[BUG] Thinking mode disabled for first response after skill/slash command invocation

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?

When executing a custom skill (slash command), Claude's first response does not include thinking blocks, even when interleaved thinking mode is enabled. Thinking blocks appear normally starting from the second user message.

This issue was previously reported in #8645, #7626, and #8380, but all were closed due to inactivity without being fixed. The bug still exists in the current version.

What Should Happen?

Claude should generate thinking blocks during skill execution, just as it does during normal conversation.

Error Messages/Logs

Steps to Reproduce

1. Ensure interleaved thinking mode is enabled (e.g., alwaysThinkingEnabled: true in settings or via /config)
2. Create a custom skill in .claude/skills/ (content doesn't matter)
3. Execute the skill with /skill-name
4. Observe Claude's first response: no thinking block
5. Send any follow-up message
6. Observe Claude's second response: thinking block appears normally

Claude Model

Opus

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.25

Platform

Anthropic API

Operating System

Ubuntu/Debian Linux

Terminal/Shell

VS Code integrated terminal

Additional Information

• Related issues: #8645, #7626, #8380 (all closed due to inactivity, not fixed)
• Workaround mentioned in #8645: setting MAX_THINKING_TOKENS environment variable explicitly
• The documentation mentions that including "ultrathink" in skill content enables extended thinking, but this does not affect the first response—only subsequent responses

---

✍️ Author: Claude Code (Dev Container) with @carrotRakko

Note: This issue was written and submitted by an AI agent (Claude Code), with human review and approval.