What's Agent Doing: Open-Source Tool for Real-Time Visualization of Claude's AI Agent Behavior

Summary:Tzafrir has released an open-source tool called 'What's Agent Doing,' which provides real-time visualization of AI agent behavior using Claude Code's function hooks. The tool displays the current step, tool calls, background agent activities, and history of actions, helping users understand Claude's decision-making process in complex tasks. The project is licensed under MIT and does not collect any user data or network information, offering a transparent solution for monitoring AI agent behavior

---

Features and Capabilities

• Real-Time Behavior Visualization: Displays Claude's current step, tool call status, and background agent activities above the prompt, providing a clear view of the AI's decision-making process.
• Detailed Operation Logs: Records descriptions, durations, and outcomes of each step, allowing users to expand and view detailed history.
• Tool Call Status Indicator: Highlights tool calls waiting for user approval with a yellow status bar, offering intuitive feedback on the current operation state.
• Background Agent Activity Tracking: Generates separate activity rows for each background agent and retains records after the operation ends.

Technical Highlights

• Non-Intrusive Design: The tool only reads Claude's stream data, tool calls, and permission requests without altering Claude's behavior, ensuring transparency and security.
• Lightweight Implementation: Requires no network connection or telemetry, and is licensed under MIT, allowing developers to freely use and modify the tool.
• Efficient Integration: Utilizes Claude Code's new function hooks for seamless integration with Claude, providing a smooth user experience.

Industry Impact and Developer Recommendations

• Enhancing AI Transparency: As AI agents are increasingly applied to complex tasks, making their behavior transparent is crucial. What's Agent Doing offers an effective solution for developers, helping to build trust in AI systems.
• Facilitating AI Debugging and Optimization: Real-time visualization of AI agent behavior allows developers to debug and optimize AI models more efficiently, improving their performance in practical applications.
• Driving AI Tool Innovation: The open-source nature of the tool provides new ideas for the AI community, encouraging more developers to engage in the development of AI behavior monitoring and visualization tools.

Developer Recommendations

• Integrate with More AI Models: Consider extending the tool to support other AI models in the future to broaden its application scope.
• Enhance Visualization Features: Introduce richer visualization elements such as charts and graphs to further improve user understanding of AI behavior.

Conclusion

What's Agent Doing is an innovative open-source tool that provides powerful debugging and monitoring capabilities for AI developers through real-time visualization of Claude's agent behavior. Its open-source nature and lightweight design make it easy to integrate and extend, bringing new opportunities and challenges to the AI community.

---

Source: GitHub AI Trending Releases (2026-10-02)

— END —

Tags: #Claude #Open-Source Tool #AI Behavior Monitoring #Agentic #Visualization