[BUG] /effort reads stale state — always reports 'max' regardless of changes via /model

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?

/effort always reports max (Maximum capability with deepest reasoning (Opus 4.6 only)) regardless of what effort level was set via /model. The two commands are out of sync.

Reproduced across:

• Model switches (Opus → Sonnet → back)
• All effort levels (low, medium, high, max)
• Plugin reloads (/reload-plugins)
• Two different terminals (Warp standalone, JetBrains Rider 2025.3.3 built-in terminal via tmux)

What Should Happen?

/effort should reflect the effort level that was set via /model. If I select "Opus 4.6 with medium effort" in /model, then /effort should report medium.

Error Messages/Logs

❯ /effort
└  Current effort level: max (Maximum capability with deepest reasoning (Opus 4.6 only))

❯ /model
└  Set model to Opus 4.6 (1M context) (default) with medium effort

❯ /effort
└  Current effort level: max (Maximum capability with deepest reasoning (Opus 4.6 only))

❯ /model
└  Set model to Opus 4.6 (1M context) (default) with low effort

❯ /effort
└  Current effort level: max (Maximum capability with deepest reasoning (Opus 4.6 only))

❯ /model
└  Set model to Sonnet 4.6 with medium effort

❯ /effort
└  Current effort level: max (Maximum capability with deepest reasoning (Opus 4.6 only))

❯ /reload-plugins
└  Reloaded: 22 plugins · 41 commands · 49 agents · 21 hooks · 0 plugin MCP servers · 4 plugin LSP servers

❯ /model
└  Set model to Opus 4.6 (1M context) (default) with high effort

❯ /effort
└  Current effort level: max (Maximum capability with deepest reasoning (Opus 4.6 only))

Steps to Reproduce

1. Start a Claude Code session (any model)
2. Run /effort — note the reported level
3. Run /model — select any model with a different effort level (e.g., medium)
4. Confirm /model output shows the new effort level
5. Run /effort again
6. Observe: /effort still reports the original level, ignoring the change made by /model

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

Unknown

Claude Code Version

2.1.76 (Claude Code)

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Reproduced in both:

• Warp (standalone)
• JetBrains Rider 2025.3.3 built-in terminal (Build #RD-253.31033.136) via tmux