Smart Image Search

Search images from text queries and return the most relevant image result, candidate images, source pages, or ready-to-open search links. Use when the user a...

This skill scrapes Bing, Baidu, and Sogou, ranks results, and downloads selected images to a local directory (/home/mumu/clawd/tmp/search-image). Before installing, consider: (1) privacy and data policy—downloads come from arbitrary external URLs and could contain inappropriate content or copyrighted material; (2) disk usage and filesystem access—the scripts create files in a hard-coded path (adjust or sandbox that path if needed); (3) network usage—scraping may trigger service blocks (the code already handles Baidu verification); (4) trust and licensing—ensure you are comfortable with the agent fetching and sending remote images. If you need stricter boundaries, require the skill not to auto-download (only return links) or change the output path to a safer, configurable location.

Type: OpenClaw Skill Name: smart-image-search Version: 0.1.0 The smart-image-search skill bundle is a well-structured tool designed to aggregate, rank, and download images from multiple search engines (Bing, Baidu, Sogou). The core logic in scripts like search_best_image.py and fetch_best_image.py implements sophisticated relevance scoring, intent detection (e.g., official logos vs. memes), and image quality validation (checking dimensions and file headers). While the skill performs network requests and writes files to a local directory (/home/mumu/clawd/tmp/search-image), these actions are strictly necessary for its stated purpose. The code includes input sanitization for filenames and uses standard libraries for all operations, with no evidence of malicious intent, data exfiltration, or prompt injection attacks.