image-collect-search

Retrieve and rank images from a Serper-compatible endpoint, returning downloaded local files.

22|3|Updated Apr 22, 2026
One-click install
npx skills add https://github.com/MadeAgents/TopoClaw --skill image-collect-search
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: image-collect-search
Source: https://github.com/MadeAgents/TopoClaw/tree/main/TopoClaw/topoclaw/skills/image-collect-search
Command: npx skills add https://github.com/MadeAgents/TopoClaw --skill image-collect-search

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

It solves the problem of quickly locating high-quality, relevant images for a topic without manually browsing and comparing results.

Core Features & Use Cases

  • High-precision image collection: Searches via a Serper-compatible image endpoint and returns downloaded local images.
  • Ranking by relevance and quality: Scores candidates using resolution, file size, MIME type, and query/keyword matching while penalizing common noise terms (e.g., logo, icon, emoji, thumbnail).
  • Parallel workflow friendly: Designed to be used alongside normal web text search so the agent can answer facts while this skill supplies image references.
  • Clear operational outputs: Produces selected image metadata (paths and detailed fields like mime, dimensions, score, and reasons) plus a report.json artifact.

Quick Start

Ask TopoClaw to collect two strong images for your query by running the image-collect-search skill with your desired search text.

Frequently Asked Questions about image-collect-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate image search and download local files for a visual reference gathering task?

To automate image search and download local files, this skill queries a Serper-compatible image endpoint using your search text, scores the results, and saves the selected images locally. It returns local file paths along with detailed metadata like dimensions and mime type.

How does image ranking work when collecting pictures from web search results?

Image ranking evaluates resolution, file size, MIME type, and query matching to score candidates. It actively penalizes noise terms like logo, icon, emoji, and thumbnail, ensuring the downloaded local files are high-quality and relevant to your query.

Do I need a specific API key to run an automated image search and retrieval workflow?

Yes, you need a SERPER_API_KEY for authentication to query the image endpoint. You can also optionally configure a SERPER_API_BASE endpoint to target a specific compatible service for your image search and download process.

Can I run parallel text and image search workflows to gather factual answers and visual references?

Yes, this skill is designed for parallel workflow friendly operations, allowing you to run standard text search for facts while it supplies visual references. It outputs selected image metadata with local paths, scores, and justifications in a report.json artifact.

What metadata is returned when downloading images from a Serper-compatible endpoint?

Downloading images via this skill returns local file paths alongside detailed metadata including MIME type, dimensions, byte size, relevance score, and a justification for the ranking. A report.json artifact is also produced for clear operational outputs.