从 Word 文档提取所有图片.
AI agents call extract_word_images to retrieve information from Office MCP Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves/extracts image data from existing Word documents. It performs no creation, modification, deletion, or execution of arbitrary code. The action is read-only with no side effects. The blast radius of misuse is minimal—an agent could extract images from a document it has access to, but this does not alter the document or have broader system consequences.
From the tool's definition Tool name 'extract_word_images' and description '从 Word 文档提取所有图片' (extract all images from Word document) indicate retrieval of content without modification.
Attacks that exploit this kind of access
从 Word 文档提取所有图片. It is categorised as a Read tool in the Office MCP Server MCP Server, which means it retrieves data without modifying state.
Register the Office MCP Server MCP server in PolicyLayer and add a rule for extract_word_images: 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 Office MCP Server. Nothing to install.
extract_word_images 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 extract_word_images 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 extract_word_images. 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.
extract_word_images is provided by the Office MCP Server MCP server (walkingzzzy/office-mcp). 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 →