claude-code-orchestra
Multi-agent coding orchestration with delegation, review, and checkpoints
All Skills in This Repository (13)
Pure Emerald Level Indicatorscontext-loader
Load shared project context from the .claude directory for Gemini CLI.
update-design
Extract conversation decisions and update DESIGN.md Markdown sections with a dated changelog entry.
init
Detect tech stack and commands from project files to update AGENTS.md top sections.
startproject
Coordinate multi-agent project kickoff and planning across Claude, Codex, and Gemini.
plan
Generate structured implementation plans with requirements, tasks, and acceptance criteria.
design-tracker
Record design decisions from conversations into DESIGN.md and the changelog.
research-lib
Research libraries from official docs and generate Markdown documentation.
gemini-system
Analyze large codebases and documents with Gemini CLI subagents.
update-lib-docs
Update library documentation in .claude/docs/libraries with web-sourced version and security changes.
checkpointing
Capture CLI logs and Git activity into centralized checkpoint files.
simplify
Analyze Python code complexity and apply safe refactoring techniques.
codex-system
Consult Codex CLI for design, debugging, and code review decisions.
Frequently Asked Questions
FAQPage SchemaHow to install claude-code-orchestra?βΌ
Run `npx skills add DeL-TaiseiOzaki/claude-code-orchestra --all -g -y` in your terminal to install all skills in this suite globally.
What does claude-code-orchestra do?βΌ
It turns Claude Code into an orchestrator that delegates coding, research, debugging, and review work to specialist sub-agents and Codex CLI, then verifies every result with automated quality gates.
How does multi-agent delegation work here?βΌ
A routing policy sends routine implementation to Sonnet agents, hard analysis to Opus agents, and design or complex code to Codex, while the main agent verifies diffs and test results before accepting anything.
Does it work with Codex CLI and Claude Code together?βΌ
Yes. Shared wrapper scripts let Claude Code consult Codex for planning, debugging, and reviews, with every call logged and sandboxed.
Can it keep context across long sessions?βΌ
Yes. Checkpointing, progress files, and design-tracking skills save session state to disk so work resumes cleanly after breaks or context compaction.
Related Repositories in Software Engineering
View All in Software Engineeringβopenclaw
Run a personal AI assistant across your devices and chat apps
superpowers
Gives coding agents a disciplined workflow from idea to merged code
react
AI agent skills for building, testing, and porting React core