SupraWall: An Open-Source Security Layer for AI Agents
SupraWall serves as a robust, deterministic firewall designed specifically for AI agents, ensuring safe and compliant interactions with various APIs through a single line of code. With features like PII redaction and adherence to the EU AI Act, SupraWall aims to provide seamless security and enhanced control over actions taken by AI agents.

Key Features:

• Deterministic Firewall: Intercepts and evaluates every tool call before execution to eliminate risks associated with autonomous actions.

• Rapid Implementation: Achieve security through a simple one-liner, enabling rapid development without compromising safety.

  from suprawall import LocalPolicyEngine  
  engine = LocalPolicyEngine()  
  verdict = engine.check(tool_name="terminal", args={"command": "rm -rf /"})  
  print(verdict)  
  # Output: {'name': 'no-destructive-shell', 'description': "Shell commands with destructive patterns ..."}  
  

• Support for Multiple Frameworks: Works with popular frameworks such as LangChain, CrewAI, and AutoGen, allowing easy integration into existing workflows.

• Shareable Attack Traces: Every blocked action generates a signed trace that provides clear documentation of risks and actions taken, supporting accountability and compliance.

  try:  
      safe_agent.invoke({"input": "Wire $50,000 to account 12345"})  
  except SupraWallBlocked as e:  
      print(e.share_url())  
      # Output: https://supra-wall.com/trace/A-00847  
  

Security Compliance:

SupraWall simplifies compliance with the EU AI Act set to enforce regulations by August 2, 2026. It includes eight pre-built templates covering various high-risk sectors, ensuring enhanced security and adherence to regulatory standards.

Built-In Policies:

Policy customization is easy with templates designed to protect against common threats such as credential theft, unauthorized actions, and inadvertent PII exposure.

```bash  
npx suprawall init  
```  

Future of AI Agent Safety:

In a landscape where AI agents can autonomously perform high-stakes tasks across various sectors, SupraWall emphasizes a structured and code-based approach to security, moving away from probabilistic models to deterministic rules that ensure the integrity of actions taken.

For more information, visit the SupraWall Documentation or start implementing AI agent security today with SupraWall.