Show model and thinking level in run-once subagent output

[rochecompaan]

Summary

Update the run-once subagent output to show the resolved model and thinking level alongside the agent name.

Current output

🤖 subagent (agent=reviewer)
🤖 subagent (agent=worker)

Desired output

🤖 subagent (agent=reviewer, model=gpt-5.6-sol, thinking=xhigh)
🤖 subagent (agent=worker, model=gpt-5.6-terra, thinking=medium)

Acceptance criteria

• Each run-once subagent output line includes the agent name, resolved model, and resolved thinking level.
• The displayed model and thinking level reflect the values actually used for that subagent run.
• The output retains the existing 🤖 subagent (...) format, adding the new fields as comma-separated values.