Kaizero — Testing skill for Claude Code
Sensei and Supervisor for an agentic fleet of Claude Codes.
How to install Kaizero
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open IvanRublev/Kaizero and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Kaizero does
Sensei and Supervisor for an agentic fleet of Claude Codes. Builds software guardrailing Spec-driven-development. It loops over the to-do list until each item is implemented, committed, and checked off. Multiple instances can claim to-dos in parallel and are synchronized through the Git repository.
Alternatives in Testing
- Cc Sdd — Spec-driven development (SDD) for your team's workflow 3.1k ★
- AI Contribution Skills (Engram) — branch-pr: clean branch and PR workflow 2.5k ★
- Spec Kitty — Spec-Driven Development for serious software developers 1.6k ★
README
Todo-list sensei for Claude Code. Zeros your list.
Loops Claude until every Task is implemented, committed, and checked off.
Restarts the coding session on a fresh context before rot.
You can spawn multiple instances to parallelize.
It's for practical Loop engineering.
Runs [`claude`](https://claude.com/product/claude-code) on a predefined prompt in a loop on the given Todo List — the file of Task lines selected to ship together as one Release. Makes it hand off each Task for review by default, or land it directly with `--local-merge`: implement, commit, and open a merge/pull request or merge to the base branch — then check the box off once the work reaches the base branch. The session stays interactive, so you can add prompts and make choices as it runs. Launches `claude` in auto permission mode by default, and restarts it on a fresh context before rot sets in.
Contents
- What it does
- Quickstart
- Install
- Todo List format
- Context Rot
- Loop engineering
- Usage
- Cleanup
- Tests
- Security
- Contributing
- Commercial Support
What it does
- Guides Claude to zero a Todo List unattended — one Task at a time until all are implemented, committed, and checked off.
- Beats context rot — session Stop hook SIGTERMs
claudeonce its context total crosses a threshold (see Context rot) and restarts clean. Fresh context, no qu
Related Skills
Crew Agentic
An AI-powered coding agent, similar to Claude Code, that autonomously handles end-to-end software development,
AI Native Sdlc Skills
Unofficial Claude Code skills adapted from Anthropic's The AI-Native SDLC playbook — the committed-artifact ch
Grounded RAG
Data-agnostic, professionally oriented chatbot prototype demonstrated on the Spanish GeSIDA HIV guidelines. Im
Canary Test Pipeline
Run the full test intelligence pipeline — assess critical areas, discover edge cases, trace failure impact, ge
OpenPitStop
A referee your AI coding agent can't cheat. PitStop scans your repo, seals the evidence, catches agents faking
Add
ADD — Agent Driven Development. Coordinated AI agent teams that ship verified software. Spec-driven TDD, trust
Related Agents
Dx Reviewer
Reviews the experience of using and extending Heinzelmännchen - CLI UX (commands, help, output, errors, exit c
Implementation Auditor
Use this agent when you need to verify that recently implemented tasks have been correctly and completely impl
Debruijn
The formalize lane's planner — de Bruijn, the Lean 4 engineer, on Opus. Launch it to plan a milestone's kernel
