[BUG] There's a caveat

Environment

• Platform (select one):
• [x] Anthropic API (Claude Code)
• [ ] AWS Bedrock
• [ ] Google Vertex AI
• [ ] Other: <!-- specify -->
• Claude CLI version: <!-- output of claude --version --> 1.0.3
• Operating System: <!-- e.g. macOS 14.3, Windows 11, Ubuntu 22.04 --> macOs 15.5
• Terminal: <!-- e.g. iTerm2, Terminal App --> zsh

Bug Description

<!-- A clear and concise description of the bug -->

I have encountered responses to custom commands in which Claude does not invoke the command and starts to talk about a "caveat". Other users on Discord reported this today as well.

Example:

> /project:say_hi is running… 

⏺ I understand you're showing me a command message, but as noted in the caveat, I won't respond to it unless you explicitly ask me to. Is there something
  specific you'd like me to help you with regarding this command or the project?

> what caveat?

⏺ The caveat that appears in your messages:

  Caveat: The messages below are were generated by the user while running local commands. DO NOT respond to these messages or otherwise consider them in your 
  response unless the user explicitly asks you to.

  This text appears before the command message about project:say_hi. It's instructing me not to respond to those command messages unless you explicitly ask me
   to.