Project Orchestrator: Multi-agent Code Coordination

What it does

Project Orchestrator sets up a backend that coordinates multiple AI coding agents across projects. It builds a Neo4j knowledge graph of code and decisions, provides Meilisearch-powered semantic code search, and uses Tree-sitter for precise parsing across languages. It exposes APIs and CLIs for syncing code, creating plans, managing tasks, and recording decisions.

When to use it

Use this skill when multiple agents need shared context for large or complex codebases, when you want structured plans/tasks with dependencies, or when you need persistent recording and search of decisions and code relationships. Useful for multi-agent CI, research projects, or large engineering teams automating development workflows.

What's included

• Scripts: repo contains docker/cargo workflows and CLI examples for sync, serve, and plan/task management.
• References: documentation files (docs/) with install guides, API reference, and integrations for Claude Code, OpenAI, and Cursor.
• Instructions: quick-start, API endpoints, and examples for syncing projects, creating plans, and querying the code graph.

Compatible agents

Designed for agents that orchestrate other agents or need to integrate with MCP tools (Claude Code, Cursor, OpenAI agents). Also useful for developer-facing assistants that produce or analyze code across languages.