recommend_config · ChunkTuner MCP: Risk & Policy

WHAT IT DOES

What recommend_config does on ChunkTuner

AI agents invoke recommend_config to trigger actions in ChunkTuner. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.

WHY IT NEEDS A POLICY

Why recommend_config needs a policy

This tool executes a tuning/benchmarking process (the 'tuner') rather than simply retrieving static data. While it appears to be a computational operation with no destructive or financial implications, and uses dummy embeddings by default (limiting real-world side effects), it still falls into Execute category because it runs a non-trivial process whose behavior depends on input parameters.

From the tool's definition The tool description states 'Run tuner' which indicates execution of a process/algorithm. This is further supported by the context that it benchmarks chunking strategies and returns ranked recommendations based on execution results.

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

recommend_config Rate-limited

recommend_config stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.

View as policy code

policy.json

{
  "version": "1",
  "default": "deny",
  "tools": {
    "recommend_config": {
      "limits": [
        {
          "counter": "recommend_config_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}

RATE-LIMIT THIS TOOL →

Instant setup, no code required.

PolicyLayer is an MCP gateway: every tool call is checked against your rule before it reaches ChunkTuner, so you can run agents against it safely.

FAQ

Questions about recommend_config

What does the recommend_config tool do? +

Run tuner and return ranked Recommendation (uses dummy embeddings by default). It is categorised as a Execute tool in the ChunkTuner MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on recommend_config? +

Register the ChunkTuner MCP server in PolicyLayer and add a rule for recommend_config: 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 ChunkTuner. Nothing to install.

What risk level is recommend_config? +

recommend_config is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.

Can I rate-limit recommend_config? +

Yes. Add a rate_limit block to the recommend_config 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.

How do I block recommend_config completely? +

Set action: deny in the PolicyLayer policy for recommend_config. 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.

What MCP server provides recommend_config? +

recommend_config is provided by the ChunkTuner MCP server (shantanu-deshmukh/chunktuner). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.