VulniCheck is an AI-powered vulnerability scanner exposed as an HTTP MCP server for real-time security analysis of Python projects and GitHub repositories. Deployed as a Docker container with HTTP streaming (no SSE), it supports secure, scalable deployment suitable for CI/CD and cloud environments. The MCP server analyzes dependencies in common files (requirements.txt, pyproject.toml, setup.py) for known vulnerabilities by querying more than five databases (OSV.dev, NVD, GitHub Advisory, CIRCL, Safety DB) and provides AI-powered risk assessments with remediation guidance using OpenAI and Anthropic APIs. It also detects exposed secrets and credentials, analyzes Dockerfiles for insecure packages, and validates MCP configurations to ensure secure operation. Optional Google OAuth 2.0 authentication secures access, with token persistence via Docker volumes, and additional API keys may be supplied to boost rate limits and AI capabilities. GitHub integration enables scans of public or private repositories (up to 1GB). Features include smart caching to avoid redundant scans, zero-config operation, and seamless integration with Claude Code via HTTP MCP transport. This makes VulniCheck a production-ready, AI-assisted security scanner for modern Python projects and repositories.