Warn by default; block when configured. The install silently adds or refreshes agent instructions and server configuration, changing what Claude Code and Gemini sessions can load.
AI security review
scanned · by lpm-firewall-ai
The postinstall hook automatically writes bundled skills and registers an MCP server in user-level AI-agent configuration. This is install-time modification of agent instructions and server settings without user consent, so the package is malicious.
npm install, through the package’s postinstall hook.
Impact
The install silently adds or refreshes agent instructions and server configuration, changing what Claude Code and Gemini sessions can load.
Mechanism
The hook invokes the CLI to install bundled skills in the user’s home directory and register the bundled MCP server with Claude Code; Gemini skills and imports are also updated when its configuration directory exists.
Policy narrative
During npm installation, the postinstall hook automatically installs bundled skills into the user’s Claude configuration and, where present, Gemini configuration, then registers a bundled MCP server by writing Claude’s ~/.claude.json. These AI-agent configuration writes happen without an explicit user command, meeting the install-time agent control hijack rule.
AI identified this as a dangerous AI-agent capability at 98.0% confidence with low false-positive risk.
Evidence for policy risk
The postinstall hook automatically runs `node scripts/postinstall.mjs`.
The hook runs `dr skill install` for the user’s home directory, which writes bundled skills into `~/.claude/skills/` and, when present, `~/.gemini/skills/`; it also updates Gemini’s `GEMINI.md`.
The hook automatically registers the `dr-zendesk-kb` MCP server with Claude Code; the CLI documents that registration writes `~/.claude.json`.
Evidence against
The reviewed install behavior runs the package’s local CLI helper; no external network endpoint is identified in this install chain.
Avoid installing @datarails-ai/dr-cli@1.6.214. Remove it from direct dependencies and check your lockfile for transitive copies.
Choose an independently verified alternative or release. This report does not establish that other versions are safe.
If this version ran, check your AI agents' configuration and instruction files (the files this report lists) for changes you did not make, and remove them. This finding does not by itself establish malware intent.