Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/HasData/google-search-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/hasdata/google-search-mcp/hasdata-google-search)<a href="https://agentmods.dev/rules/hasdata/google-search-mcp/hasdata-google-search"><img src="https://agentmods.dev/badge/rules/hasdata/google-search-mcp/hasdata-google-search/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/rules/hasdata/google-search-mcp/hasdata-google-search"><img src="https://agentmods.dev/badge/rules/hasdata/google-search-mcp/hasdata-google-search.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00018 | $0.00543 |
| Opus 5.5 | $0.00007 | $0.00217 |
| Sonnet 5.5 | $0.00004 | $0.00109 |
| Haiku 4.5 | $0.00002 | $0.00054 |
Grade A, and why
hasdata-google-search scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 33 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Google Search through HasData
Ten tools over one endpoint. Pick by what the answer needs, because they differ in price and in payload.
Which tool
hasdata_google_serp_serp_getSearchResults is the default: organic results, People Also Ask, related searches and the inline AI Overview when the page has one. hasdata_google_serp_serp_light_getSearchResults returns organic results only, which makes it the lighter call, so reach for it when the task is ten blue links.
hasdata_google_serp_news_getGoogleNews for news, hasdata_google_serp_shopping_getSearchResults for products with prices, hasdata_google_serp_short_videos_getShortVideosSearchResults for Shorts and Reels, hasdata_google_serp_ai_mode_getAiModeResponse for the conversational answer with its citations.
Tokens, not ids
Three tools expand something a search already returned, and they take a token rather than a query. ai_overview takes the pageToken of an AI Overview block, immersive_product takes the token on a shopping card, events takes the pageToken on an item in eventsResults. Run the search first, read the token out of the response, then call the tool. Tokens go stale, so a token from yesterday will not work today.
Localisation
location, gl and hl decide what Google shows. A query without them is answered from wherever the request leaves from, which is rarely what the user meant. Ask for the market when the task is about a country, and pass location as Google writes it, for example Austin,Texas,United States.
What to watch
An empty result set is a valid answer, not an error. num caps results per page and Google treats it as a hint. Every successful call spends credits from the connected HasData account, so one well-aimed search beats three broad ones.
Prompts and resources
The server also serves resources that spell out the accepted values of every parameter that takes a fixed list, each one at hasdata://google_serp/<parameter>. Reading one spends no credits and makes no upstream request, which beats discovering the vocabulary from a rejected call. Listing them needs no API key either, so a client can read them before the user has connected an account.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago Changed · +6 lines 93a94060449f
- 16d ago Changed 4a2b28044ca9
- 23d ago First seen · 27 lines · 18 tokens per session scan A a8ab670838af
hasdata-google-search is a cursor rule published in the GitHub repository HasData/google-search-mcp (17 stars, last pushed 3d ago), licensed MIT. It adds 18 tokens to every session and 543 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-23.
Other cursor rules, from other repositories
nativescript
NativeScript best practices and patterns for mobile applications.
python
Python best practices and patterns for modern software development with Flask and SQLite.
cursor
You are working on the checkout service. Preserve transaction integrity and auditability.
error-handling-and-validation
A code-quality guide for handling errors, invalid inputs, edge cases, preconditions, logging, and user-friendly error messages.
archcore-files
Enforce MCP-only operations when working with .archcore/ files.
dev_workflow
Guide for using Taskmaster to manage task-driven development workflows.