检测异常使用情况
AI agents call detect_anomalies to retrieve information from Claude Stats 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 analyzes existing usage data to identify patterns or outliers. It has no side effects, does not modify data, does not execute external operations, and does not access financial systems. The worst-case misuse would be an AI agent generating false anomaly reports or excessive queries, neither of which causes material harm. Classification as Read is appropriate.
From the tool's definition Tool name 'detect_anomalies' and description 'detect anomalies in usage' indicate passive data analysis. The server context emphasizes 'query', 'analyze', 'compare', and 'generate reports'—all read-only operations.
Attacks that exploit this kind of access
检测异常使用情况. It is categorised as a Read tool in the Claude Stats MCP MCP Server, which means it retrieves data without modifying state.
Register the Claude Stats MCP server in PolicyLayer and add a rule for detect_anomalies: 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 Claude Stats MCP. Nothing to install.
detect_anomalies 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 detect_anomalies 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 detect_anomalies. 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.
detect_anomalies is provided by the Claude Stats MCP server (upjiang/ai-mcp-study). 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 →