![]()
Codenotary, leaders in assuring safe and secure use of AI, today released an update to its free, open source software AgentMon Live that expands its coverage to OpenAI Codex and AWS Kiro helping developers to identify problems requiring human attention and reduce unnecessary token consumption.
Initially made available for Anthropic’s Claude Code, AgentMon Live now provides centralized visibility across three widely used AI-assisted coding agents – Claude Code, OpenAI Codex and AWS Kiro enabling developers to inspect their actions, identify execution failures and better understand how autonomous coding tools operate across their development workflows.
As organizations adopt AI coding agents to accelerate software development, developers face growing challenges understanding what autonomous agents actually do. AI agents can execute shell commands, modify source code, access files, interact with external services and delegate tasks to other agents. Without sufficient visibility, developers may struggle to identify failures, investigate unexpected behavior or determine whether an agent has successfully completed its assigned work.
“AI coding agents are evolving extremely quickly, and developers are adopting these new tools almost as soon as they become available,” said Dennis Zimmer, chief technology officer and co-founder of Codenotary. “Monitoring and security cannot afford to lag behind that rapid adoption. Our goal with AgentMon Live is to provide visibility across the tools developers use, without forcing them into a particular AI ecosystem.”
Unified AI Coding Agent Monitoring Across Claude Code, Codex and Kiro
Observing AI coding agent activity from a single monitoring interface reduces the need to inspect separate logs or rely exclusively on information reported by individual agents.
The AgentMon Live update includes:
- OpenAI Codex monitoring – AgentMon Live reads Codex’s native session files, allowing developers to inspect Codex activity alongside Claude Code sessions. Developers can also install AgentMon Live as a Codex plugin.
- AWS Kiro Crew monitoring – Developers can view active Kiro Crew conversations alongside sessions from other supported AI coding environments. AgentMon Live can also run as a Kiro Crew application, presenting monitoring information directly beside the agent conversation.
- Remote and containerized agent monitoring – Support for agents running in virtual machines, containers and headless environments extends visibility beyond local development workstations.
These capabilities help developers maintain a consistent view of AI agent behavior as they move between coding tools, projects and execution environments.
AgentMon Live Improves Visibility Into AI Agent Execution and Task Completion
In addition to the new integrations, AgentMon Live introduces monitoring and diagnostic capabilities to make autonomous software development more transparent and easier to troubleshoot.
- Step-by-step AI agent execution monitoring – Developers can inspect model activity, shell commands, file operations, sub-agent execution, network interactions, memory access and hooks in execution order. This provides a clearer record of how an agent approaches and performs a task.
- Visibility across sessions and environments – A unified web interface presents supported AI coding agent sessions, including activity originating in containers, virtual machines and remote development environments.
- AI agent completion verification – Enhanced checks help identify failed tests, unfinished tasks and unsupported claims of successful execution before an agent completes its turn. These checks provide additional evidence for evaluating reported outcomes.
- Integrated developer feedback – Developers submit feedback to the AI coding agent via AgentMon Live and then review the behavior and performance of the AI coding agent.
- Resource usage and performance insights – Visibility into token consumption, execution time and recurring problems helps developers investigate inefficient workflows and diagnose issues across sessions.
Open-Source AI Agent Monitoring Available Immediately
AgentMon Live is available as open-source software under the Apache 2.0 license, allowing developers to inspect, use and extend the monitoring tool without requiring a proprietary enterprise monitoring platform.
The AgentMon Live update is available immediately with OpenAI Codex and AWS Kiro integrations. Source code and installation instructions are available through the project’s GitHub repository. For additional information about AgentMon Live and its monitoring capabilities, visit https://codenotary.com/blog/agentmon-live-turn-steps-feedback-and-sessions-everywhere.
About Codenotary
Used by hundreds of customers worldwide – including the world’s leading banks, governments, and defense organizations – Codenotary delivers technology that enables secure and trusted agentic networks in the modern organization. For more information, visit www.codenotary.com.
View source version on businesswire.com: https://www.businesswire.com/news/home/20261117115093/en/
Media gallery