Assistant text preceding an AskUserQuestion tool call is not shown to the user
Status Fixed / completed
Maintainer reply None cached
Workaround ✓ Mentioned in description ↑
Activity 0 comments · opened Aug 6, 2026 · closed Aug 7, 2026
Description
When the assistant emits explanatory text in the same turn as (immediately before) an AskUserQuestion tool call, that text is sometimes not displayed — the user only sees the multiple-choice picker. Context the question depends on (in our case, a list of trade-offs the user was being asked to weigh) is silently lost, and the user has to ask for it to be repeated.
Environment
- Claude Code on Windows 11 (desktop/CLI session)
- Model: claude-fable-5
- Observed 2026-08-06
Repro sketch
- Have the assistant produce a turn of the form: several paragraphs of markdown analysis, then an
AskUserQuestioncall whose question refers to "the consequences above". - The user-facing UI renders the picker, but the preceding paragraphs are not visible to the user ("they seem to get eaten sometimes with the multiple choice thing" — user's words).
Expected
Text emitted before the tool call renders normally above the picker.
Workaround
Embedding the full context inside the question field of the picker itself.