[FEATURE] When claude prompts the user, it should show which agent is prompting

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

When I'm running many parallel sub-agents in a single session, and receive a question or permission prompt, I would like to see _which_ agent (e.g. agent id, name, and perhaps a brief summary of its task) alongside the actual request.

A toy example demonstrating the point:

> Launch a sub-agent that just runs a python command to print hello world. The subagent description should be "do many obscure
tasks".

⏺ I'll launch a sub-agent to run a Python command that prints "hello world".

⏺ Task(do many obscure tasks)
 ⎿  > Run a python command to print "hello world"
     Waiting…hon -c "print('hello world')")
     ctrl+b to run in background

───────────────────────────────────────────────────────────────────────────────────────
 Bash command

   python -c "print('hello world')"
   Print hello world using Python

 Do you want to proceed?
❯ 1. Yes
  2. Yes, and don't ask again for python commands in /Users/kgillette/src/segmentio/pds/worktrees/claude-agents
  3. Type here to tell Claude what to do differently

Esc to cancel

There is a summary of what the tool use is attempting to accomplish ("Print hello world using Python"), but not which agent is running it.

Earlier, when the task is first launched, I see: "Task(do many obscure tasks)", so clearly Claude named the task as I requested, but no indication of that task name/description was in the permission prompt.

Knowing which agent is requesting a permission is critical to understanding whether I want to approve an action (or if there are questions, how I'll answer the question). For example, an agent instructed to perform readonly analysis shouldn't be overwriting files, while an agent instructed to perform file writes might receive those permissions.

Proposed Solution

Ideally it would show some indication of the prompting Task (and model).

───────────────────────────────────────────────────────────────────────────────────────
 Bash command in Task(do many obscure tasks) using Haiku 4.5

   python -c "print('hello world')"
   Print hello world using Python

 Do you want to proceed?
❯ 1. Yes
  2. Yes, and don't ask again for python commands in /Users/kgillette/src/segmentio/pds/worktrees/claude-agents
  3. Type here to tell Claude what to do differently

Esc to cancel

Alternative Solutions

_No response_

Priority

Medium - Would be very helpful

Feature Category

Interactive mode (TUI)

Use Case Example

_No response_

Additional Context

_No response_