Autonomous Data QA Layer for Google Workspace. Built for the Google Cloud Rapid Agent Hackathon

Teams relying on spreadsheets for complex scientific modeling, financial indexing, or tokenomics constantly fight human error. A single missing negative sign or misaligned row can skew an entire dataset (e.g., ruining a Shannon-Weiner biodiversity index). Manual QA is slow, expensive, and error-prone.

SanitySync is a lightweight Model Context Protocol (MCP) server written in Rust. It exposes specific deterministic tools (like flag_anomaly) to a Gemini-powered QA Agent.

Instead of acting as a conversational chatbot, Gemini operates as a silent backend auditor. It ingests massive rows of spreadsheet data, autonomously checks the mathematical logic, and when it detects an anomaly, it formulates a structured JSON-RPC call demanding the Rust server physically update the sheet with a correction.

1. Data Ingestion: Rust fetches target data ranges.
2. Cognitive Audit: Gemini 2.5 Flash evaluates the logic (e.g., verifying pi * ln_pi calculations).
3. Tool Execution: If an error is found, Gemini triggers the flag_anomaly Rust function, passing the exact row, column, explanation, and auto-fix parameters.

###
Note: Ensure your .env contains a valid GEMINI_API_KEY.


***

### 3. Push to GitHub safely
You already ran `echo ".env" >> .gitignore`, so your new key is safe. 
Run the standard Git flow:

```bash
git init
git add .
git commit -m "Initial commit: Core MCP architecture and Gemini tool-calling"
Then create the sanity_sync repo on your GitHub and push the code:

Bash
git branch -M main
git remote add origin https://github.com/YOUR_USERNAME/sanity_sync.git
git push -u origin main
git clone [https://github.com/YOUR_GITHUB/sanity_sync.git](https://github.com/YOUR_GITHUB/sanity_sync.git)
cd sanity_sync
cargo run