spur-loop-authoring

Create and manage multi-agent coding loops with a doctor-preview approval workflow.

2|Updated Apr 15, 2026
One-click install
npx skills add https://github.com/getspur/spur --skill spur-loop-authoring
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: spur-loop-authoring
Source: https://github.com/getspur/spur/tree/main/assets/skills/spur-loop-authoring
Command: npx skills add https://github.com/getspur/spur --skill spur-loop-authoring

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill solves the complexity of managing long-running, multi-agent coding tasks by providing a structured, preview-first workflow for defining and scheduling autonomous loops.

Core Features & Use Cases

  • Durable Loop Authoring: Transforms natural-language intent into structured, validated loop configurations.
  • Preview-First Safety: Ensures every loop is reviewed and approved via a doctor-produced preview before execution.
  • Use Case: Use this to automate a recurring weekly refactoring task where a brain plans the work, delegates to specialized agents, and reports results back to a single review lane.

Quick Start

Initiate the loop authoring process by typing /spur-loop followed by your natural language request for a recurring coding task.

Frequently Asked Questions about spur-loop-authoring

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

FAQPage Schema
How do I automate recurring multi-agent coding workflows safely?

Multi-agent orchestration for coding tasks is managed through durable, multi-agent coding loops. The system uses a structured doctor-preview workflow that transforms natural-language intent into validated loop configurations, ensuring safe and predictable autonomous execution.

How do I schedule autonomous coding tasks with cadence-based triggers?

Scheduling autonomous coding tasks with cadence-based triggers involves initiating the loop authoring process with a natural language request. The system generates a structured, validated loop configuration that requires preview and approval before execution begins.

What is the best way to delegate complex software engineering tasks to specialized agents?

Delegating complex software engineering tasks to specialized agents is best handled through multi-agent coding loops. A primary agent plans the work, delegates to specialized agents, and consolidates results back into a single review lane.

Can I use natural language to define autonomous escalation logic for code refactoring?

Yes, natural language can be used to define autonomous escalation logic for code refactoring. The durable loop authoring process transforms your natural-language intent into structured, validated configurations that include autonomous escalation logic.

Do I need to approve multi-agent workflows before they execute?

Yes, multi-agent workflows require approval before execution. The system mandates strict adherence to a preview-approval-submission lifecycle, utilizing a doctor-produced preview to ensure safe and predictable agent execution.

When should I not use durable loops for software engineering operations?

You should not use durable loops for software engineering operations when bypassing the preview-approval-submission lifecycle is necessary. The system's safe and predictable agent execution fundamentally depends on this strict review process.