bigquery-mcp CLAUDE.md

bigquery-mcp CLAUDE.md is an instructions file for coding agents from pvoo/bigquery-mcp. It costs 1,465 tokens per session, scanned A, original, MIT.

Project instructions for a Python MCP server that lets an agent work with Google BigQuery, a service for storing and querying large datasets.

In plain words
What is it for?
Use them when developing the BigQuery MCP server, setting up its environment, running quality checks, or launching it for a chosen Google Cloud project and location.
Why use it?
They give contributors a consistent way to install, run, inspect, format, type-check, and test the server.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/pvoo/bigquery-mcp/claude-md
Clone the repo
git clone --depth 1 https://github.com/pvoo/bigquery-mcp

Wrote 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.

agentmods badge for bigquery-mcp CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/pvoo/bigquery-mcp/claude-md.svg)](https://agentmods.dev/instructions/pvoo/bigquery-mcp/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/pvoo/bigquery-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/pvoo/bigquery-mcp/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 1,465 This file is loaded in full into every session.
When invoked 1,465 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.01465 $0.01465
Opus 5 $0.00732 $0.00732
Sonnet 5 $0.00293 $0.00293
Haiku 4.5 $0.00146 $0.00146

Measured 5d ago against content hash bd2cd98af1d4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

bigquery-mcp CLAUDE.md 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 5d 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.

CLAUDE.md · 150 lines

How it starts

The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

BigQuery MCP server - A Python implementation using FastMCP to provide BigQuery operations through the Model Context Protocol. The project is optimized for navigating large datasets efficiently while keeping LLM context minimal.

Development Commands

# Setup (recommended approach)
make install                # Setup environment and pre-commit hooks
# OR manual setup:
uv sync                     # Install all dependencies
uv run pre-commit install   # Setup git hooks

# Run the server
make run                    # Runs bigquery-mcp console script
# OR:
uv run bigquery-mcp --project YOUR_PROJECT --location US

# Development workflow
make check                  # Run all quality checks (lint, type, format)
make test                   # Run pytest test suite
make inspect                # Launch MCP inspector for testing

# Individual tools
uv run ruff check .         # Lint code
uv run ruff format .        # Format code
uv run mypy                 # Type checking (configured for src/ directory)
uv run pytest              # Run tests

# Build and distribution
make build                  # Build wheel file
make clean                  # Clean build artifacts

Architecture

Modular Package Structure

  • src/bigquery_mcp/server.py: Main entry point and CLI argument handling
  • src/bigquery_mcp/bigquery_tools.py: Core MCP tool implementations
  • src/bigquery_mcp/auth.py: Authentication helpers and error formatting
  • src/bigquery_mcp/query_safety.py: SQL query validation and safety checks
  • FastMCP decorators for clean tool definitions
  • Async operations for BigQuery interactions

Tool Implementation Pattern

@mcp.tool()
async def tool_name(param: str) -> dict:
    """Tool description for MCP."""
    # Input validation with helpful error messages
    # BigQuery operation with proper client handling
    # Comprehensive error handling with context
    # Return structured response optimized for LLM consumption

Read the full file on GitHub · 150 lines

Changes

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.

  1. 5d ago First seen · 150 lines · 1,465 tokens per session scan A bd2cd98af1d4

Subscribe to this mod's changes

bigquery-mcp CLAUDE.md is an instructions file published in the GitHub repository pvoo/bigquery-mcp (9 stars, last pushed 3mo ago), licensed MIT. It adds 1,465 tokens to every session, about $0.0073 per session on Opus 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 2026-08-31.