A simple CLI for orchestrating Claude Code, Codex, and OpenCode.
# review loop
cook "Implement dark mode" review
# 3 passes
cook "Implement dark mode" x3
# race 3, pick best
cook "Implement dark mode" v3 "least code"
# two approaches, pick one
cook "Auth with JWT" vs "Auth with sessions" pick "best security"
# task list
cook "Work on next task in plan.md" review ralph 5 "DONE if all tasks complete, else NEXT"Two ways to get it:
Skill (no CLI) — your agent orchestrates subagents directly:
mkdir -p .claude/skills/cook && curl -fsSL https://raw.githubusercontent.com/rjcorwin/cook/main/no-code/SKILL.md -o .claude/skills/cook/SKILL.mdCLI — standalone tool for terminal or CI:
npm install -g @let-it-cook/cli
mkdir -p .claude/skills && cp -r $(npm root -g)/@let-it-cook/cli/skill .claude/skills/cookRequires Node.js 20+ and Claude Code, Codex, or OpenCode.
By default, Docker sandbox restricts outbound traffic to only what each agent requires. Configure in .cook/config.json:
{
"docker": {
"network": {
"mode": "restricted",
"allowedHosts": ["api.mycompany.com"]
}
}
}mode |
Behavior |
|---|---|
"restricted" (default) |
Blocks all outbound traffic except DNS and HTTPS to a whitelist of required agent hosts |
"unrestricted" |
No network restrictions |
allowedHosts adds extra hosts to the whitelist when mode is "restricted".
cook init # Set up COOK.md, config, and Dockerfile
cook rebuild # Rebuild the sandbox Docker image
cook doctor # Check Docker + auth readiness
cook shell / sandbox # Interactive shell in Docker sandbox
cook shell <command> # Run a command in the sandbox
cook shell --unrestricted # Shell with unrestricted networkingWork — a prompt string. One agent call.
Loop operators wrap work with iteration:
xN— N sequential passesreview— review → gate → iterate loop (up to 3 by default)ralph N "<gate>"— repeat until a gate says DONE
Composition operators run parallel branches then resolve:
vN— N identical branches, pick the bestvs— two different branches, pick the bestpick,merge,compare— resolvers
Operators compose left to right. Each wraps everything to its left.
cook "work" x3 review # 3 passes, then review loop
cook "work" review x3 # review loop, repeated 3 times
cook "work" review v3 # race 3, each with a review loopcook "Add dark mode" review # default prompts, up to 3 iterations
cook "Add dark mode" review 5 # up to 5 iterations
cook "Add dark mode" "Review for accessibility" "DONE if WCAG AA, else ITERATE"
cook "Add dark mode" review --work-agent codex --review-agent claude --review-model opusWhen an agent hits a token quota or rate limit, cook automatically waits and retries instead of bailing. A countdown is shown in the TUI. Enabled by default.
cook "Build the feature" review --no-wait # disable: fail immediately on rate limitConfigure in .cook/config.json:
{
"retry": {
"enabled": true,
"pollIntervalMinutes": 5,
"maxWaitMinutes": 360
}
}cook initCreates COOK.md (project instructions + prompt template) and .cook/config.json:
{
"agent": "claude",
"sandbox": "agent",
"steps": {
"work": { "agent": "codex", "model": "gpt-5-codex" },
"review": { "agent": "claude", "model": "opus" }
},
"env": ["CLAUDE_CODE_OAUTH_TOKEN"]
}