cub
♥Cherished
Cub is a CLI task orchestrator that structures AI coding work into discrete, agent-sized tasks and runs them through Claude Code or other AI harnesses in a reliable autonomous loop. It helps developers plan complex changes upfront, verify task results, record execution history, and retry failures with context—eliminating manual context-switching between agent sessions.
Clauded With Love Rating
8.3 / 10
Cub is a CLI orchestrator that structures AI coding work into discrete tasks and runs them autonomously through various AI harnesses like Claude Code. It addresses the real problem of context-switching and session management when working with AI coding agents by providing planning, execution tracking, and retry mechanisms.
Code Quality7.5
Usefulness8.5
Claude Usage9.0
Documentation8.0
Originality8.5
Highlights
- ✓Multi-harness abstraction allows vendor-agnostic AI agent orchestration across Claude Code, Codex, Gemini, and OpenCode
- ✓Comprehensive task lifecycle with planning, execution tracking, ledger recording, and automatic retry with failure context
- ✓Well-structured CLI with clear installation methods, budget controls, and both single-iteration and continuous execution modes
To Improve
- →Add comprehensive test coverage and error handling examples to demonstrate production readiness
- →Include performance benchmarks and cost analysis examples to help users understand resource implications of different harness choices