Show subagent model in agent activity display (alongside tool uses and tokens)

Resolved 💬 2 comments Opened Apr 3, 2026 by jacob-balslev Closed May 12, 2026

Feature Request

When Claude spawns subagents via the Agent tool, the activity display shows the agent description, tool use count, and token count in grey metadata:

— Audit OWASP security skill · 18 tool uses · 61.3k tokens

Request: Add the model name (Opus, Sonnet, Haiku, etc.) to this metadata line, so it reads:

— Audit OWASP security skill · Sonnet · 18 tool uses · 61.3k tokens

or equivalently in the grey section:

— Audit OWASP security skill · 18 tool uses · 61.3k tokens · Sonnet

Motivation

The Agent tool already accepts a model parameter ("sonnet", "opus", "haiku", etc.). As orchestration patterns become more common — parallel dispatch, heterogeneous model routing, cost-aware delegation — it becomes important for the user to see which model is actually being used for each spawned subagent without having to inspect the raw tool call.

This is a transparency request: the same way token counts help users understand cost, the model name helps users understand which capability tier is being invoked and verify that the orchestrator is routing correctly.

Current workaround

Embedding the model in the description string (e.g. "[Sonnet] Solve SH-4643 CI fix"), but this pollutes the human-readable description and is inconsistent.

Suggested implementation

The model field is already present on the Agent tool call. Surfacing it in the activity line would require no API change — only a rendering change in the CLI's agent activity display.

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗