AI agents call search_directory to retrieve information from Largefile without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Search operations that query file systems or data structures without modifying them are Read category. The absence of keywords like 'delete', 'edit', 'execute', or 'modify' in the name, combined with the server's stated purpose of navigation and the Read-only nature of sibling tools, strongly suggests this is a retrieval operation with no side effects.
From the tool's definition Tool name is 'search_directory' and belongs to a server described as 'Navigate, search, and edit large codebases, logs, and data files'. The sibling tools (read_content, get_overview, list_directory, search_content) are all Read operations.
Documented attack patterns abuse exactly the kind of access search_directory gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Largefile, and nothing reaches the server without passing your rules. This is the rule we recommend for search_directory:
{
"version": "1",
"default": "deny",
"tools": {
"search_directory": {}
}
}search_directory 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.
search_directory. It is categorised as a Read tool in the Largefile MCP Server, which means it retrieves data without modifying state.
Register the Largefile MCP server in PolicyLayer and add a rule for search_directory: 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 Largefile. Nothing to install.
search_directory 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 search_directory 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 search_directory. 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.
search_directory is provided by the Largefile MCP server (peteretelej/largefile). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Largefile, 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.
8 Largefile tools catalogued and risk-classified — across an index of 43,000+ MCP servers.