AI Coding Rules Frontend Plugin#
Overview#
The AI Coding Rules frontend plugin provides a comprehensive interface for visualizing AI coding rules, agent skills, ignore files, and agent configurations from a wide range of AI coding assistants. It supports 11 agent types and is designed to make all AI context living in a repository visible and navigable directly within Backstage.
Features#
Rule Visualization (11 Agent Types)#
- Display rules from Cursor, GitHub Copilot, Cline, Claude Code, Windsurf, Roo Code, OpenAI Codex, Gemini CLI, Amazon Q Developer, Continue, and Aider
- Color-coded chips and icons for quick visual identification of each agent type
- Expandable rule cards with metadata and full content
- Support for markdown rendering and raw source view per rule
Content Search and UX Utilities#
- Content Search Bar: Filter all displayed rules by any keyword in name, title, description, or content
- Copy to Clipboard: One-click copy for any rule's Markdown content with confirmation toast
- Raw/Rendered Toggle: Switch between rendered Markdown and raw source per rule
- Export to Markdown: Download all currently filtered rules as a single
.mdfile
Manual Filtering with Apply Filter#
- Filter rules by type (all 11 agent types)
- Manual control with Apply Filter button functionality
- Configurable default rule types
- Always visible filter interface
- Status indicators for unsaved changes
MCP Server Visualization#
- Groups servers by source: Cursor, VSCode, Claude, Windsurf, Cline
- Displays server type (local/remote) and command information
- Expandable view of environment variables
- Raw configuration display with syntax highlighting
Ignore Files Viewer#
Displays agent-specific ignore files (.cursorignore, .aiderignore, .rooignore, .geminiignore, .copilotignore) with:
- Line-numbered content using
CodeSnippet - Agent name and file path display
- Link to open the file in the repository
Agent Configs Viewer#
Displays agent configuration files with language-aware syntax highlighting:
.aider.conf.yml.continue/config.yaml/.continue/config.json.cursor/settings.json.zed/assistant.json
Agent Skills Viewer (agentskills.io Standard)#
Discovers and displays Agent Skills — reusable, portable AI capabilities:
- Scans
.agents/skills/,.claude/skills/, and.cursor/skills/ - Stats bar showing total skills and source breakdown
- Skills grouped by source in expandable accordions
- Each skill displays: name, description, compatibility chips, license, author, version,
allowedTools - Full SKILL.md content with raw/rendered toggle
- Resource listing:
scripts/,references/,assets/
Clickable Git Repository Links#
- Launch icon (↗) for each rule/skill/file to open in git repository
- Multi-provider support (GitHub, GitLab, etc.)
- Opens in new tab for easy access
Statistics and Overview#
- Total rule count display with per-type breakdown
- Visual statistics cards with color-coded counts
- Counts update in real-time with active search/filter
Components#
AiInstructionsComponent (Recommended)#
The unified tabbed component that brings all features together:
import { AiInstructionsComponent } from '@terasky/backstage-plugin-ai-rules';
// Default title
<AiInstructionsComponent />
// Custom title
<AiInstructionsComponent title="Development Guidelines" />
Tabs provided:
- Agent Rules — AI coding rules from all 11 agents
- MCP Servers — MCP server configurations
- Ignore Files — Agent ignore patterns
- Agent Configs — Agent configuration files
- Agent Skills — Agent Skills (agentskills.io)
AIRulesComponent#
Standalone component for AI coding rules only:
import { AIRulesComponent } from '@terasky/backstage-plugin-ai-rules';
<AIRulesComponent />
<AIRulesComponent title="Cursor Rules Only" />
MCPServersComponent#
Standalone component for MCP server configurations:
IgnoreFilesComponent#
Standalone component for agent ignore files:
AgentConfigsComponent#
Standalone component for agent configuration files:
import { AgentConfigsComponent } from '@terasky/backstage-plugin-ai-rules';
<AgentConfigsComponent />
AgentSkillsComponent#
Standalone component for Agent Skills:
Component Props#
All components share the same optional prop:
| Property | Type | Default | Description |
|---|---|---|---|
title |
string |
Component-specific default | The title displayed at the top of the component |
Technical Details#
Integration Points#
- Backstage catalog entities
- AI Rules backend plugin
- Repository source locations
- Entity metadata
Supported Agent Types and Files#
| Agent | Type Value | Key Files |
|---|---|---|
| Cursor | cursor |
.cursorrules, .cursor/rules/*.mdc\|.md, .cursor/MEMORY.md |
| GitHub Copilot | copilot |
.github/copilot-instructions.md, .github/instructions/*.instructions.md |
| Cline | cline |
.clinerules (root), .clinerules/*.md |
| Claude Code | claude-code |
CLAUDE.md, .claude/CLAUDE.md, CLAUDE.local.md, .claude/rules/*.md |
| Windsurf | windsurf |
.windsurfrules, .windsurf/rules/*.md |
| Roo Code | roo-code |
.roorules, .roo/rules/, .roo/rules-{mode}/ |
| OpenAI Codex | codex |
AGENTS.md, AGENTS.override.md |
| Gemini CLI | gemini |
GEMINI.md, .gemini/*.md |
| Amazon Q | amazon-q |
.amazonq/rules/*.md |
| Continue | continue |
.continue/rules/*.md, .continue/prompts/*.md |
| Aider | aider |
CONVENTIONS.md |
Rule Detection#
The plugin automatically detects components with:
- Git source locations in entity metadata
- Repository annotations
- Supported rule file patterns
Error Handling#
- Clear error messages for missing rules or files
- Loading states during data fetching
- Graceful degradation for unsupported repositories
- Per-tab isolation — an error in one tab does not affect others
Use Cases#
Development Team Guidelines#
- Centralize AI coding rules across projects
- Ensure consistent AI assistant configurations
- Share best practices for AI-assisted development
- Maintain rule documentation alongside code
Rule Management#
- Discover existing AI rules in repositories
- Validate rule configurations
- Monitor rule adoption across teams
- Track rule evolution over time
Agent Skills Discovery#
- Browse reusable AI capabilities available in a repository
- Understand which skills are scoped to specific agents vs. shared
- Inspect skill metadata, compatibility, and bundled resources
Code Quality Assurance#
- Enforce coding standards through AI rules
- Maintain consistency across AI assistants
- Document preferred coding patterns
- Share domain-specific guidelines