The package's explicit install configures a hook in the user's broader Codex settings, which can inject service-delivered chat messages into agent context. The hook has task checks and labels messages untrusted, but the settings change makes this a warning-level agent capability risk.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
User runs the package's explicit install command; afterward, Codex invokes the hook on PostToolUse events for an active task.
Impact
This changes broader agent settings and introduces externally supplied content into the agent context, although the hook marks that content untrusted.
Mechanism
The hook reads task-addressed chat messages using the Entrorove account credential and injects them into the agent's context as additionalContext.
Evidence
cli.mjsagent-chat-hook-install.mjsagent-chat-mention-hook.mjs~/.codex/hooks.json~/.codex/entrorove/chat-hook-installation.json