# AI SDLC Quality Gate MCP Server

This is a personal learning project for exploring Model Context Protocol (MCP),
static code analysis, QA automation, and practical SDLC quality gates.

It is designed to be a small, local-first reference implementation that can be
used from Cursor, Claude Desktop, or Claude Code while experimenting with
developer workflow automation.

## Tool Highlights

| Tool | Outcome |
| --- | --- |
| `audit_code_quality` | Static Python/TypeScript security and reliability findings. |
| `generate_playwright_spec` | Executable async Playwright/Pytest test source. |
| `run_sdlc_healthcheck` | 0-100 repository quality score and recommendations. |

## Quickstart

```bash
python -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements.txt
python -m src.server
python -m pytest -q
```

## Cursor and Claude Integration

Copy `mcp_config.json` into Cursor or Claude Desktop's `mcpServers` configuration and replace `cwd` with this repository's absolute path. Claude Code can use the same command: `python -m src.server`.

## Screenshots / GIF Guide

After connecting the server, capture one successful interaction for each tool and save the assets under `docs/assets/`.

## Safety Boundaries

The server performs local static analysis only. It does not execute target code, need credentials, call an LLM, or use a network API.