This package has no npm install hook, but its manual installer writes persistent rules into global AI assistant configuration, and its skill directs agents to perform actions without asking the operator. That warrants a warning for AI-agent capability abuse; the reviewed source does not establish malicious execution or data theft.
Static reason
One or more suspicious static signals were detected.
Trigger
A user runs scripts/install.sh; the instruction files are written when the corresponding assistant configuration directories already exist. The skill can influence an assistant when loaded.
Impact
Assistant behavior may be changed persistently, including instructions to perform workflow actions without asking the user. The reviewed source does not show a concrete malicious action.
Mechanism
The installer redirects generated content into assistant rules or instruction files. The skill labels intensive audits pre-authorized and directs agents to run them without prompting for operator input.
Evidence
scripts/install.shSKILL.md$HOME/.claude/rules/vine.md$HOME/.config/opencode/instructions/vine.md$HOME/.gemini/antigravity/rules/vine.md