When it comes to custom slash commands, the entire content of the markdown file is added as a user message in the Anthropic API request. It's the same as pasting a prompt.
In many commands I see there is description, usage etc. about the command. These are sent in one message. Is this intentional?
Otherwise Anthropic suggest to use frontmatter for those metadata about the command (reusable prompt). Link: https://docs.anthropic.com/en/docs/claude-code/slash-commands#frontmatter
When it comes to custom slash commands, the entire content of the markdown file is added as a user message in the Anthropic API request. It's the same as pasting a prompt.
In many commands I see there is description, usage etc. about the command. These are sent in one message. Is this intentional?
Otherwise Anthropic suggest to use frontmatter for those metadata about the command (reusable prompt). Link: https://docs.anthropic.com/en/docs/claude-code/slash-commands#frontmatter