AI agents call Confidence to retrieve information from MkDocs MCP Search Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves or filters already-retrieved search results by confidence threshold. It has no side effects, creates no data, executes no code, and cannot destroy or modify content. It is a read-only operation on search results. Severity is low because misuse (e.g., setting threshold too high or low) merely affects search quality, not security or data integrity.
From the tool's definition Tool named 'Confidence' with description 'based filtering (configurable threshold)' appears to filter search results based on confidence scores.
Documented attack patterns abuse exactly the kind of access Confidence gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and MkDocs MCP Search Server, and nothing reaches the server without passing your rules. This is the rule we recommend for Confidence:
{
"version": "1",
"default": "deny",
"tools": {
"Confidence": {}
}
}Confidence 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.
based filtering (configurable threshold). It is categorised as a Read tool in the MkDocs MCP Search Server MCP Server, which means it retrieves data without modifying state.
Register the MkDocs MCP Search Server MCP server in PolicyLayer and add a rule for Confidence: 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 MkDocs MCP Search Server. Nothing to install.
Confidence 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 Confidence 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 Confidence. 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.
Confidence is provided by the MkDocs MCP Search Server MCP server (serverless-dna/mkdocs-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from MkDocs MCP Search Server, 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.
3 MkDocs MCP Search Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.