replication-25 -empirical-study

MCP server from SAILResearch/replication-25-mcp-server-empirical-study

Installation

Run this command in your terminal to add the MCP server to Claude Code.

Run in terminal:

Command

claude mcp add --transport stdio sailresearch-replication-25-mcp-server-empirical-study docker run -i replication-25-empirical-study-image \
  --env LOG_LEVEL="info" \
  --env DATASET_DIR="path/to/datasets" \
  --env MCP_SERVER_CONFIG="default"

How to use

This MCP server entry corresponds to the replication-25 empirical study data package. It provides a containerized environment to explore, reproduce, and validate the analyses described in the associated paper. The server is designed to host and interact with a suite of MCP-compatible components (datasets, tools for static analysis, and maintainability metrics) used in the replication effort. You can spin up the container, point it at the datasets included in the repository, and run the provided tooling to inspect server metadata, contributor statistics, issue counts, commit activity, and CI/CD-related metrics that were collected for the study. The tooling supports querying server metadata, listing MCP servers in the replication set, and applying static analysis or poisoning reports to assess risk vectors described in the MCP literature.

How to install

Prerequisites:

Docker installed and running on your system
Access to the replication-25 empirical-study dataset (as described in the repository)

Step-by-step:

Ensure Docker is up to date:
- Windows/macOS: install Docker Desktop from the official site
- Linux: follow your distro's Docker installation guide
Pull or build the container image for the replication-25 empirical study:
- If using a prebuilt image name: replication-25-empirical-study-image docker pull replication-25-empirical-study-image
- If you need to build from a Dockerfile in the repo: docker build -t replication-25-empirical-study-image .
Run the MCP server container: docker run -i replication-25-empirical-study-image
(Optional) Configure environment variables for dataset location and logging:
- DATASET_DIR: Path inside the container where datasets are mounted or located
- LOG_LEVEL: Information/debug level for logs
- MCP_SERVER_CONFIG: Path or flag for the server configuration if applicable
Verify the server is running:
- Check container logs for startup messages
- Use any provided CLI or API endpoints described in the repository to list MCP servers and statistics
Optional: Mount datasets from your host to the container at runtime: docker run -i -v /host/path/to/datasets:/container/datasets replication-25-empirical-study-image

Additional notes

Notes and tips:

The replication-25 empirical study focuses on datasets of MCP servers and maintainability metrics; the container is tailored to access these datasets and run the study's analyses.
If you encounter dataset path issues, ensure DATASET_DIR points to the mounted location inside the container (e.g., /container/datasets).
For reproducibility, record the exact container image version or SHA and any environment variable configurations used to run the analysis.
Some environments require elevated permissions to access Docker or to mount host directories; ensure appropriate access rights.
If you need to customize analysis pipelines, look for configuration files within the container or repository that specify which analyses to run (e.g., static analysis, code smells, vulnerability scans).
This repository contains multiple MCP-related datasets (server lists, vulnerability reports, poisoning reports). Use the provided tools to cross-reference server metadata with vulnerability findings for comprehensive replication workflows.