# harishsg993010/damn-vulnerable-MCP-server

Damn Vulnerable MCP Server

Repository: https://github.com/harishsg993010/damn-vulnerable-MCP-server
Canonical: https://ross.abutalabs.com/products/damn-vulnerable-mcp-server
Language: Python
License Family: other
Last push: 2025-12-08T00:14:14+00:00

## Health v2 (maintenance only)
Score: 45/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 56, release rhythm 35, longevity 36
- inputs: {"age_days": 504, "days_push": 269, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1339, forks 173 (observed 2026-08-28T04:04:25.870315+00:00)

## What it is
Damn Vulnerable MCP Server (DVMCP) is a deliberately vulnerable implementation of the Model Context Protocol built as an educational security lab containing 10 challenges of increasing difficulty. It demonstrates MCP-specific attack vectors such as prompt injection, tool poisoning, excessive permissions, rug pull attacks, tool shadowing, indirect injection, and token theft.

## Use cases
- practice exploiting prompt injection and tool poisoning in MCP servers
- learn LLM and AI agent security through hands-on challenges
- set up a vulnerable MCP target for security training or a CTF
- study MCP attack vectors like rug pull, tool shadowing, and token theft
- test AI agent security tooling against realistic vulnerable MCP implementations
- teach developers how to secure MCP tools, resources, and prompt implementations

## When to choose
- You want a safe, self-hosted lab to practice attacking MCP/LLM integrations
- You are a security researcher, developer, or AI safety professional studying MCP-specific vulnerabilities
- You need challenge-based material with escalating difficulty for security education

## When to avoid
- You need a production-ready or hardened MCP server for real applications
- You want a clean reference implementation of the MCP protocol
- You need a native Windows environment (the project recommends Docker or Linux)

## Facets
- artifact type: learning-resource
- maturity: active
- function: security, mcp, penetration-testing
- domain: security, artificial-intelligence, large-language-models, penetration-testing, education
- platform: python, self-hosted
- tags: mcp-security, dvmcp, prompt-injection, tool-poisoning, llm-security, ai-agent-security, security-training, ctf-challenges, deliberately-vulnerable, hands-on-lab, ai-safety, docker, linux

## Member repositories
- harishsg993010/damn-vulnerable-MCP-server (main) score 45

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:25.870315+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:43:54.951452+00:00, confidence not recorded.
  - readme: https://github.com/harishsg993010/damn-vulnerable-MCP-server (fetched 2026-08-28T04:04:25.870315+00:00, sha 150e79c77aaa)
- Data as of 2026-08-30T08:39:29.467469+00:00.