🏗️ ERNE — Everything React Native & Expo

AI agent harness with 13 specialized agents, autonomous worker, video debugging, adaptive fix, and 29 auto-updating knowledge rules. Adaptive stack detection, multi-agent orchestration, and a pixel-art dashboard. Every config is generated to match your exact project setup.

⚡ Quick Start

Prerequisite: Claude Code is required for the full experience (13 agents, pipeline orchestration, dashboard, hooks). Other IDEs receive adaptive rules and configuration — see IDE Support.

Install

# Install globally (recommended)
npm i -g erne-universal

Setup

# One-time setup in your React Native / Expo project
erne init

# Launch the dashboard
erne dashboard

# Or do both at once
erne start

Without global install: use npx erne-universal instead of erne (slower — downloads each time).

What init does:

1. 🔍 Deep-scans your project — detects 15 stack dimensions (state management, navigation, styling, lists, images, forms, storage, testing, build system, component style, monorepo, New Architecture, and more)
2. 🎚️ Lets you choose a hook profile (minimal / standard / strict)
3. 🔌 Select MCP integrations (simulator control, GitHub, etc.)
4. ⚙️ Generates adaptive configuration — selects from 24 variant templates matched to your exact stack (Zustand vs Redux, Expo Router vs React Navigation, NativeWind vs StyleSheet, etc.)

CLI Commands

erne init                  # Setup ERNE in your project
erne dashboard             # Launch dashboard
erne start                 # Init + dashboard
erne doctor                # Health check (22 checks)
erne audit                 # Generate project documentation
erne worker --config w.json  # Start autonomous worker
erne sync-configs          # Export rules to other IDEs
erne update                # Update to latest version
erne uninstall             # Clean removal

📦 What's Included

🎮 Agent Dashboard

ERNE includes a real-time dashboard with 6 pages, pixel-art agent HQ, and adaptive fix integration.

erne dashboard              # Start on auto-detected port, open browser
erne dashboard --port 4444  # Custom port
erne start                  # Init + dashboard in one command

Command Center

Health score, issues summary, agent status, system health checks.

Issues — Agent Fix

Severity filtering, Agent Fix buttons (auto-detects Claude Code), "Needs planning" for complex issues, real-time fix output.

Tasks — Worker & Providers

6 provider integrations (ClickUp, GitHub, Linear, Jira, GitLab, Local), Kanban board, interactive setup guides.

Agents — Pixel Art HQ

13 animated agents in 4 rooms, real-time status, custom naming, activity history.

Project — Stack & Docs

Stack detection, MCP servers, generated documentation (12 doc types), recommendations.

Commands

23 commands including /erne-debug-video, click to copy.

🎬 Video Debugging

Analyze screen recordings to find temporal UI bugs that screenshots cannot capture.

/erne-debug-video recording.mp4

Catches: animation glitches, race conditions, gesture issues, scroll jank, keyboard overlap, navigation transitions.

• Extracts key frames via ffmpeg scene detection (any format: mp4, mov, webm, avi, mkv, gif)
• Claude analyzes frames as a timeline with frame references
• No additional API keys needed — uses Claude you already have

📚 29 Knowledge Rules (Auto-Updating)

ERNE ships with 29 comprehensive rule files covering the entire React Native ecosystem:

Rules auto-update weekly via GitHub Action — checks npm for new SDK/RN versions, analyzes changelogs with Claude API, and opens a PR.

🎯 Multi-Agent Orchestration

ERNE supports coordinated multi-agent workflows through the pipeline orchestrator:

/orchestrate "build user profile screen"

Pipeline phases:

  🏗️ Plan          architect decomposes the task
       ↓
  ⚡ Implement     senior-developer + feature-builder (parallel)
       ↓
  🧪 Test          tdd-guide writes and runs tests
       ↓
  🔍 Review        code-reviewer validates with evidence
       ↓
  📊 Validate      performance-profiler checks performance

Features retry logic (max 3 attempts), escalation to user on persistent failures, and structured handoff templates for context passing between agents. See Pipeline Documentation for details.

🤖 Agents

Each agent has a distinct personality, quantified success metrics, and memory integration for cross-session learning.

🧠 Context Optimization

ERNE includes a built-in context intelligence system (auto-enabled with dashboard) that compresses tool outputs by 97-100%, indexes content with FTS5 search, and manages your context budget. See BENCHMARK.md for the full 21-scenario breakdown.

Pairs well with context-mem — ERNE gives your AI deep React Native knowledge, context-mem gives it persistent memory across sessions. Together: an AI assistant that knows RN best practices AND remembers your project patterns. npm i context-mem && npx context-mem init

💰 Token Efficiency

ERNE minimizes token usage through two complementary systems: architecture-level savings (what gets loaded into context) and runtime context optimization (how tool outputs and session state are compressed).

Architecture savings

Runtime context optimization (benchmark-verified)

Result: Architecture saves 60–67% on what enters context. Runtime optimization achieves 97–100% compression on tool outputs (verified across 21 benchmark scenarios with 537 KB of real data). In a full debugging session, 99% of tool output tokens are eliminated — leaving 99.6% of your context window free for actual problem solving. See BENCHMARK.md for complete results.

🛡️ Hook Profiles

Change profile: set ERNE_PROFILE env var, add <!-- Hook Profile: standard --> to CLAUDE.md, or use /vibe context.

⚡ Commands

🖥️ IDE & Editor Support

ERNE generates adaptive config files for multiple IDEs, but the full agent experience requires Claude Code:

Generated config files:

🏗️ Architecture

Claude Code Hooks ──▶ run-with-flags.js ──▶ Profile gate ──▶ Hook scripts
                                                │
                                     ┌──────────┴──────────┐
                                     │   Only hooks for    │
                                     │   active profile    │
                                     │   are executed      │
                                     └─────────────────────┘

erne dashboard ──▶ HTTP + WS Server ──▶ Browser Canvas
                        ▲
Claude Code PreToolUse ─┤  (Agent pattern)
Claude Code PostToolUse ┘

Key design principles:

• 🪶 Zero runtime dependencies for the harness itself (ws package only for dashboard)
• 🎯 Conditional loading — rules, skills, and hooks load based on project type and profile
• 🧹 Fresh subagent per task — no context pollution between agent invocations
• 🔇 Silent failure — hooks never block Claude Code if something goes wrong

🤝 Contributing

We welcome contributions from everyone — from typo fixes to new agents and skills.

git checkout -b feat/your-feature
npm run validate && npm test   # Must pass before PR

🤝 Partnerships

Skills, agents, and MCP configs are open source — anyone can add them via PR. Partnerships are for deeper collaboration:

If you maintain a React Native library, Expo tool, or developer service — let's talk.

📦 Available On

• npm — npx erne-universal init
• SkillsMP — Auto-indexed from GitHub
• BuildWithClaude — Plugin directory
• VoltAgent/awesome-agent-skills — Curated skills list

📖 Documentation

📜 License

MIT License — use freely, commercially or personally.