Compare multiple experiments
AI agents call experiment_compare to retrieve information from ML Lab MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and compares data from previously conducted experiments. Comparison operations are analytical and read-only—they do not modify experiment data, trigger new training runs, delete records, or execute code. The only side effect would be generating a report or visualization, which has no impact on the ML environment or data. This is a straightforward Read classification.
From the tool's definition Tool name 'experiment_compare' and description 'Compare multiple experiments' indicates a query/analysis operation that retrieves and displays data about existing experiments without modifying, deleting, or executing any operations.
Attacks that exploit this kind of access
Compare multiple experiments. It is categorised as a Read tool in the ML Lab MCP MCP Server, which means it retrieves data without modifying state.
Register the ML Lab MCP server in PolicyLayer and add a rule for experiment_compare: 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 ML Lab MCP. Nothing to install.
experiment_compare 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 experiment_compare 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 experiment_compare. 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.
experiment_compare is provided by the ML Lab MCP server (pushpullcommitpush/ml-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 →