Skip to main content
WardenOpen-source AI scannerExplore →
View on GitHubMIT Licensed

How governed are your AI agents?

One command. One score.

Quick Install

$ pip install warden-ai
$ warden scan

30 seconds. No API key. No signup.

What It Finds

MCP Server Inventory

Discovers all connected tools and MCP servers in your environment.

Policy Gap Analysis

Identifies which tools and actions lack governance policies.

Deny-by-Default Check

Determines if your posture is open (allow-all) or closed (deny-by-default).

Environment Exposure

Detects leaked secrets, API keys, and exposed environment variables.

Log Hygiene

Evaluates audit trail completeness and log retention configuration.

17 Scoring Dimensions

Core Governance

Tool Inventory25/10
Risk Detection20/10
Policy Coverage20/10
Credential Mgmt20/10
Log Hygiene10/10
Framework Coverage5/10

Advanced Controls

Human-in-the-Loop15/10
Agent Identity15/10
Threat Detection20/10

Ecosystem

Prompt Security15/10
Cloud / Platform10/10
LLM Observability10/10
Data Recovery10/10
Compliance Maturity10/10

Unique Capabilities

Post-Exec Verify10/10
Data Flow Gov.10/10
Adversarial Resilience10/10

Market Comparison

VendorWarden ScoreCategory
SharkRouter91/100Tool-Call Gateway
Zenity55/100AI Security Posture
Wiz41/100Cloud Security
Noma Security40/100AI Security Posture
Oasis Security38/100NHI Lifecycle
HiddenLayer34/100ML Security
Market Average28/100
Portkey32/100LLM Gateway
Protect AI (Palo Alto)32/100ML Security
Lasso / Intent Security30/100AI Security
Kong27/100API Gateway
Rubrik26/100Data Recovery

Warden Scoring v4 — 17 dimensions, /100 scale — April 2026

Star on GitHub

MIT Licensed · 100% Open Source

Warden finds the gaps. SharkRouter fills them.

Want to enforce what Warden found?

Explore the Platform →

We use cookies for analytics to understand how visitors use our site. No advertising cookies. Privacy Policy