AI agents call find_references to retrieve information from LSP MCP Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool searches for and returns all locations where a symbol is referenced in code. It has no side effects—it does not modify, delete, or execute anything. It is purely a retrieval/analysis operation typical of language server protocol query features. The low severity reflects that misuse (e.g., looking up references to sensitive symbols) poses minimal risk compared to write, execute, or destructive operations.
From the tool's definition Tool name 'find_references' and description 'Find all references to a symbol using LSP' indicate a query-only operation that retrieves information without modifying code or executing arbitrary operations.
Documented attack patterns abuse exactly the kind of access find_references gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and LSP MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for find_references:
{
"version": "1",
"default": "deny",
"tools": {
"find_references": {}
}
}find_references 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.
Find all references to a symbol using LSP. It is categorised as a Read tool in the LSP MCP Server MCP Server, which means it retrieves data without modifying state.
Register the LSP MCP Server MCP server in PolicyLayer and add a rule for find_references: 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 LSP MCP Server. Nothing to install.
find_references 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 find_references 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 find_references. 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.
find_references is provided by the LSP MCP Server MCP server (sminnee/lsp-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from LSP MCP 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.
5 LSP MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.