[BUG] Desktop app Remote Control: slash-command autocomplete force-corrects /clear to nearest match instead of allowing literal input

Resolved 💬 0 comments Opened Jul 12, 2026 by samdawes Closed Jul 13, 2026

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?

In the Claude Desktop app (macOS) with Remote Control active into a Claude Code CLI session running on a separate machine, typing /clear into the Remote Control input box triggers the autocomplete overlay to fuzzy-match it against Desktop's own known slash commands and force-correct it toward an unrelated match (in my case, /clear → /simplify). There is no way to keep the literal /clear text and submit it — the command never reaches the underlying CLI session at all. This appears related to #28379, #68102, and #68252, which describe recognized slash commands being routed to the model as plain text instead of executing — but this is a distinct, earlier-stage bug: the autocomplete UI prevents typing/submitting the literal command text in the first place, so it never even reaches the plain-text-fallback stage those issues describe.

What Should Happen?

Typing an unrecognized-but-valid-looking slash command — especially one the target CLI session supports, even if Desktop doesn't have it in its own command list — should be typeable and submittable as literal text, not silently coerced to a fuzzy autocomplete match.

Error Messages/Logs

None — no error is shown; the input is silently force-corrected in the UI.

Steps to Reproduce

  1. Open the Claude Desktop app on macOS.
  2. Start/join a Remote Control session into a Claude Code CLI session running on a separate machine.
  3. In the Remote Control input box, type /clear.

Observe: the autocomplete overlay fuzzy-matches this to an unrelated known command (/simplify in my case) and force-corrects the input toward that suggestion.

There is no way to submit the literal /clear text — it cannot reach the CLI session.

Claude Model

Not sure / Multiple models

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.207

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

<img width="277" height="135" alt="Image" src="https://github.com/user-attachments/assets/8076ef6d-524e-4f8e-8ec4-796f3c3022e4" />
<img width="275" height="141" alt="Image" src="https://github.com/user-attachments/assets/8beab665-d420-4cfd-8922-421e8b6fddbc" />

View original on GitHub ↗