AI agents invoke run_doctor to trigger actions in NPM Helper MCP. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool runs arbitrary test commands and executes package installations, which are external operations with effects dependent on the project's test suite and installed packages. While not destructive on its own (upgrades are reversible via version rollback), the execution of tests and installations qualifies as Execute category.
From the tool's definition Tool description states 'Iteratively install upgrades and run tests', indicating execution of package installation commands and test suites. The tool name 'run_doctor' combined with 'install upgrades' demonstrates code execution capabilities.
Documented attack patterns abuse exactly the kind of access run_doctor gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and NPM Helper MCP, and nothing reaches the server without passing your rules. This is the rule we recommend for run_doctor:
{
"version": "1",
"default": "deny",
"tools": {
"run_doctor": {
"limits": [
{
"counter": "run_doctor_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}run_doctor stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Iteratively install upgrades and run tests. It is categorised as a Execute tool in the NPM Helper MCP MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the NPM Helper MCP server in PolicyLayer and add a rule for run_doctor: 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 NPM Helper MCP. Nothing to install.
run_doctor is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the run_doctor 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 run_doctor. 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.
run_doctor is provided by the NPM Helper MCP server (pinkpixel-dev/npm-helper-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from NPM Helper MCP, 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.
10 NPM Helper MCP tools catalogued and risk-classified — across an index of 43,000+ MCP servers.