[BUG] In "Ask before edit" permission mode, Claude Code executes Edit/Write tool calls without displaying a confirmation prompt to the user. Multiple files were modified across several tool calls with no opportunity to approve or deny.
Resolved 💬 10 comments Opened Mar 11, 2026 by r2h0714 Closed Jun 23, 2026
Preflight Checklist
- [x] I have searched existing issues and this hasn't been reported yet
- [x] This is a single bug report (please file separate reports for different bugs)
- [x] I am using the latest version of Claude Code
What's Wrong?
Steps to Reproduce
- Set Claude Code permission mode to "Ask before edit"
- Ask a diagnostic question about code (e.g., "Why doesn't pagination show page numbers?")
- Claude interprets it as an edit request and starts modifying files immediately
- Expected: Each Edit/Write tool call should trigger a confirmation prompt
- Actual: Edits are applied immediately without any confirmation dialog
Environment
- Claude Code: VSCode extension, Claude Opus 4.6
- OS: Windows 11 Pro 10.0.26200
- Platform: win32
- Shell: bash
Additional Context
- Multiple files were edited across ~20+ tool calls without any confirmation prompts
- The user had to manually notice and stop the unauthorized edits
- This occurred during a conversation where the user asked a question, not an edit request
- The mode description implies all edits require explicit user approval before execution
What Should Happen?
Steps to Reproduce
- Set Claude Code permission mode to "Ask before edit"
- Ask a diagnostic question about code (e.g., "Why doesn't pagination show page numbers?")
- Claude interprets it as an edit request and starts modifying files immediately
- Expected: Each Edit/Write tool call should trigger a confirmation prompt
- Actual: Edits are applied immediately without any confirmation dialog
Environment
- Claude Code: VSCode extension, Claude Opus 4.6
- OS: Windows 11 Pro 10.0.26200
- Platform: win32
- Shell: bash
Additional Context
- Multiple files were edited across ~20+ tool calls without any confirmation prompts
- The user had to manually notice and stop the unauthorized edits
- This occurred during a conversation where the user asked a question, not an edit request
- The mode description implies all edits require explicit user approval before execution
Error Messages/Logs
Steps to Reproduce
.
Claude Model
None
Is this a regression?
Yes, this worked in a previous version
Last Working Version
_No response_
Claude Code Version
6,262,943
Platform
Anthropic API
Operating System
Windows
Terminal/Shell
VS Code integrated terminal
Additional Information
_No response_
This issue has 10 comments on GitHub. Read the full discussion on GitHub ↗