fusion-rules

Route editor-specific rule authoring requests and scaffold instruction files.

1|Updated Feb 18, 2026
One-click install
npx skills add https://github.com/equinor/fusion-skills --skill fusion-rules
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fusion-rules
Source: https://github.com/equinor/fusion-skills/tree/main/skills/fusion-rules
Command: npx skills add https://github.com/equinor/fusion-skills --skill fusion-rules

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Many repositories need editor-specific instruction and rule files (GitHub Copilot, Cursor, Claude Code) but teams struggle to choose the right format, validate frontmatter/globs, and keep equivalent guidance consistent across editors. Fusion Rules centralizes intent detection, routes requests to the correct editor agent, and ensures drafts, validations, and diffs are shown before any file writes.

Core Features & Use Cases

  • Editor detection and routing: Parse user intent to select Copilot, Cursor, or Claude Code workflows and load only the routed agent.
  • Scaffolding and validation: Create or improve root and path-scoped instruction files, validate frontmatter (applyTo, globs, paths), and enforce quality gates (length, structure).
  • Multi-editor orchestration: Run agents sequentially, pass scan summaries to avoid redundant scans, and generate equivalent files per editor without cross-editor duplication.
  • Use case: Onboard a mixed team by detecting the preferred editor and scaffolding repository-specific Copilot instructions, Cursor rules, and a CLAUDE.md with consistent content and safe defaults.

Quick Start

Ask fusion-rules to detect the target editor and scaffold or review project rules for GitHub Copilot, Cursor, or Claude Code.

Frequently Asked Questions about fusion-rules

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

FAQPage Schema
How do I scaffold editor-specific rule files for GitHub Copilot, Cursor, and Claude Code?

Scaffolding editor-specific rule files involves routing your authoring request to the correct agent based on detected editor intent, then generating validated instruction files with proper frontmatter and glob patterns. Fusion Rules automates routing, validation, and file creation.

How does multi-editor rule orchestration keep instructions consistent across different AI assistants?

Multi-editor rule orchestration runs editor-specific agents sequentially and passes scan summaries between them. This prevents redundant scans and generates equivalent instruction files for Copilot, Cursor, and Claude Code without cross-editor duplication.

Can I validate frontmatter and glob patterns in existing Cursor rules or Copilot instructions?

Yes, you can validate frontmatter and glob patterns in existing rule files. The review process checks frontmatter fields like applyTo, globs, and paths, enforcing quality gates for length and structure before requiring diffs and approval for any writes.

What is the best way to manage path-scoped rules for multiple AI coding assistants in one repository?

Managing path-scoped rules for multiple AI assistants requires centralizing intent detection and routing requests to editor-specific agents. This ensures equivalent guidance is created consistently across Copilot, Cursor, and Claude Code instruction files.

Do I need to manually select an agent when creating Claude Code or Cursor rules?

No, you do not need to manually select an agent. The system parses user intent to detect the target editor automatically and loads only the routed agent file, streamlining the creation of Claude Code or Cursor rules.

Why should I use a routing orchestrator instead of writing AI assistant instruction files directly?

Using a routing orchestrator solves the problem of choosing the correct format, validating frontmatter, and keeping guidance consistent across editors. It ensures drafts and validations are shown with diffs before any file writes occur.