minimax-vision-captcha_skill

Overview

This skill uses the MiniMax visual model to recognize and analyze CAPTCHA images, slider positions, and embedded text. It helps automate visual verification tasks by returning precise object locations, missing-piece positions, and extracted text from screenshots or image files. The skill is optimized for scenarios like web and mobile verification flows where automated visual interpretation is required.

How this skill works

The skill inspects an image file or a browser screenshot and runs MiniMax visual understanding to detect CAPTCHA elements, slider gaps, and readable text. It returns structured information such as coordinates, distance-to-edge for sliders, and OCR text results. Results are suitable for driving automated actions (click, drag, or input) or for human verification.

When to use it

• Automating slider CAPTCHA detection and computing drag distances.
• Extracting visible text from screenshots or images via OCR.
• Analyzing web page screenshots for visible UI elements and their positions.
• Preprocessing verification steps in login or form workflows.
• When you need machine-readable coordinates and textual content from images.

Best practices

• Provide server-accessible absolute image paths or pass screenshots directly from the browser tool.
• Capture full context screenshots (include surrounding UI) to improve position accuracy.
• Validate returned coordinates against image dimensions before performing automated clicks or drags.
• Prefer PNG format for screenshots to preserve clarity for OCR and edge detection.
• Confirm MiniMax MCP configuration and access before batch processing images.

Example use cases

• Detect the missing-piece location in a WeChat slider CAPTCHA and compute the drag distance to solve it.
• Run OCR on a webpage screenshot to extract account numbers, labels, or verification codes.
• Analyze a snapshot to locate a specific button or input field by visible appearance for automated testing.
• Process a series of screenshots to build an index of visible text and UI change logs.