AI agents call rag_search to retrieve information from Ragi without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves or queries data from a RAG (Retrieval-Augmented Generation) index. Semantic search is a read-only operation that returns matching documents or context based on query input. It has no side effects, does not create, modify, delete, or execute code. The server's stated purpose is 'document retrieval and context-aware queries,' confirming the read-only nature.
From the tool's definition Tool name is 'rag_search' and description states 'Semantic search in the specified project RAG' — performs document retrieval and queries without modifying data.
Attacks that exploit this kind of access
Semantic search in the specified project RAG. It is categorised as a Read tool in the Ragi MCP Server, which means it retrieves data without modifying state.
Register the Ragi MCP server in PolicyLayer and add a rule for rag_search: 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 Ragi. Nothing to install.
rag_search 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 rag_search 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 rag_search. 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.
rag_search is provided by the Ragi MCP server (susutawar/ragi). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →