AI agents call monitoring_custom_metrics to retrieve information from Infra Ops without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves monitoring metrics from Prometheus for observability purposes. Querying metrics is a read-only operation that does not create, modify, delete, or execute any system changes. The blast radius of misuse is minimal — an attacker could only gain visibility into infrastructure metrics without the ability to alter infrastructure state or trigger operations.
From the tool's definition Tool name 'monitoring_custom_metrics' and description 'Query custom metrics from a Prometheus endpoint' — the verb 'Query' indicates data retrieval with no modification.
Attacks that exploit this kind of access
Query custom metrics from a Prometheus endpoint. It is categorised as a Read tool in the Infra Ops MCP Server, which means it retrieves data without modifying state.
Register the Infra Ops MCP server in PolicyLayer and add a rule for monitoring_custom_metrics: 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 Infra Ops. Nothing to install.
monitoring_custom_metrics 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 monitoring_custom_metrics 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 monitoring_custom_metrics. 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.
monitoring_custom_metrics is provided by the Infra Ops MCP server (skyvanguard/infra-ops-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 →