:kotlin_emoji::robot_face: New blog post: <https:/...

# announcements

[Alina Dolgikh [JB]]

K🤖 New blog post: Building AI Agents in Kotlin – Part 2
The next part of the AI agent series is out. This one shows how to extend Koog agents with tools, using a shell execution tool as an example.
It covers:
• The anatomy of a Koog tool (name, description, args, result, execute logic).
• How to build an ExecuteShellCommandTool.
• Handling confirmation, sandboxing concerns, and Brave mode.
• How timeouts work and why they matter for LLMs.
• How the agent’s system prompts change to encourage test-driven behavior.
• Benchmark results after adding execution capability.
If you’re experimenting with agents, this post provides a good overview of how tool design works in Koog.
➡️ Read it here: https://kotl.in/koog-tutorial-2