弹出图片选择对话框，让用户选择图片文件或从剪贴板粘贴图片。
AI agents call pick_image to retrieve information from MCP Feedback Collector without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs image selection and retrieval only—it reads user input through a GUI dialog box. There are no side effects, no data modification, no code execution, and no irreversible actions. It is a standard Read operation that facilitates gathering user-provided input as part of the feedback collection workflow.
From the tool's definition Tool allows users to select and upload image files or paste from clipboard. The description indicates it 'pops up an image selection dialog' for users to 'choose image files or paste images from clipboard.' This is a file selection interface that retrieves…
Attacks that exploit this kind of access
弹出图片选择对话框，让用户选择图片文件或从剪贴板粘贴图片。. It is categorised as a Read tool in the MCP Feedback Collector MCP Server, which means it retrieves data without modifying state.
Register the MCP Feedback Collector MCP server in PolicyLayer and add a rule for pick_image: 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 Feedback Collector. Nothing to install.
pick_image 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 pick_image 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 pick_image. 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.
pick_image is provided by the MCP Feedback Collector MCP server (zhaobu/mcp-feedback-collector). 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 →