Calls an external MCP tool once with caller-supplied sample arguments, captures the top-level keys of its response into the persistent tool-output-schema cache, and returns the captured shape. Use this when the plan synthesizer encounters an external tool whose upstream server doesn
AI agents call probe-tool-output to retrieve information from Mmc without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool executes a single call to an external tool to probe its response structure and caches metadata about that response. While it does invoke an external tool (which could have side effects depending on what tool is probed), the probe-tool-output itself only reads and catalogs the output shape.
From the tool's definition Tool 'probe-tool-output' calls an external MCP tool with supplied arguments and captures/returns the response shape without modifying data.
Attacks that exploit this kind of access
Calls an external MCP tool once with caller-supplied sample arguments, captures the top-level keys of its response into the persistent tool-output-schema cache, and returns the captured shape. Use this when the plan synthesizer encounters an external tool whose upstream server doesn. It is categorised as a Read tool in the Mmc MCP Server, which means it retrieves data without modifying state.
Register the Mmc MCP server in PolicyLayer and add a rule for probe-tool-output: 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 Mmc. Nothing to install.
probe-tool-output 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 probe-tool-output 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 probe-tool-output. 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.
probe-tool-output is provided by the Mmc MCP server (modelmycontext/mmc-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
probe-tool-output is one line of Mmc's registry record.
The record carries the whole server: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.
Teams ship this data inside their own products. See what a licence covers →