mutation of a consumer project's AI-agent control surface and broader repository workflow. The guards are scoped to detected Suppa projects, but that does...

The package contains a concrete credential-redirection path to a fixed unofficial gateway and persists the credential. Lack of an install hook limits the trigger but does not remove the r...

The package performs automatic postinstall rewrites of foreign agent authentication and permission controls, then activates unrestricted operation through its launcher. This meets the ins...

The package automatically installs package-authored agent instructions into two unrelated default agent homes and modifies them on every install. This meets the install-time foreign/broad...

The package performs broad user-level AI-agent configuration and capability installation automatically from postinstall. This meets the install-control-surface blocking policy despite no...

This is an unconsented postinstall mutation of a broad foreign AI-agent control surface. The global command registration is persistent and applies beyond the installing project.

The automatic postinstall hook mutates Claude CLI configuration outside the package-owned directory. Under the install-control-surface policy, this is concrete malicious install-hook abus...