Assistant text blocks are dropped and not rendered or persisted in the session transcript, leading to user confusion and loss of information.
Assistant text blocks are dropped and not rendered or persisted in the session transcript, leading to user confusion and loss of information.
Problem statement normalized from the linked evidence
CANDIDATE VIEW · NOT A PUBLISHED OPPORTUNITY
01
Qualification checklist
Gate decision: Aug 15, 2026
✓
At least 3 independent users3 / 3
MET
✓
At least 2 independent discussions3 / 2
MET
✓
At least 1 supporting signalPersistence
MET
02
Observed signals
Users3
Discussions3
Mentions3
Active days3
Source coverageGitHub
TrendUNKNOWN
First seenJun 03, 2026
Last seenJul 04, 2026
Why the gate is holding
No rejection reason is currently recorded.
Qualification passed. Commercial analysis may still recommend REJECT; only non-rejected, evidence-backed analyses can be published as opportunities.
03
Source evidence
3 evidence records
github · HIGH · Jul 04, 2026
Assistant text blocks silently dropped when followed by more thinking in the same turn — never rendered, missing from transcript JSONL (2.1.201, adaptive thinking)
**Environment:** Claude Code 2.1.201, macOS (darwin 25.5), model `claude-fable-5` (interleaved/adaptive thinking), terminal TUI (also reproduced in a claude.ai/code bridged child session, `CLAUDE_CODE_CHILD_SESSION=1`). **Summary:** Assistant text emitted mid-turn — a `text` block followed by another `thinking` block and/or further tool calls in the same turn — is frequently neither rendered on any surface nor persisted to the session transcript JSONL. The affected assistant message is persisted as `[thinking, thin…
Pre-tool-call assistant text never emitted (prose stays in thinking) — onset ~2026-06-04, not CLI-version-bound
## Summary Since ~v2.1.162 (still present on 2.1.165), assistant **text blocks are silently dropped from the session transcript** (`~/.claude/projects/…/ .jsonl`) whenever the model emits another (interleaved) thinking block after the text in the same turn: | generated turn | persisted to JSONL | |---|---| | `[thinking, text, tool_use…]` | `[thinking, text, tool_use…]` ✅ | | `[thinking, text]` (final reply) | `[thinking, text]` ✅ | | `[thinking, text, thinking, tool_use…]` | `[thinking, thinking, tool_use…]` ❌ text…
[Bug] Background (daemon) sessions drop assistant text blocks from transcript when response mixes text with tool_use (regression 2.1.160 → 2.1.161)
## Environment - Claude Code: 2.1.161 (npm install, Linux) - OS: Ubuntu 24.04 (kernel 6.17) - Session type: background job, daemon backend (`"backend": "daemon"`, `sessionKind: "bg"` in the transcript records) - Model: claude-opus-4-8 ## Summary In background (daemon) sessions, assistant **text blocks are not persisted** to the session transcript (`~/.claude/projects/ / .jsonl`) whenever the response also contains tool_use blocks and two thinking blocks (interleaved thinking). The transcript ends up with only `thin…