[BUG] Built-in slash command (/stickers) intercepted as conversation when sent during task execution

[markspivey]

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?

When a slash command (e.g., /stickers) is sent while Claude Code is actively executing a task, the command is incorrectly intercepted and handled as conversational input instead of being executed as a system command.

Sequence:

1. Claude Code was executing a task (creating documents)
2. A tip about /stickers appeared in the output
3. User sent /stickers while Claude was still processing/responding
4. Bug: Claude responded conversationally saying the command wasn't recognized
5. User sent /stickers again (after task completed)
6. Command executed correctly

This suggests slash commands are being routed to Claude's conversation handler when sent during active task execution, instead of being handled by the system command processor.

What Should Happen?

Slash commands should be handled by the system command processor regardless of whether Claude is currently executing a task.

Expected behavior options:

1. Immediate execution: Command executes immediately, interrupting current task
2. Queued execution: Command is queued and executes after current response completes
3. Clear rejection: System shows "busy, please wait" message (not conversational response)

Should NOT happen:

• Slash commands should never be interpreted as conversational input
• Claude should never respond as if the command doesn't exist when it actually does

Error Messages/Logs

∴ Thinking…

  The user typed "/stickers" which appears to be a slash command. Let me check if this is a valid command I should handle with the SlashCommand tool.

  Looking at my instructions, the SlashCommand tool is for custom slash commands that appear in the "Available Commands" list. The instructions say:

  "IMPORTANT: Only use this tool for custom slash commands that appear in the Available Commands list below. Do NOT use for:
  - Built-in CLI commands (like /help, /clear, etc.)
  - Commands not shown in the list
  - Commands you think might exist but aren't listed"

  I don't see an "Available Commands" section in my context, which means there are no custom slash commands configured for this project/user.

  The user previously used "/clear" which was shown as a built-in command. "/stickers" doesn't sound like a built-in command either - it's not listed in the system instructions as a built-in command.

  Given that:
  1. There's no custom slash commands list provided
  2. "/stickers" is not a documented built-in command
  3. The instruction says to ask users to check the slash command file if their command isn't listed

  I should let the user know that "/stickers" is not a recognized command. However, I should be helpful and ask if they meant something else or if they'd like help with something specific.

  Actually, looking more carefully at the context, I see a system reminder that was injected:

  "
  The user sent the following message:
  /stickers

  Please address this message and continue with your tasks.
  "

  This seems to be instructing me to acknowledge the user's message. Perhaps "/stickers" is meant to be some kind of encouragement or celebration command? Or maybe they just wanted to test something?

  I should acknowledge it briefly and ask if they need anything else or if they're ready to proceed with the homework preparation.

● I don't recognize /stickers as a valid command. It's not in the list of built-in commands or custom slash commands configured for this project.

  Did you mean to:
  - Test a command?

Steps to Reproduce

1. Start a task that causes Claude Code to generate a lengthy response
   • Example: Ask Claude to create detailed documentation or analyze multiple files
2. While Claude is still generating the response, type a slash command
   • Example: /stickers (works with other commands too, likely)
3. Press Enter while Claude is mid-task

Observed: Claude responds conversationally, claiming the command doesn't exist

4. Wait for Claude to finish the task completely
5. Send the same slash command again

Observed: Command now executes correctly

Expected: Command should either execute or queue in step 3, never be misinterpreted as conversation

Claude Model

Sonnet (default)

Is this a regression?

I don't know

Last Working Version

N/A

Claude Code Version

2.0.33 (Claude Code)

Platform

Other

Operating System

Windows

Terminal/Shell

PowerShell

Additional Information

No response