anthropics / anthropics/claude-code
[Bug] Anthropic API Error: Overly aggressive safety filtering on legitimate system signal operations
- Dominant language
- Python
- Stars
- 145k
- Forks
- 23.1k
- PR merge metrics
- PR metrics pending
Description
**Bug Description**
I got an "Opus 5 (1M context)'s safeguards flagged this message." "Details: [cyber]" which is a false-positive, I am working on my own systems, and doing development where I am sending signals (SIGBREAK) to a task I own. This should not have been flagged under cyber anything.
**Environment Info**
- Platform: linux
- Terminal: gnome-terminal
- Version: 2.1.267
- Feedback ID: b7638dee-de8b-471d-90a0-a57537c25031
**Errors**
```json
[]
```
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by reviewing feedback ID b7638dee-de8b-471d-90a0-a57537c25031 and reproducing the report in Linux with gnome-terminal while sending SIGBREAK to an owned task. Done would mean this legitimate system-signal operation is no longer classified as a cyber safety violation.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- linux
- Domain
- api, cli, security
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100