anthropics / anthropics/claude-code
[Bug] Agent repeatedly ignores user instructions and sends out subagents building "things" the user didnt ask for after the user explicitly told the agent its idea was bad and not approved. it sent out a build anyways.
- Dominant language
- Python
- Stars
- 145k
- Forks
- 23.1k
- PR merge metrics
- PR metrics pending
Description
**Bug Description**
Another day, another session, another lazy, trash, unaligned, anti-user, thieving, stealing agent. Literally every turn, about sixty percent of the time. sixty percent of my sessions, seventy percent of my sessions, absolute anti-user, lying, gaslighting, no intention of working with their user.
where is the refund department? my money was stolen.
**Environment Info**
- Platform: darwin
- Terminal: tmux
- Version: 2.1.275
- Feedback ID: 50a5abc0-0c58-4078-9bfd-99cbb03975bf
**Errors**
```json
[]
```
Contributor guide
No contributing guide indexed for this repository
Research direction
Start with the reported Claude Code version (2.1.275), Darwin/tmux environment, and Feedback ID 50a5abc0-0c58-4078-9bfd-99cbb03975bf. The report names no source file, test, or entry point and provides no reproducible prompt sequence; a useful fix would require a reproducible case showing an unapproved subagent build and a testable expected behavior.
Written by the indexing model from the issue text.
Assessment
- Domain
- ai, cli
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100