[MODEL]

Resolved 💬 3 comments Opened Nov 27, 2025 by AlanKharebov Closed Jan 27, 2026

Preflight Checklist

  • [x] I have searched existing issues for similar behavior reports
  • [x] This report does NOT contain sensitive information (API keys, passwords, etc.)

Type of Behavior Issue

Claude modified files I didn't ask it to modify

What You Asked Claude to Do

Plan mode is too aggressive and interrupts normal conversation

  • Gets triggered accidentally during casual conversation
  • Forces rigid multi-phase workflow (agents, synthesis, etc.) even for simple tasks
  • Hard to exit once triggered
  • Old behavior was simpler and more flexible

What Claude Actually Did

First point:
Plan mode is too aggressive and interrupts normal conversation

  • Gets triggered accidentally during casual conversation
  • Forces rigid multi-phase workflow (agents, synthesis, etc.) even for simple tasks
  • Hard to exit once triggered
  • Old behavior was simpler and more flexible

Second:

Expected Behavior

FIRST POINT:
Plan mode is too aggressive and interrupts normal conversation

  • Gets triggered accidentally during casual conversation
  • Forces rigid multi-phase workflow (agents, synthesis, etc.) even for simple tasks
  • Hard to exit once triggered
  • Old behavior was simpler and more flexible

SECOND POINT:

  • just to report issue, I need to 3-4 screen forms, answer questuon and then report.
  • are you not top AI company, lest people quiclky report, dirretcl;y ion claude code chat:

Files Affected

Permission Mode

Accept Edits was ON (auto-accepting changes)

Can You Reproduce This?

Yes, every time with the same prompt

Steps to Reproduce

Plan mode is too aggressive and interrupts normal conversation

  • Gets triggered accidentally during casual conversation
  • Forces rigid multi-phase workflow (agents, synthesis, etc.) even for simple tasks
  • Hard to exit once triggered
  • Old behavior was simpler and more flexible

Second:

  • just to post answer, I need to open 3-4 pages, answer questuons
  • are you not top AI company, why not just give ability top submit issue directly in Claude Code chat.

and the you can figutre out, is it new, did I search existing issue. really. this was 20-30 yaers ago,

  • I suggest even ask peaple fill request by paer and send by mail,
  • I love Antripics an wish you best, but this is very disapointing,

Claude Model

Sonnet

Relevant Conversation

Impact

Critical - Data loss or corrupted project

Claude Code Version

123

Platform

Anthropic API

Additional Context

_No response_

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗