AI agents call perform_ocr to retrieve information from MCP OCR Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
perform_ocr retrieves and extracts text content from images—a read-only operation with no ability to modify, delete, or execute code. The blast radius of misuse is minimal; worst case, an AI agent unnecessarily processes many images, consuming compute resources but causing no data loss or security breach. This is a classic Read category tool.
From the tool's definition The tool is part of an OCR server that 'Extracts text from images' with 'production-grade OCR capabilities.' The sibling tools (get_supported_languages, image_to_data, perform_batch_ocr, perform_pdf_ocr) confirm this is a data extraction service with no side…
Documented attack patterns abuse exactly the kind of access perform_ocr gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and MCP OCR Server, and nothing reaches the server without passing your rules. This is the rule we recommend for perform_ocr:
{
"version": "1",
"default": "deny",
"tools": {
"perform_ocr": {}
}
}perform_ocr 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.
perform_ocr. It is categorised as a Read tool in the MCP OCR Server MCP Server, which means it retrieves data without modifying state.
Register the MCP OCR Server MCP server in PolicyLayer and add a rule for perform_ocr: 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 MCP OCR Server. Nothing to install.
perform_ocr 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 perform_ocr 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 perform_ocr. 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.
perform_ocr is provided by the MCP OCR Server MCP server (rjn32s/mcp-ocr). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 5 MCP OCR Server tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
5 MCP OCR Server tools catalogued and risk-classified — across an index of 42,500+ MCP servers.