mcp-standards

The MCP Standards Server provides a context-aware interface for applying software development standards. It loads a collection of 25 standards and 25 detection rules, and exposes a set of tools that can be invoked by language models or external clients to select, validate, and improve code according to those standards. The server supports multiple languages for analysis and can optionally leverage a Redis caching layer to speed up repeated queries. Typical usage involves starting the server in stdio or HTTP mode and then calling the provided MCP tools to retrieve applicable standards, validate code, and search standards.

Key tools exposed by the MCP server include: get_applicable_standards to determine which standards apply given a project context; validate_against_standard to verify code against a specific standard; suggest_improvements to receive improvement recommendations; search_standards for semantic search across all standards; get_compliance_mapping to map standards to NIST controls; analyze_code to analyze code files for standard compliance; and get_standard_content to retrieve full or compressed standard content. These tools enable closed-loop interactions with LLMs, allowing automated guidance and compliance checks during development.

Prerequisites:

• Python 3.10+ installed on your system
• Git for cloning repositories
• Optional: Redis installed and running if you want caching

Install from PyPI (recommended):

# Install the latest release
pip install mcp-standards-server

# Or install with specific feature sets (if applicable)
pip install "mcp-standards-server[full]"         # All features including web API

Install from source (development):

# Clone the repository
git clone https://github.com/williamzujkowski/mcp-standards-server.git
cd mcp-standards-server

# Create and activate a virtual environment
python -m venv venv
# On macOS/Linux
source venv/bin/activate
# On Windows
venv\Scripts\activate

# Install in development mode
pip install -e .

# Optional: install with extra features
pip install -e ".[full]"         # All features including web API
pip install -e ".[dev]"          # Development tools (linting, formatting)

Running the MCP server:

# Start the server (stdio mode suitable for tool integration)
python -m src

# Or run via the provided CLI for HTTP/daemon modes (depending on package support)
mcp-standards serve --stdio
mcp-standards serve --port 3000

Verifications:

# Basic checks
pytest tests/unit/core/standards/test_rule_engine.py -v
python -c "import src; print('Installation successful')"

Environment variables and configuration tips:

• REDIS_URL: Set when enabling Redis caching to improve performance. Example: redis://localhost:6379/0
• MCP_HTTP_PORT: Use to specify the port for the HTTP server if running in HTTP mode. Default port is implementation-specific.
• If you see import errors after installation, ensure you installed in editable mode (pip install -e .) in a active virtual environment.
• The 25 standards and 25 detection rules are loaded on startup; ensure data files are accessible in the installed package path.
• For web UI usage, Node.js is optional but may be required for the web UI tooling; ensure Redis is running if you enable caching.
• When upgrading from an older version, review the release notes for changes to tools and API compatibility.