使用说明

brd-enhancer-mcp

A package that ships seven MCP servers for an AI-assisted SDLC workflow. Each is a separate executable you can register on its own in your IDE's MCP config:

Server	Executable	Purpose
enhance-prompt	`prompt-enhancer-mcp`	Enhance dev tasks with project documentation context
test-workflow	`test-workflow-mcp`	Generate & submit Gherkin test cases from Confluence
pipeline-analyzer	`pipeline-analyzer-mcp`	Analyze Harness pipeline failures with RAG context from past incidents
code-documentation	`code-documentation-mcp`	Generate Markdown code documentation and publish it to Confluence
unit-test	`unit-test-mcp`	Generate unit tests + coverage; plus mutation testing (fabrication-resistant)
code-quality	`code-quality-mcp`	Complexity + lint/quality (whole-repo or changed files) + Copilot-as-judge review + architecture structure (golden-graph diff)
security	`security-mcp`	Scan SAST (semgrep) + dependency CVEs (trivy) + secrets (gitleaks); finding-led

Prerequisites

⚠️ Use Official Python (NOT Microsoft Store Python)

Download Python from https://www.python.org/downloads/

During installation, make sure to check:

✅ "Add Python to PATH"

Verify you have the correct Python:

where python

Output	Status
`C:\Python313\python.exe`	✅ Official Python — Good
`C:\Users\...\WindowsApps\python.exe`	❌ Microsoft Store Python — Reinstall from python.org

Installation

Option A — Global Install (Recommended for most developers)

pip install git+https://github.com/arushsingh17/mcp.git

Verify installation:

pip show brd-enhancer-mcp
where brd-enhancer-mcp        # Windows
which brd-enhancer-mcp        # Mac/Linux

Option B — Virtual Environment Install

# Step 1: Create and activate a virtual environment
python -m venv venv
venv\Scripts\activate          # Windows
source venv/bin/activate       # Mac/Linux

# Step 2: Install the package
pip install git+https://github.com/arushsingh17/mcp.git

# Step 3: Get the exact executable path (needed for config)
python -c "import shutil; print(shutil.which('brd-enhancer-mcp'))"

Example output:

C:\Users\YourName\Desktop\myproject\venv\Scripts\brd-enhancer-mcp.exe

Copy this path — you will need it in the config below.

Configuration

🌍 Global Install Config

Since brd-enhancer-mcp is registered in system PATH, no file path is needed.

Claude Desktop → %APPDATA%\Claude\claude_desktop_config.json (Windows) Claude Desktop → ~/Library/Application Support/Claude/claude_desktop_config.json (Mac) Claude Code → ~/.claude.json

{
    "mcpServers": {
        "brd-enhancer": {
            "command": "brd-enhancer-mcp",
            "env": {
                "API_KEY": "your_api_key",
                "PROJECT_ID": "your_project_id",
                "API_URL": "https://your-backend.com"
            }
        }
    }
}

📦 Virtual Environment Config

Use the full path you got from the command above.

⚠️ On Windows replace every \ with \\ in the path

{
    "mcpServers": {
        "brd-enhancer": {
            "command": "C:\\Users\\YourName\\Desktop\\myproject\\venv\\Scripts\\brd-enhancer-mcp.exe",
            "env": {
                "API_KEY": "your_api_key",
                "PROJECT_ID": "your_project_id",
                "API_URL": "https://your-backend.com"
            }
        }
    }
}

Mac/Linux venv config:

{
    "mcpServers": {
        "brd-enhancer": {
            "command": "/Users/yourname/myproject/venv/bin/brd-enhancer-mcp",
            "env": {
                "API_KEY": "your_api_key",
                "PROJECT_ID": "your_project_id",
                "API_URL": "https://your-backend.com"
            }
        }
    }
}

Environment Variables

Shared (all servers — only three vars total)

Variable	Required	Default	Description
`API_KEY`	✅ Yes	—	Your personal backend API key
`PROJECT_ID`	✅ Yes	—	Your project ID/GUID
`API_URL`	❌ No	`http://localhost:8000`	Backend URL

enhance-prompt only (optional tech-stack hints)

Variable	Required	Default	Description
`FRONTEND_REQUIREMENTS`	❌ No	—	e.g. `react18` — appended to the enhanced prompt
`BACKEND_REQUIREMENTS`	❌ No	—	e.g. `redis,postgresql,express` — appended to the enhanced prompt

pipeline-analyzer — no Harness env vars needed

The pipeline-analyzer used to require five HARNESS_* env vars. It no longer does — the backend resolves your stored Harness credentials by project owner. Link Harness once in Settings → Harness on the SDLC frontend; the MCP picks it up automatically (cached 5 min, so changes propagate within ~5 minutes without a restart).

code-documentation — uses the project owner's Atlassian credentials

No Confluence env vars needed. The backend uses the project owner's stored Atlassian credentials and the project's linked Confluence space to publish documentation pages.

Same config shape for all four servers

{
    "mcpServers": {
        "enhance-prompt": {
            "command": "prompt-enhancer-mcp",
            "env": {
                "PROJECT_ID": "your_project_id",
                "API_KEY":    "your_backend_api_key",
                "API_URL":    "https://your-backend.com"
            }
        },
        "test-workflow": {
            "command": "test-workflow-mcp",
            "env": {
                "PROJECT_ID": "your_project_id",
                "API_KEY":    "your_backend_api_key",
                "API_URL":    "https://your-backend.com"
            }
        },
        "pipeline-analyzer": {
            "command": "pipeline-analyzer-mcp",
            "env": {
                "PROJECT_ID": "your_project_id",
                "API_KEY":    "your_backend_api_key",
                "API_URL":    "https://your-backend.com"
            }
        },
        "code-documentation": {
            "command": "code-documentation-mcp",
            "env": {
                "PROJECT_ID": "your_project_id",
                "API_KEY":    "your_backend_api_key",
                "API_URL":    "https://your-backend.com"
            }
        }
    }
}

Quick Reference

Scenario	Command in config
Global install (Official Python)	`"command": "brd-enhancer-mcp"`
Global install (Microsoft Store Python)	Full path from `where brd-enhancer-mcp`
Virtual environment (any OS)	Full path from `python -c "import shutil; print(shutil.which('brd-enhancer-mcp'))"`