Delete a cell from a running notebook.
AI agents call delete_cell to permanently remove resources in Marimo — typically in cleanup and lifecycle workflows. It does its job in a single call, and there is no undo.
Deleting a cell from a notebook irreversibly removes code and its associated state. This is a destructive action that cannot be undone through normal tool use and will impact the notebook's execution flow and saved state. While not as severe as a critical system-level deletion, it represents permanent data loss within the notebook context, placing it in the Destructive category rather than Write.
From the tool's definition Tool name is 'delete_cell' with description 'Delete a cell from a running notebook.' The verb 'delete' is explicit and irreversible—once removed, the cell content is gone unless manually recovered.
Attacks that exploit this kind of access
Delete a cell from a running notebook. It is categorised as a Destructive tool in the Marimo MCP Server, which means it can permanently delete or destroy data. Block by default and require explicit approval.
Register the Marimo MCP server in PolicyLayer and add a rule for delete_cell: 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 Marimo. Nothing to install.
delete_cell is a Destructive tool with critical risk. Critical-risk tools should be blocked by default and only enabled with explicit human approval.
Yes. Add a rate_limit block to the delete_cell 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 delete_cell. 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.
delete_cell is provided by the Marimo MCP server (vladisluv12/marimo-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →