AI agents invoke lean_run_code to trigger actions in Lean Lsp. 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.
Running code in a theorem prover environment, even if sandboxed, is an Execute action that can trigger external operations and produce effects dependent on the code arguments. Misuse could cause DoS, resource exhaustion, or unintended proof state modifications.
From the tool's definition Tool name 'lean_run_code' combined with context of 'Lean theorem prover via LSP' indicates code execution capability.
Documented attack patterns abuse exactly the kind of access lean_run_code gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Lean Lsp, and nothing reaches the server without passing your rules. This is the rule we recommend for lean_run_code:
{
"version": "1",
"default": "deny",
"tools": {
"lean_run_code": {
"limits": [
{
"counter": "lean_run_code_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}lean_run_code 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.
Free to start. No card required.
lean_run_code. It is categorised as a Execute tool in the Lean Lsp MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Lean Lsp MCP server in PolicyLayer and add a rule for lean_run_code: 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 Lean Lsp. Nothing to install.
lean_run_code is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the lean_run_code 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 lean_run_code. 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.
lean_run_code is provided by the Lean Lsp MCP server (project-numina/lean-lsp-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Lean Lsp, 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.
22 Lean Lsp tools catalogued and risk-classified — across an index of 43,000+ MCP servers.