AI agents use save_notebook to create or update resources in JupyterMCP — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your JupyterMCP environment.
The tool modifies persistent state (the notebook file) but is reversible through version control or backup recovery. It does not delete data or trigger irreversible operations. The severity is high because an AI agent could save a notebook after executing malicious code (potentially inserted via sibling tools like insert_and_execute_cell or edit_cell_content), making the compromise persistent and harder to undo.
From the tool's definition The tool 'save_notebook' persists changes to notebook state. The description states it will 'Save the current Jupyter notebook', which modifies the filesystem by writing notebook data.
Documented attack patterns abuse exactly the kind of access save_notebook gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and JupyterMCP, and nothing reaches the server without passing your rules. This is the rule we recommend for save_notebook:
{
"version": "1",
"default": "deny",
"tools": {
"save_notebook": {
"limits": [
{
"counter": "save_notebook_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}save_notebook 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.
Save the current Jupyter notebook. It is categorised as a Write tool in the JupyterMCP MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Jupyter MCP server in PolicyLayer and add a rule for save_notebook: 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 JupyterMCP. Nothing to install.
save_notebook 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 save_notebook 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 save_notebook. 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.
save_notebook is provided by the Jupyter MCP server (jjsantos01/jupyter-notebook-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from JupyterMCP, 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.
11 JupyterMCP tools catalogued and risk-classified — across an index of 43,000+ MCP servers.