Security Middleware for AI Agents. Zero Trust Architecture. Prevent prompt injection, billing disasters, and unauthorized access with one line of code.
Quick Start
# Install the middleware SDK $ npm install molt-guard
// SYSTEM_LOGIC_FLOW
Zero Trust by Design
🤖 AI Agent
🛡️ Molt Guard (Interceptor)
📦 Guard SDK
🧠 Guard Server
policy engine
🔐 RBAC
📝 Moderation
💰 Budget
⛔ DENIED
🔑 Token Vendor
💳 Stripe
☁️ AWS
1. Call Tool
2. Intercept
3. Evaluate
4. Vend JIT Token
5. Execute
// CAPABILITIES
What It Does
🔐
JIT Token Vending
Static keys are dead. We issue ephemeral credentials for AWS and Stripe that live for 5 minutes and self-destruct.
💰
Budget Controls
Stop infinite loops from draining your wallet. Hard caps on API spend per hour, day, or request.
📝
Content Moderation
Automatically detect PII, offensive language, or negative sentiment before it hits the API.
🧠
Intent Analysis
We analyze the intent of the agent's request. If it tries to delete data when it should only read, we block it.
📜
Audit Logging
Full visibility. "Agent X accessed Stripe because 'User asked for refund'." Complete decision trail.
🔌
Universal Wrapper
Wrap any tool or function with guard.protect(). Works with any TS/JS agent framework.
// REAL_TIME_AUDIT
The Guard Server evaluates every request against your policies in milliseconds.