MCPFast / Tools / Agentmetry: Local SIEM for AI coding agents with MITRE audit
Agentmetry is a local SIEM for AI coding agents, providing flight recording and MITRE-mapped auditing for sequence detection and DLP.
View on GitHub→Agentmetry provides a robust, local Security Information and Event Management (SIEM) solution specifically designed for AI coding agents. It offers critical visibility and auditing capabilities for developers working with AI agents, ensuring better security, compliance, and debugging. By capturing detailed "flight recordings" of agent activity and mapping it to the MITRE ATT&CK framework, Agentmetry enables precise sequence detection and Data Loss Prevention (DLP) for your AI development workflows.
Agentmetry acts as a central logging and analysis platform for your AI coding agents. It intercepts and records all interactions, decisions, and code executions performed by these agents. This detailed log data is then processed to identify suspicious patterns, potential security threats, and compliance violations. The MITRE ATT&CK mapping allows for a standardized approach to understanding and categorizing agent behaviors, making it easier to detect advanced threats and ensure responsible AI development.
Agentmetry is an essential tool for AI developers , security engineers working with AI systems, and DevOps teams responsible for deploying and managing AI agents. It is particularly beneficial for projects that handle sensitive data, require strict compliance, or involve complex AI agent interactions where understanding the exact sequence of operations is crucial for troubleshooting and security auditing. If you are building, testing, or deploying AI coding agents, Agentmetry provides the necessary visibility to ensure their secure and reliable operation.