Cook: A simple CLI for orchestrating Claude Code
rjcorwin.github.io
March 19, 2026
4 min read
🔥🔥🔥🔥🔥
62/100
Summary
Workflow loops for Claude Code, Codex, and OpenCode streamline task management by enabling iterative reviews and selections of implementation strategies. Users can execute tasks with specific commands such as reviewing, passing, racing options, and composing freely to optimize results.
Key Takeaways
- Cook allows users to implement iterative workflows for coding tasks, utilizing commands like "xN" for sequential passes and "vN" for parallel execution.
- The review loop in Cook enables quality checks after each iteration, determining whether to mark a task as DONE or to ITERATE for further refinement.
- Users can specify different agents and models for each step of the workflow, enhancing flexibility in task execution and review processes.
- Cook supports composition operators to run multiple coding tasks in isolated environments and combine results based on specified criteria.
Community Sentiment
Positives
- The CLI tool Cook is praised for its concise design, making it approachable for users who want to experiment with AI functionalities.
- The enthusiasm from the developer about getting featured on the front page indicates a positive reception and community interest in the tool.
Concerns
- Concerns were raised about the lack of clarity on how Cook handles user input scenarios, which is critical for effective interaction with AI models.
- The website's design was criticized for being visually unappealing, which could deter potential users from engaging with the tool.
Related Articles
Get Shit Done: A Meta-Prompting, Context Engineering and Spec-Driven Dev System
Mar 17, 2026
Claude Code – Everything You Can Configure That the Docs Don't Tell You
May 29, 2026
/architect: Reduce Fable tokens by 80%, Fable orchestrates/reviews, Codex builds
Jun 12, 2026
Agent-Manager: A Tmux TUI for Running Claude Code, Codex and OpenCode
Jul 30, 2026