Compute Anthropic prompt cache hit rate from the latest Claude Code session log. Low hit rate signals wasted tokens — usually caused by reordering tools or system prompt mid-session.
AI agents call cache_stats to retrieve information from Mcp Token Saver 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 session log data to calculate cache metrics. It performs read-only analysis with no ability to modify, delete, or execute external operations. The data queried is diagnostic/telemetry information about token usage, not sensitive secrets or control data. Misuse would only surface inaccurate metrics to inform decisions, posing minimal security risk even if an agent is misled.
From the tool's definition Tool name 'cache_stats' and description 'Compute Anthropic prompt cache hit rate from the latest Claude Code session log' — the verb is 'compute' from existing log data with no modification or side effects.
Documented attack patterns abuse exactly the kind of access cache_stats gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Mcp Token Saver, and nothing reaches the server without passing your rules. This is the rule we recommend for cache_stats:
{
"version": "1",
"default": "deny",
"tools": {
"cache_stats": {}
}
}cache_stats is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Compute Anthropic prompt cache hit rate from the latest Claude Code session log. Low hit rate signals wasted tokens — usually caused by reordering tools or system prompt mid-session. It is categorised as a Read tool in the Mcp Token Saver MCP Server, which means it retrieves data without modifying state.
Register the Mcp Token Saver MCP server in PolicyLayer and add a rule for cache_stats: 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 Token Saver. Nothing to install.
cache_stats 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 cache_stats 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 cache_stats. 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.
cache_stats is provided by the Mcp Token Saver MCP server (talap-creator/mcp-token-saver). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Mcp Token Saver, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
5 Mcp Token Saver tools catalogued and risk-classified — across an index of 43,000+ MCP servers.