Skip to main content
CC Commander is an AI project manager that sits on top of Claude Code. Instead of typing prompts into a blank terminal and hoping for the best, you get a guided, arrow-key menu with 450+ on-demand skills, automatic model and budget selection, and a knowledge engine that compounds everything it learns across your sessions. One install. No configuration. The longer you use it, the smarter it gets.

​
Why CC Commander exists

Stock Claude Code starts each session from zero — no memory of what worked, no awareness of your stack, and no guidance on which tool to reach for. It also burns context fast: verbose tool output consumes tokens you could be spending on actual work. CC Commander fixes all of this. Before dispatching a single token, it reads your project, scores your task, pulls relevant lessons from past sessions, and auto-selects the right model and budget. When the session ends, it extracts patterns and errors so the next dispatch is better-informed than the last.

​
Key benefits

Intelligence Layer

Scores every task 0–100, reads your stack, injects past lessons, and auto-selects model and budget — no flags, no YAML.

450+ skills

On-demand expertise across 11 domains. Load only what you need; skills do not consume context unless you use them.

Knowledge compounding

Every completed session extracts patterns and errors. Future dispatches are informed by everything that worked before.

Interactive TUI

Arrow-key menus, 10 themes, a live Cockpit Dashboard, and tmux Split Mode so each task gets its own window.

​
What’s included

CC Commander ships everything in a single install. Here is what you get out of the box.
ComponentCountWhat it does
Skills450+On-demand expertise, deduplicated
Domains11Domain routers with specialized sub-skills
Commands80+Slash commands with the /ccc: prefix
Hooks25Lifecycle automation (pre-dispatch, post-session, and more)
Themes10Cyberpunk, Fire, Ocean, Aurora, Sunset, Monochrome, Rainbow, Dracula, Graffiti, Futuristic
Vendor packages19Best-in-class community tools, auto-updated weekly

​
How the Intelligence Layer works

The Intelligence Layer runs four modules before every dispatch. Weighted complexity scoring — every task is scored 0–100 using 47 keyword signals, word count, and fuzzy regex matching. The score drives model selection and token budget automatically: 
"fix typo"             → score  0  → 10 turns, $1 budget, Haiku
"add dark mode"        → score 25  → 20 turns, $3 budget, Sonnet
"refactor auth module" → score 60  → 35 turns, $6 budget, Sonnet
"build SaaS platform"  → score 100 → 50 turns, $10 budget, Opus
Stack detection — CC Commander reads package.json, Dockerfile, go.mod, and requirements.txt before every dispatch, detects your framework, reads your current git branch, and flags monorepo patterns. Skill recommendations — skills are ranked by stack match, task keyword overlap, and your personal usage history. Skills that led to successful sessions rank higher over time. Knowledge compounding — every completed session extracts a lesson (keywords, stack, error patterns, success patterns) stored in ~/.claude/commander/. The next dispatch searches this store with time-decay relevance: lessons under 7 days old carry 2× weight, under 30 days 1.5×, older lessons 1×.

​
What you need

  • Node.js 18 or later — required for the ccc CLI and the installer
  • Claude Code CLI — installed automatically if not already present
tmux is recommended for Split Mode, which runs the CCC menu and each Claude session in separate terminal windows. Install with brew install tmux on macOS or apt install tmux on Linux. Split Mode is optional — CC Commander works fully without it.

​
Ready to install?

Get started in 60 seconds

Install CC Commander with one command and launch your first session.