AI agents call get_cell_outputs to retrieve information from Marimo without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves pre-computed outputs and console streams from notebook cells without executing code, modifying state, or affecting the notebook. It is a pure query operation analogous to reading logs or fetching results. No side effects or destructive capabilities are present.
From the tool's definition Tool name 'get_cell_outputs' and description 'Get visual output and console streams for cells' indicate data retrieval with no modification or execution of code. The word 'Get' and 'for cells' (querying existing outputs) confirm read-only behavior.
Attacks that exploit this kind of access
Get visual output and console streams for cells. It is categorised as a Read tool in the Marimo MCP Server, which means it retrieves data without modifying state.
Register the Marimo MCP server in PolicyLayer and add a rule for get_cell_outputs: 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.
get_cell_outputs is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the get_cell_outputs 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 get_cell_outputs. 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.
get_cell_outputs 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 →