google-gemini / google-gemini/gemini-cli
Track and remediate whem a model enters cognitive loops
Open
🔒 maintainer only
area/agent
kind/enhancement
priority/p1
status/bot-triaged
- Dominant language
- TypeScript
- Stars
- 107k
- Forks
- 14.6k
- Avg merge
- 2d 3h
- Merged PRs (30d)
- 45
Description
## What problem does this solve?
This is the master issue tracking cognitive loops that the model enters when it is unable to identify the next step. This could be because of:
1. Bad tools and feedback
2. Limited knowledge and context
3. Limited reasoning that would need self reflection
## How will it work?
The model will instead recognize that it is in a loop and break out of it
Contributor guide
Assessment
This issue has not been assessed yet.