Actualis

Actualis is a specialized tool designed to provide crucial insights into the activity of your coding agents. It processes existing agent logs to deliver clear answers across three key areas: • Exposure: Identifies critical credentials and secrets exposed during agent operations, detailing their usage and associated projects. • Activity: Audits every shell command actually executed, distinguishing between proposed and truly ran commands, and flagging destructive or network-reaching actions. • Cost: Tracks and attributes token spend by project, branch, model, and ticket, providing a detailed breakdown of expenditures including cache performance. This robust analysis helps development teams understand and manage the security posture and operational efficiency of their coding environments. It ensures that sensitive information is not inadvertently shared with models and provides a transparent view of agent actions. By reading existing log files locally and without network interaction, Actualis guarantees privacy and security, offering a read-only, non-intrusive monitoring solution. It's built for developers and security professionals who need to verify what their coding agents are doing, how much it's costing, and whether any sensitive data has been compromised. Compatible with macOS, Linux, and Windows, it operates using only the Python standard library for minimal overhead and easy deployment.