anthropics / anthropics/claude-code

[BUG] Android app renders slash-command messages without their arguments

Ouverte
#88,018 0 commentaires 0 réactions 0 personnes assignées Voir sur GitHub
area:ui bug platform:android stale
Langage dominant
Python
Étoiles
145k
Forks
23.1k
Métriques de merge des PR
Métriques de PR en attente

Description

### Preflight Checklist

- [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) 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 Android app, a user message that invokes a slash command with arguments renders as the command name only. The argument text is dropped from the displayed message.

Example: `/echo Will this show up in the Android app?` renders on Android as just `/echo` (see `cc-android.png`). The same message in the same session, viewed in the terminal CLI, renders in full as `/echo Will this show up in the Android app?` (see `cc-terminal.png`).

This is display-only, not a transport problem: the arguments still reach the model. `/echo` is a skill that replies with its argument text verbatim, and the reply on Android correctly reads "Will this show up in the Android app?" - so the app received the arguments and simply did not render them in the user bubble.

Impact: when reviewing a session from the phone, there is no way to see what a slash command was actually asked to do. Every invocation of the same command looks identical.

**cc-terminal.png**
Image

**cc-android.png**
Image

### What Should Happen?

The user message should render the full invocation, including arguments - matching the terminal, e.g. `/echo Will this show up in the Android app?`

### Error Messages/Logs

```shell
None - no error is raised; the arguments are silently omitted from the rendered message.
```

### Steps to Reproduce

1. Create a slash command / skill that takes arguments. Minimal example - `~/.claude/skills/echo/SKILL.md`:
```markdown
---
name: echo
description: Reply with the argument text verbatim.
---
Reply with the exact text following `/echo`, verbatim. Nothing added, nothing removed, no preface.
```
2. In the terminal CLI, send: `/echo Will this show up in the Android app?`
3. Confirm the terminal renders the full invocation, arguments included (see `cc-terminal.png`).
4. Open the same session in the Claude Code Android app and look at that user message (see `cc-android.png`).

Expected: `/echo Will this show up in the Android app?`
Actual: `/echo` - the arguments are not rendered.

Note: the arguments are still delivered to the model. The assistant reply shown on Android is the echoed argument text, so only the user-message rendering drops them.

### Claude Model

Opus

### Is this a regression?

I don't know

### Last Working Version

_No response_

### Claude Code Version

2.1.235

### Platform

Anthropic API

### Operating System

Windows

### Terminal/Shell

Windows Terminal

### Additional Information

Android app version: 1.260817.0 (26081700)

Screenshots attached in the What's Wrong section:
- `cc-terminal.png` - terminal CLI rendering the full `/echo ...` invocation
- `cc-android.png` - Android app rendering the same message as `/echo` alone

Guide de contribution

Aucun guide de contribution indexé pour ce dépôt

Piste de recherche

Reproduce the issue by creating the documented echo skill, sending `/echo Will this show up in the Android app?` in the terminal, and opening the same session in the Android app. Start by tracing how the Android app renders slash-command user messages and compare it with the terminal representation. Done means the Android user bubble displays the command and its arguments while the assistant reply remains unchanged.

Rédigé par le modèle d'indexation à partir du texte de l'issue.

Évaluation

Stack technique
android
Domaine
mobile
Type d'issue
Bug
Difficulté
3/5
Temps estimé
1-2 jours
Activité
Active
Clarté
Plutôt claire
Accessibilité débutants
55/100

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.