Google ai search
Token-efficient public-web research through Gemini API with Google Search grounding and source links. Use when the user asks to search the web, google something, verify current public information, compare options, or get a compact research summary with citations.From its SKILL.md
npx -y skills add MattoYuzuru/yuzuru-skills --skill google-ai-searchAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
2.3 KB, as published. Nobody here has run it
Google AI Search
Workflow
- Rewrite the request into a precise search query.
- Pick the answer language: English for broad technical topics, Russian for Russia-specific topics, or the user's requested source language.
- Resolve this skill directory and run the configuration check:
python3 scripts/setup.py check
On Windows, use py -3 scripts\setup.py check.
- If
readyisfalse, tell the user to:- create a key at https://aistudio.google.com/apikey;
- run the
next_actioncommand in their own terminal; - paste the key into the hidden prompt and confirm when complete.
- Never ask the user to paste an API key into chat, and never pass it as a command-line argument or print it in logs.
- After confirmation, rerun
setup.py check. If ready, run:
python3 scripts/search.py \
--query "OpenAI latest model announcements 2026" \
--include-sources \
--lang en
Use the installed google-ai-search launcher when available. The default model
is the stable gemini-3.1-flash-lite; override it with
GOOGLE_AI_SEARCH_MODEL when Google announces a newer compatible model.
Result Handling
- Summarize the JSON
answerrather than pasting it blindly. - Cite URLs from
sourcesnear the claims they support. - Keep the default 10-source bound; raise
--max-sourcesonly when comparison breadth requires it. - Request
--include-usageonly when diagnosing cost or quota behavior. - Mention uncertainty when sources are weak or missing.
- If authentication, quota, regional access, or API availability blocks the request, report it clearly and use another search path.
- Prefer primary sources for documentation, policy, pricing, legal text, and API behavior.
- Do not use this as the only source for medical, legal, financial, or other high-stakes answers.
- When the user explicitly asks not to search the web, do not use this skill.
Read references/setup.md only for OS-specific setup, configuration overrides, or free-tier details.