Feedback
AI assistant tool use now stops safely at generous limits
- Status
- Fixed
- Fixed in
- v26.7.5
- Source discussion
- -
- Last updated
- 2026-07-25
Upvotes
0 upvotes
Uses your Objo forum account.
Public summary
Objo Studio’s AI assistant will apply generous safety ceilings to model-requested tool calls so malformed or looping responses cannot trigger an effectively unbounded sequence of operations. Normal codebase exploration remains supported with up to 64 tool calls in one provider round and 256 across one user request.