The package provides explicit or prompted setup for its own AI client integrations, which is a guarded capability that warrants a warning. Its postinstall hook only prints instructions, and the inspected source does not show a concrete attack.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review
Trigger
A user runs the plugin add or update command, or accepts the plugin installation prompt during interactive setup.
Impact
The clients gain gmax tools and instructions. The inspected code does not establish credential theft, broad configuration takeover, or remote code execution.
Mechanism
The commands register the gmax MCP integration and write its marked instructions to Codex AGENTS.md, or install the gmax plugin through the Claude CLI.
Evidence
dist/commands/setup.jsdist/commands/plugin.jsdist/commands/codex.jsdist/commands/claude-code.jsCodex AGENTS.mdClaude Code plugin configuration