anthropics / anthropics/claude-code

Desktop app Code tab: effort picker changes never reach the running session (model receives max every turn)

Open
#95,638 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

area:desktop area:tui bug platform:windows
Dominant language
TypeScript
Stars
147k
Forks
24k
PR merge metrics
PR metrics pending

Description

Environment: Claude desktop app, Code tab, Windows 11 Pro 10.0.26200, model claude-fable-5-1, auto permission mode.

Steps:

  1. Start a Code tab session.
  2. During the session, change the effort picker: Extra -> High -> Low (each followed by a new user turn).

Actual: on every subsequent turn the model still receives a reasoning_effort marker of max. The picker label and the value delivered to the session disagree, and no change to the picker has any effect for the life of the session.

Expected: the selected effort applies from the next turn, as the picker implies.

Notes:

  • The picker showed Extra when the session already reported receiving max, so the mismatch may exist from session start, not only after a change.
  • The session-management tool set_session_effort refuses to change the current session by design, so there is no in-session workaround.
  • Impact: turns run at maximum effort regardless of the setting, making small tasks very slow.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No source file, test, or entry point is named. First reproduce the effort-picker sequence in a Windows 11 desktop Code tab session and trace where the selected value is sent on each new turn; done means Low, High, and Extra selections reach the running session from the next turn onward.

Written by the indexing model from the issue text.

Assessment

Tech stack
typescript
Domain
desktop
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.