The install hook automatically writes this package's bundled skill into user-level agent directories when its lifecycle conditions match. That is a meaningful agent-extension lifecycle risk, but the inspected behavior is limited to the package's own skill, so it does not support a malicious verdict.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
Supported global npm/pnpm installation or a qualifying pnpm add postinstall.
Impact
The package changes user-level agent skill configuration without a separate explicit skill-install command; the behavior is limited to its own bundled skill.
Mechanism
The hook loads the bundled skill and synchronizes it under ~/.agents/skills/gd-cli and into detected agent-host skill directories, including ~/.codex/skills/gd-cli and ~/.claude/skills/gd-cli.
Evidence
package.jsondist/bin/postinstall.jsdist/src/features/skill/installer.js~/.agents/skills/gd-cli~/.claude/skills/gd-cli~/.codex/skills/gd-cli~/.cursor/skills/gd-cli~/.config/opencode/skills/gd-cli~/.openclaw/skills/gd-cli