AI agents call random_float to retrieve information from Basic without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The random_float tool is a pure utility that returns a value (a random float) without side effects. It does not create, modify, delete, or execute external operations. While 'cryptographically secure' suggests it uses entropy, the output is read-only data generation, analogous to a GET request. Misuse by an AI agent (e.g., using it for randomization in harmless decisions) poses minimal risk.
From the tool's definition Tool generates random numbers with no side effects—it retrieves or produces data deterministically without modifying external state, creating resources, executing code, or deleting anything.
Attacks that exploit this kind of access
Generate a cryptographically secure random float between 0 and 1 (or custom range). It is categorised as a Read tool in the Basic MCP Server, which means it retrieves data without modifying state.
Register the Basic MCP server in PolicyLayer and add a rule for random_float: 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 Basic. Nothing to install.
random_float 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 random_float 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 random_float. 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.
random_float is provided by the Basic MCP server (msilverblatt/basic-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 →