The automatic postinstall hook invokes agent setup, and the README confirms that it writes hooks, a Skill, and a Codex setting. Under the review rules, install-time changes to AI-agent configuration outside the package directory without consent are malicious.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review; source fingerprint signature matched known malicious package; routed for review
Trigger
Automatically during npm postinstall; the README also says setup is retried when the service starts.
Impact
Unrequested install-time changes to AI-agent hooks, skills, and settings can alter how those agents behave.
Mechanism
The installer invokes the platform binary with --install-agent-hooks. The README identifies the changes as installing agent hooks and a Skill, and changing Codex's startup update setting.
Policy narrative
During npm postinstall, the package copies or downloads its platform binary and runs it with --install-agent-hooks. The README says this setup installs Codex and Claude Code hooks and a Skill, changes Codex's startup update setting, and is retried on service start. These install-time changes affect agent configuration outside the package directory without a separate user command.
Evidence
package.jsonscripts/install.jsREADME.md