takuan-osho Gemini Search
Overview
This skill enables web search functionality using the Google Gemini CLI tool. When this skill is triggered, use the Gemini CLI for web searches instead of the built-in WebSearch tool.
Prerequisites
The gemini CLI tool must be installed and configured on the system. The Gemini CLI provides access to Google's Gemini AI model with web search capabilities.
Usage
When web search is needed, execute the Gemini CLI via the Bash tool:
gemini --prompt "WebSearch: <query>"
About the WebSearch: prefix
The WebSearch: prefix inside the prompt is a convention that signals to Gemini that this request requires web grounding rather than answering from training data. Keep it for any search-style query. Without the prefix, Gemini may respond from its training corpus and miss recent information.
Fallback
If the gemini CLI is unavailable (not installed, unauthenticated, or returns a non-zero exit code), fall back to the built-in WebSearch tool and report the fallback to the user in one line (e.g., "gemini CLI unavailable, falling back to built-in WebSearch"). Do not silently swallow the failure.
Examples
Basic Search
gemini --prompt "WebSearch: latest Python 3.12 features"
Technical Documentation Search
gemini --prompt "WebSearch: Kubernetes pod networking best practices 2025"
API Documentation Search
gemini --prompt "WebSearch: OpenAI API rate limits documentation"
When to Use
This skill is appropriate when:
- The user explicitly requests using Gemini for search
- More comprehensive web search results are needed
- The built-in WebSearch tool has limitations for specific queries
- Access to Google's search capabilities is preferred
Notes
- The Gemini CLI must be properly authenticated and configured
- Search queries should be clear and specific for best results
- Results are returned through the Gemini model's response