[BUG] Tool call / agent card renders above its own introduction text — message order reversed in TUI

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 Claude says something like "Let me invoke the XXX skill" or "I'll spawn a sub-agent to help" and then makes a tool call or spawns an agent in the same turn, the tool call / agent card renders ABOVE the introduction text instead of below it.

For example, Claude first displays: "I'll help you search with the exa-search skill."

Then the tool call card appears — but it renders ABOVE that sentence, pushing it down. The visual order becomes:

[Tool call card] ← should be BELOW
"I'll help you search..." ← should be ABOVE

The text is not lost, it's still there — it's just rendered in the wrong position. This makes conversation flow confusing because the explanation appears after the action it's explaining.

This happens on macOS (all terminals: Ghostty, iTerm2, Terminal.app) but NOT on Windows. Both tool calls and sub-agent spawns are affected.

What Should Happen?

The introduction text should render ABOVE the tool call / agent card, in the same order the model emits them. Within a single assistant message, the text always comes before the tool_use block, so the TUI should preserve that natural order:

"I'll help you search with the exa-search skill." ← text first
┌──────────────────────────┐
│ 🔧 Skill: exa-search │ ← tool card below
└──────────────────────────┘

Error Messages/Logs

Steps to Reproduce

1. Open Claude Code on macOS in any terminal (Ghostty, iTerm2, Terminal.app)

1. Type any prompt that causes Claude to say it will invoke a skill or spawn a sub-agent, then actually do it. Examples:

• Ask Claude to "search for the latest React documentation online"
• Ask Claude to "do a deep research on WebAssembly component model"
• Any prompt that triggers Claude to use the Skill tool or Agent tool

1. Observe the conversation display:

• Claude's introductory text appears first (e.g., "I'll invoke the exa-search skill for you...")
• Then the tool call / agent card renders — but it appears ABOVE the introductory text

1. Compare with the expected order:

• Expected: text above, tool card below
• Actual: tool card above, text below

1. (Optional) Try the same prompt on Windows — the bug does NOT reproduce there, confirming it's a macOS-specific TUI rendering issue.

Claude Model

None

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.195

Platform

Other

Operating System

macOS

Terminal/Shell

iTerm2

Additional Information

_No response_