Slash Commands: Using CC Commander Inside Claude Code
Reference for all /ccc slash commands available inside Claude Code sessions, plus how to invoke CC Commander headlessly from an AI agent using ccc —dispatch.
CC Commander works inside any Claude Code session without the standalone CLI. Type /ccc in a Claude Code session and the full interactive menu appears — the same 15 options, the same sub-menus, the same guided flows. This also works in Claude Desktop Cowork and VS Code with the Cursor extension.
This displays the main interactive menu with all 15 options. You navigate with a numbered prompt or by typing the key shown next to each choice. Type “back”, “cancel”, or press Escape at any point to return to the previous menu.
Theme switching (/ccc theme) controls the visual skin of the ccc terminal UI. It has no visible effect inside a Claude Code session, where rendering is handled by Claude’s own interface. To change the theme, run ccc in a terminal and press t from the main menu.
When you need CC Commander to run a task without an interactive session — for example, from an AI agent or CI script — use the --dispatch flag:
ccc --dispatch "Add JWT authentication to the login endpoint"
The Intelligence Layer scores the task, detects your stack, selects the right model and budget, and dispatches Claude without any prompts. Add --json to receive structured output:
ccc --dispatch "Build auth with JWT" --json --model opus --budget 5
# Run a task headlessly and capture JSON outputresult=$(ccc --dispatch "Refactor the auth module" --json)# Check for errorsecho $result | jq '.error // "success"'
You can override the Intelligence Layer’s auto-selection with explicit flags:
Flag
Example
What it does
--model
--model opus
Force a specific model
--max-turns
--max-turns 50
Cap the number of turns
--budget
--budget 5
Set a USD budget cap
--cwd
--cwd /path/to/project
Run in a different directory
For AI orchestrators, read docs/BIBLE-AGENT.md from the CC Commander repo. It is a 268-line machine-readable reference covering the full CLI API, dispatch patterns, JSON schemas, and integration points — written so any AI agent can manage CCC immediately after reading it.