aliyun

Maxcompute Semantic — Data skill for Claude Code

Data community

Agent-agnostic semantic skill for MaxCompute SQL — CLI + bare SKILL.md bundle.

How to install Maxcompute Semantic

This entry records only its repository, not the path inside it, so there is no exact command to give. Open aliyun/maxcompute-semantic and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Maxcompute Semantic does

Agent-agnostic semantic skill for MaxCompute SQL — CLI + bare SKILL.md bundle.

Alternatives in Data

Private GPT — Complete API layer for private AI applications on local models: RAG, skills, tools, MCP, text-to-sql, and more 57.5k ★
Privacy & Data Handling — This repository contains: - a plugin package (plugins/compound-engineering) made of markdown/config content 10.8k ★
SQL Queries — Generate SQL queries from natural language across major dialects 7.8k ★

README

maxcompute-semantic (`mcs`)

[![PyPI](https://img.shields.io/pypi/v/maxcompute-semantic)](https://pypi.org/project/maxcompute-semantic/) [![Python](https://img.shields.io/pypi/pyversions/maxcompute-semantic)](https://pypi.org/project/maxcompute-semantic/) [![License](https://img.shields.io/github/license/aliyun/maxcompute-semantic)](LICENSE) [![CI](https://github.com/aliyun/maxcompute-semantic/actions/workflows/ci.yml/badge.svg)](https://github.com/aliyun/maxcompute-semantic/actions/workflows/ci.yml) [![codecov](https://codecov.io/gh/aliyun/maxcompute-semantic/graph/badge.svg)](https://codecov.io/gh/aliyun/maxcompute-semantic) [![English](https://img.shields.io/badge/lang-English-blue)](README.md) [![中文](https://img.shields.io/badge/lang-中文-red)](README.zh-cn.md)

**Give AI agents the context they need to query your MaxCompute data.**

`mcs` is a CLI for building a local semantic package from your MaxCompute project: table and column descriptions, JOIN relationships, UDFs, business metrics, verified SQL, and notes from past queries. Your agent reads that package before it writes SQL, then uses `mcs` to review, cost-check, and execute the query.

[Documentation](https://aliyun.github.io/maxcompute-semantic/) · [PyPI](https://pypi.org/project/maxcompute-semantic/) · [Changelog](CHANGELOG.md)

Why mcs?

AI agents can call MaxCompute, but they do not know your data warehouse. They guess table names, miss JOIN keys, and write SQL that runs but answers the wrong question.

`mcs` gives the agent a local source of truth:

Semantic package - mcs build scans table metadata and produces a structured package the agent can read before writing SQL.
Memory - verified queries, failed patterns, and domain notes accumulate over time. Similar questions can reuse known-good SQL instead of starting cold.
SQL guard rails - cost estimation, write protection, dialect review, and tier-aware schema resolution run before the query reaches MaxCompute.
**Agent