light-orchestrator

Coordinate multi-stage AI workflows with checkpointing and context restoration.

514|67|Updated Jun 7, 2026

Search Tags:#workflow #memory #pipeline #handoff #checkpoint #resumption #orchestrator

Ready to run

npx skills add https://github.com/Light0305/Light-skills --skill light-orchestrator

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: light-orchestrator
Source: https://github.com/Light0305/Light-skills/tree/main/skills/light-orchestrator
Command: npx skills add https://github.com/Light0305/Light-skills --skill light-orchestrator

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates cross-stage AI workflows by coordinating multiple skills and preserving context across interruptions.

Core Features & Use Cases

Phase-based planning and scheduling across stages (planning, execution, review).
Checkpointing and forced phase boundaries to prevent drift and ensure artifact handoffs.
Context restoration from passport memory and stage logs to resume interrupted tasks.

Quick Start

Describe a multi-stage task (e.g., "from data to paper") to the orchestrator to generate a phase-by-phase plan and resume after interruptions.

Frequently Asked Questions about light-orchestrator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I coordinate multi-stage AI workflows and preserve context across interruptions?▼

To coordinate multi-stage AI workflows, you use an orchestrator that enforces stage contracts, tracks artifacts in a passport, and restores context from checkpoint logs to resume interrupted long-running pipeline tasks seamlessly.

What is the best way to resume an interrupted long-running AI pipeline?▼

The best way to resume an interrupted AI pipeline is through context restoration from passport memory and stage logs, allowing stepwise handoffs to pick up exactly where the previous execution left off.

How do I prevent context drift when handing off artifacts between AI planning and execution stages?▼

To prevent context drift during AI stage handoffs, apply forced phase boundaries and checkpointing to ensure artifacts are strictly validated against stage contracts before moving to the next phase.

Can I use this orchestrator for projects that span fewer than three stages?▼

This orchestrator is designed for long-running projects that span three or more stages, such as planning, execution, and review, making it unsuitable for simpler, shorter tasks without phase boundaries.

Do I need specific conventions to define a stage pipeline for AI workflow orchestration?▼

Yes, AI workflow orchestration requires a CONVENTIONS-compliant stage pipeline to enforce phase boundaries, interface with optional scripts, and successfully track artifacts across stepwise handoffs.