A high-performance search integration providing grounded AI answers with citations, raw web results, and image search for AI agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install google-search-grounding
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install google-search-grounding using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Google Search for AI Agents?
This skill provides advanced Google Search capabilities for AI coding agents, leveraging Gemini 2.5 Flash for high-fidelity search grounding. Unlike standard web search tools, it offers synthesized answers with direct, numbered citations, ensuring high accuracy and reliability for complex technical or general queries.
By integrating this into your Openclaw Skills library, you can empower your agents to access real-time information with superior multilingual support, particularly for Hebrew, and the ability to retrieve raw metadata or images. It serves as a robust alternative to built-in search functions, offering deeper control over localization and output formats.
Google Search for AI Agents Use Cases
- Obtaining synthesized answers to complex questions with verified citations.
- Generating reference lists and retrieving raw URLs for deep technical research.
- Searching for images and visual assets directly through an AI agent workflow.
- Performing localized searches in specific languages like Hebrew or Arabic with high precision.
How Google Search for AI Agents Works
- The AI agent initiates a search request using the script, specifying the query and the desired mode (search, raw, or image).
- For the search mode, the skill utilizes Gemini 2.0 Flash with Search Grounding to synthesize a response based on live web data.
- For raw or image modes, the tool interfaces with the Google Custom Search JSON API to fetch specific snippets or image URLs.
- The skill processes the results, applying language and country filters as defined in the Openclaw Skills environment configuration.
- The final output is returned to the agent as formatted text with sources or a structured JSON object for programmatic use.
Google Search for AI Agents Setup
To get started, install the required library and run the installation script:
pip install google-genai
bash skills/google-search/install.sh
Configure your environment variables in your configuration file to enable the Openclaw Skills integration:
{
"env": {
"GOOGLE_API_KEY": "YOUR_GOOGLE_API_KEY",
"GOOGLE_CSE_CX": "YOUR_CUSTOM_SEARCH_ENGINE_ID",
"GOOGLE_SEARCH_LANG": "he",
"GOOGLE_SEARCH_COUNTRY": "IL"
}
}
Google Search for AI Agents Data Schema & Taxonomy
The skill organizes its output based on the selected mode to ensure the data is easily consumable by agents.
| Mode | Data Type | Description |
|---|---|---|
| search | Markdown | Synthesized answer followed by a numbered list of sources |
| raw | List | Page titles, URLs, and text snippets from the top search results |
| image | List | Image URLs and source page titles for visual queries |
| --json | JSON | Full structured response including grounding metadata and confidence scores |
Google Search for AI Agents Advanced Features
- Smart fallback mechanism that can trigger alternative web search tools if Google quotas are exceeded.
- Automatic 429 rate limit handling with a 5-second retry delay for high-reliability automation.
- Comprehensive JSON output mode for developers building complex multi-agent Openclaw Skills workflows.
- Direct control over search localization using the language and country resolution hierarchy.
SKILL.md
Loading
Related Openclaw Skills
Sentinel Shield
Sentinel Shield provides robust runtime security and monitoring for Openclaw Skills to prevent token theft, session hijacking, and malicious prompt injections.
Mixlab Solo Scope
Automates the retrieval and intelligent summarization of the Mixdao Solo Scope RSS feed into structured, categorized Markdown briefs.
Daily Briefing Agent
An automated workflow that fetches daily news, categorizes content via MiniMax AI, and publishes briefings to mixdao.
Balance Management
A comprehensive toolset for managing Routstr API credits, Lightning Network invoices, and Cashu token redemptions.
Hebrew Nikud Reference for AI Agents
A comprehensive technical reference for AI agents to apply selective Hebrew vowel points for accurate pronunciation and Text-to-Speech output.
Tavily Search Pro
A comprehensive AI-powered search and research engine that enables agents to search, crawl, extract, and map web data with high precision.
METADATA
Featured*
