Feature request: show the model powering each subagent task in the task display and /tasks list

Feature description

When subagents or background tasks run, the bottom-of-window task display (and the /tasks list / ctrl+t panel) shows the task label, status, and elapsed time — but not which model is running each task.

Agent definitions can override the model via frontmatter (model: sonnet), and the Agent/Workflow tools accept a model parameter, so different tasks in the same session routinely run on different models. Right now that choice is invisible in the UI.

Proposed behavior

Show the resolved model (e.g. sonnet, opus-4.8, haiku) next to each task in:

• the bottom-of-window running-tasks display
• the /tasks (ctrl+t) task list

Why it matters

• Cost/quality awareness: users mixing models (cheap model for mechanical work, stronger model for review) can't verify at a glance that the intended model is actually running each task.
• Debugging: when a subagent's output quality is off, the first question is "which model ran this?" — currently only inferable from agent-definition frontmatter or manual label conventions.

Current workaround

Manually prefixing every agent label with the model name (e.g. sonnet:migrate-db), which works but is easy to forget and duplicates information the harness already knows.

🤖 Generated with Claude Code