CorvixMCP: Air-Gapped Security Engine for AI Agents

CorvixMCP is a high-performance, compiled native security engine exposed via the Model Context Protocol (MCP). It empowers AI coding assistants (Cursor, Claude Desktop, Antigravity, and autonomous agents) to locally analyze, traverse, and detect multi-cloud attack paths, circular privilege escalation backdoors, and IaC misconfigurations in complete air-gapped isolation with zero data exfiltration.

🚀 Key Capabilities & Architecture

• Sub-Millisecond Native Engine: High-performance compiled C++ core capable of resolving deep multi-hop infrastructure graphs in under 15ms without sending private infrastructure data off-premises.
• Visual Mermaid.js Attack Path Diagrams: Every vulnerability finding includes clean, render-ready Mermaid graph syntax for immediate visual comprehension in AI assistant chats and markdown viewers.
• Automated Surgical Remediation Snippets: Generates exact least-privilege configuration code (Terraform HCL, AWS IAM Deny blocks, GCP gcloud commands, Azure CLI, Kubernetes YAML, GitHub Actions configs, Ansible tasks) to sever toxic edges without disrupting operations.
• Model Context Protocol (FastMCP): Instant standard integration for AI agents to query infrastructure scope, diagnose root causes, and audit security in real time.
• Hardware-Bound Air-Gapped Licensing: Enterprise-grade licensing via Lemon Squeezy with local machine fingerprinting and offline validation.

🔍 Supported Multi-Cloud & IaC Parsers

1. Google Cloud (GCP IAM): Ingests project IAM exports, detecting over-privileged service account delegations and project-wide editor/owner bindings.
2. Amazon Web Services (AWS IAM): Evaluates complex IAM policies, wildcard actions, PassRole, and AssumeRole lateral movement paths.
3. Microsoft Azure RBAC: Maps subscription role assignments, Service Principals, and administrative escalation vectors.
4. Kubernetes (K8s RBAC): Audits ClusterRoles, RoleBindings, and high-privilege ServiceAccount container escape routes.
5. Terraform Infrastructure as Code (tfplan.json): Evaluates planned infrastructure changes and flags public internet ingress exposure (0.0.0.0/0).
6. CI/CD Pipelines (GitHub Actions): Audits workflow files for untrusted triggers (pull_request_target), vulnerable runner configurations, and OIDC cloud credential assumptions.
7. Ansible Playbooks: Ingests automation playbooks to detect unsafe root escalations (become: true), sudo misconfigurations, and server privilege abuse.

🛠️ Quick Start Guide

1. Installation

Install the pre-compiled binary wheel (zero C++ compiler or build toolchain required):

pip install corvixmcp

Or run locally inside your virtual environment:

python3 -m venv venv
source venv/bin/activate
pip install -e .

2. Run the Verification Test Suite

Run the test suite verifying all 7 cloud/IaC parsers, Mermaid visual generation, and surgical remediation:

python3 test_cpp_core.py
python3 tests/test_visualizer_remediation.py

3. Activate & Manage Pro License

To activate, check status, or deactivate your hardware-bound Pro license key via the CLI:

corvixmcp activate <YOUR_LICENSE_KEY>
corvixmcp status
corvixmcp deactivate

📚 Documentation & Project Links

• PROJECT_STATE.md: Complete project tracking dashboard, phase board, and active milestone status.
• PROJECT_PROGRESS.md: Detailed development changelog and completed phases.
• PRODUCT_DESCRIPTION.md: Commercial store listing copy and pricing tiers.
• PRICING.md: Detailed pricing and feature tier comparison.