Where does the bug appear (feature/product)?
Grok Bot
Describe the Bug
On Grok Bot Auto-review / Always allow approval cards, the Details body sometimes shows an internal truncation marker as if it were part of the approval copy.
In this card, Details showed a literal “[3397 chars omitted]” in the middle of a routine prompt. The header is also cut off: “Use automation_write to update the LinkedIn watch automation when directed by trusted use…”
Expected: the full text, or a clean user-facing truncation. Not an internal omit placeholder visible to the user.
Steps to Reproduce
- Use update_state to update an existing routine whose prompt is long (example: “LinkedIn watch”, weekdays 9:59 AM and 3:59 PM).
- When the Always allow card appears for automation_write, read the Details body.
- Observe a literal “[N chars omitted]” marker in the middle of the prompt, and a truncated header.
Expected Behavior
Approval copy should show the real text, or a clean truncation. Internal markers like “[3397 chars omitted]” must not appear as user-visible copy.
Screenshots / Screen Recordings
Operating System
Other
Version Information
Grok Bot: unknown - no About version shown in this session
Conversation id: unknown - not shown
Surface: desktop
Additional Information
Related but not the same: Grok Bot: let the user set the timeout on Auto-review approval cards (timeout on these cards, not leaked omit markers).
Searched Grok Bot bug reports for “chars omitted” / Auto-review truncation; no match.
Does this stop you from using Cursor
No - Cursor works, but with this issue