Skip to main content
With no flags, in a terminal, the CLI opens the catalog browser. Without a terminal it prints a short usage text and exits with code 0, so agents and CI never hang waiting for input.

Notes

  • Paths. Every install flag takes category/name, the path shown on the component’s page. Separate several with commas: --agent security/security-auditor,development-team/backend-architect. Since 1.33.0 a name alone also works when exactly one component of that type has it (for example the value frontend-developer for an agent); otherwise the summary prints the closest paths as Did you mean: commands.
  • --search "<words>". Since 1.33.0. Prints up to 5 components whose name, path, keywords or description match every word, best text match first (never by downloads), each with its page and install command. Installs nothing. Only the fact that a search ran is counted, never the words.
  • After an install. Each installed component’s page is printed as a Page: line, and when two or more were installed, a Stack: line with the one command that installs them all again.
  • System folders. The CLI refuses to install into C:\Windows (and anything under it) or the root of a drive, where an administrator terminal often starts, and says which folder to open instead.
  • Combining. Install flags combine in one run: one plan, one location prompt for settings and hooks.
  • --yes. Skips every prompt. Settings and hooks go to .claude/settings.local.json.
  • --prompt. After installing, starts claude with that prompt as a single argument (never through a shell). Without a terminal it prints the command to run instead.
  • --analytics, --chats, --health-check. Local tools, see Local dashboards.
  • Removed flags. Seventeen flags were removed in 1.31.0 and now print their replacement. See Removed in 1.31.0.
This table is generated from --help by docs-site/scripts/generate_snippets.py, and CI fails when it falls behind the code. Check yours with npx claude-code-templates@latest --version.