AI agents use upload_rag to create or update resources in RAGFlow MCP — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your RAGFlow MCP environment.
Upload actions create or modify data within a system (Write category). Severity is medium because upload to a RAG system could introduce malicious training data, poison search results, or consume storage, but lacks the irreversibility of Destructive operations or the financial impact of Financial ones.
From the tool's definition Tool name 'upload_rag' indicates data upload/ingestion. Description is empty, limiting precision. Context suggests RAGflow dataset/RAG system operations.
Documented attack patterns abuse exactly the kind of access upload_rag gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and RAGFlow MCP, and nothing reaches the server without passing your rules. This is the rule we recommend for upload_rag:
{
"version": "1",
"default": "deny",
"tools": {
"upload_rag": {
"limits": [
{
"counter": "upload_rag_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}upload_rag stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
upload_rag. It is categorised as a Write tool in the RAGFlow MCP MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the RAGFlow MCP server in PolicyLayer and add a rule for upload_rag: 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 RAGFlow MCP. Nothing to install.
upload_rag is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the upload_rag 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 upload_rag. 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.
upload_rag is provided by the RAGFlow MCP server (oraichain/ragflow-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from RAGFlow MCP, 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.
4 RAGFlow MCP tools catalogued and risk-classified — across an index of 43,000+ MCP servers.