[BUG] v2.1.69: thinking summaries empty in transcript and TUI — undocumented behavior change

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

After upgrading from v2.1.68 to v2.1.69, thinking block content becomes empty. The type: "thinking" blocks are present in the session transcript (JSONL), but the thinking field is an empty string (0 characters). No thinking summaries are displayed in the TUI.

Downgrading to v2.1.68 immediately restores normal behavior — thinking summaries appear in both the TUI and the transcript.

This behavior change is not documented in the v2.1.69 release notes.

What Should Happen?

Thinking summaries should be recorded in the transcript and displayed in the TUI, as they are in v2.1.68.

If this is an intentional change, it should be documented in the release notes with information on how to opt out (e.g., showThinkingSummaries setting).

Error Messages/Logs

# v2.1.68 session — thinking works
$ grep -c '"thinking"' session-v2.1.68.jsonl
19

$ python3 -c "..." # extract thinking content length
line 2: thinking length=437 chars  (content present)
line 5: thinking length=892 chars  (content present)
...

# v2.1.69 session — thinking empty
$ grep -c '"thinking"' session-v2.1.69.jsonl
3

$ python3 -c "..." # extract thinking content length
line 2: thinking length=0 chars  (EMPTY)
line 9: thinking length=0 chars  (EMPTY)
line 13: thinking length=0 chars  (EMPTY)

Both sessions used the same prompt, same model (Opus 4.6), same environment. Only the CLI version differs.

Steps to Reproduce

1. Install v2.1.68: npm install -g @anthropic-ai/claude-code@2.1.68
2. Start a new interactive session with Opus 4.6
3. Send any prompt — observe thinking summaries in TUI and verify in transcript JSONL
4. Install v2.1.69: npm install -g @anthropic-ai/claude-code@2.1.69
5. Start a new interactive session with Opus 4.6
6. Send the same prompt — no thinking summaries displayed, transcript shows empty thinking blocks

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

2.1.68

Claude Code Version

2.1.69

Platform

Anthropic API

Operating System

Ubuntu/Debian Linux

Terminal/Shell

iTerm2

Additional Information

• Related: #24285 reports a similar symptom ("Can't see Claude's thinking anymore") but was observed at v2.1.45→v2.1.47, likely a different root cause
• Related: #8477 requests an option to always show thinking

✍️ Author: Claude Code with @carrotRakko (AI-written, human-approved)