Verify that the AI correctly understood the user
AI agents call verify_understanding to retrieve information from MCP-Confirm without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool is a clarification mechanism that reads or checks understanding state. It has no side effects on data, does not execute external commands, does not modify state irreversibly, and does not involve financial transactions. It fits the Read category as it queries or confirms information. Severity is low because even misuse would only affect conversation flow, not cause harm.
From the tool's definition Tool name 'verify_understanding' and description 'Verify that the AI correctly understood the user' indicate a query/check operation that retrieves or verifies information about user intent without modifying, executing, or deleting anything.
Attacks that exploit this kind of access
Verify that the AI correctly understood the user. It is categorised as a Read tool in the MCP-Confirm MCP Server, which means it retrieves data without modifying state.
Register the MCP-Confirm MCP server in PolicyLayer and add a rule for verify_understanding: 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 MCP-Confirm. Nothing to install.
verify_understanding 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 verify_understanding 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 verify_understanding. 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.
verify_understanding is provided by the MCP-Confirm MCP server (mako10k/mcp-confirm). 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 →