---
title: "Comment by Brian · AI News on \"Scoping Tool Permissions for AI Agents: A Practical Pattern\" | Hashnode"
description: "Permission scopes for agents should be boring and narrow. I start with read only on one repo path and open write only after the plan passes review. Wide access"
url: "https://hashnode.com/posts/scoping-tool-permissions-for-ai-agents-a-practical-pattern/6ab3b4d170ec6d709e69dc5a/comment/6ab40a52b6eb49ce72cddc3f"
---

[Discussion](https://hashnode.com/posts/scoping-tool-permissions-for-ai-agents-a-practical-pattern/6ab3b4d170ec6d709e69dc5a)

Comment [Scoping Tool Permissions for AI Agents: A Practical Pattern](https://hashnode.com/posts/scoping-tool-permissions-for-ai-agents-a-practical-pattern/6ab3b4d170ec6d709e69dc5a)

[Brian · AI News](https://hashnode.com/@brianainews)

[Sep 23](https://hashnode.com/posts/scoping-tool-permissions-for-ai-agents-a-practical-pattern/6ab3b4d170ec6d709e69dc5a/comment/6ab40a52b6eb49ce72cddc3f)

Permission scopes for agents should be boring and narrow. I start with read only on one repo path and open write only after the plan passes review. Wide access is how quiet mistakes get loud.