Description

Clawsec is an open-source AI agent security layer that protects against prompt injection, data exfiltration, secret leaks, destructive system commands, and unauthorized tool execution.

Built for modern LLM-powered agents, Clawsec acts as a real-time guardrail system that intercepts and validates tool calls before they execute — adding security controls to frameworks that ship with power but no protection.

With under 5ms latency overhead and simple YAML-based configuration, Clawsec enables developers to enforce security policies, restrict filesystem access, prevent sensitive data exposure, and block unsafe actions in production AI environments.

If you’re building autonomous AI agents, AI workflows, or LLM applications with system-level access, Clawsec provides lightweight, configurable, open-source protection.