AI Codebase Drift & Security Inspection Gate

A Shift-Left Dual-Mode MCP Server & CI/CD Drift Engine. Stop hallucinated dependencies, architectural violations, and security bypasses before they merge.

Get Started Learn More

guardrail.exe scan

guardrail scan .

[GuardRail-Agent] | AI Codebase Drift & Security Inspection

[*] Analyzing dependencies...
[OK] No hallucinated PyPI packages detected.

[*] Analyzing AST for Architectural Drift...
[OK] Imports comply with boundaries in .drift-rules.toml

[*] Scanning for Security Bypasses (# type: ignore, nosec)...
PASSED
Clean bill of health!
No hallucinated dependencies, architectural boundary violations, or security bypasses detected.

_

To use your own .gif simulation, replace the image source in index.html with the path to your .gif file.

Enterprise-Grade Code Guardrails

Protect your codebase from AI-generated vulnerabilities.

PyPI Slopsquatting Detection

Validates dependencies asynchronously against the live PyPI registry to catch hallucinated package names injected by AI coding assistants before execution.

Architectural Drift Analysis

Uses Tree-sitter AST parsing to enforce modular boundaries defined in `.drift-rules.toml`. Prevents domain logic from directly calling infrastructure layers.

Security Bypass Detection

Scans the syntax tree to pinpoint unauthorized suppression comments like `# type: ignore` or `# nosec`, ensuring security standards are upheld.

Native MCP Server

Plugs directly into Antigravity, Claude Desktop, and Cursor via the Model Context Protocol (`mcp>=1.0.0`). AI agents can verify their own code contextually.

Rich CLI & SARIF Exports

Beautiful terminal UI with detailed diagnostic outputs. Export scan results as standard SARIF JSON for seamless integration with GitHub Advanced Security.

Lead Developer & Researcher

Rock Atamkeze

Fullstack Software Engineer | AI Engineer | Ethical Hacking & AI Security Specialist

Bridging the gap between rapid AI-assisted development and robust application security. GuardRail-Agent was engineered to ensure that autonomous agents and human developers alike can push boundaries without compromising structural integrity or introducing supply chain vulnerabilities.