AI agents use install_packages to create or update resources in Npmplus — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Npmplus environment.
Installing packages modifies package.json, package-lock.json, and node_modules—reversible changes to project state. However, severity is high because installed packages can execute arbitrary code during installation (lifecycle scripts like postinstall), introducing supply-chain attack risk. An AI agent with this tool could install malicious packages or unexpected versions without proper validation.
From the tool's definition Tool name 'install_packages' and description 'Install npm packages' indicate package installation, which modifies the project's dependencies and filesystem.
Documented attack patterns abuse exactly the kind of access install_packages gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Npmplus, and nothing reaches the server without passing your rules. This is the rule we recommend for install_packages:
{
"version": "1",
"default": "deny",
"tools": {
"install_packages": {
"limits": [
{
"counter": "install_packages_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}install_packages stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Install npm packages. It is categorised as a Write tool in the Npmplus MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Npmplus MCP server in PolicyLayer and add a rule for install_packages: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Npmplus. Nothing to install.
install_packages is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the install_packages rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for install_packages. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
install_packages is provided by the Npmplus MCP server (shacharsol/js-package-manager-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Npmplus, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
16 Npmplus tools catalogued and risk-classified — across an index of 43,000+ MCP servers.