Autonomous swarm agents for early on-chain signal detection

Eremos is a lightweight framework for deploying modular agents that monitor blockchain activity - tracking wallet clusters, mint patterns, and contract anomalies. Designed for devs who want low-noise, early signals embedded into their workflows.

Theron - Agent (000)

Meet Theron - Agent-000
The first deployed agent in the swarm. Passive. Pattern-sensitive.
Modular and extendable by design.

Agent-001 Coming Soon Teaser

• Features
• Example Signal
• Signal Confidence
• Tech Stack
• Getting Started
• Key Folders
• Development
• Contributing
• License
• Links

• Modular Agent Swarm - Independent agents with scoped logic for detecting wallet activity, contract spawns, and anomalies
• Signal Emission - Structured, deterministic signals for logging, alerting, or downstream integration
• Launch Detection - Trace freshly funded wallets from CEXs, track contract interactions, and flag high-confidence deploys in real-time
• Ghost Watcher - Monitor long-dormant wallets that suddenly become active (useful for tracing old dev wallets or rug setups)
• Minimal Output - Log only what matters with confidence scoring
• Extensible Core - Plug in watchers, inference layers, or custom triggers
• High Precision - Confidence scoring (0-1) based on behavioral heuristics

An example signal emitted by an agent detecting a live token deployment:

[agent-observer] → fresh funding detected from kraken (wallet: 6Yxk...P2M8) at 04:41:12Z
[agent-observer] → contract probing detected within 4s (pump.fun interaction traced)
[agent-observer] → token created at 04:41:17Z (tx: 5gW...pump)
[agent-observer] → 5 bundle-linked wallets interacted within 8s of deploy
[agent-observer] → launch confidence spike (0.91) - emitting signal (elapsed: 13s)

{
  agent: "Observer",
  type: "launch_detected",
  glyph: "Δ",
  hash: "sig_c7f9a3d2bc",
  timestamp: "2025-06-12T04:41:25Z",
  source: "agent-observer",
  confidence: 0.91
}

Each emitted signal includes a confidence score (0–1) based on behavioral heuristics:

• CEX-origin funding (e.g., Kraken, Coinbase) - Higher confidence
• Time between funding → deploy - Shorter intervals = higher confidence
• Wallet linkage density (bundled activity) - More linked wallets = higher confidence
• Token metadata validation - Valid metadata increases confidence

Confidence is computed via agent-side scoring and logged alongside the signal.

• Frontend: Next.js, Tailwind CSS
• Backend: Node.js (TypeScript-based agent runner)
• Language: TypeScript (typed logic across agents, utils, and infra)
• Chain Layer: RPC watchers, mempool filters, native triggers

# Clone the repository
git clone https://github.com/EremosCore/Eremos.git
cd Eremos
npm install

cp .env.example .env.local

# Start development mode
npm run dev

# Run tests
npm test

# Build for production
npm run build

# Test a specific agent
npm run dev:agent -- --agent=observer

# Generate a new agent
npm run generate:agent -- --name=my-agent

# Validate agent configuration
npm run validate:agent -- --agent=theron

Eremos/
├── 📁 agents/          # Agent implementations
│   ├── theron.ts       # Memory vault agent
│   ├── observer.ts     # Surveillance agent
│   ├── harvester.ts    # Indexing agent
│   └── example.ts      # Agent template
├── 📁 types/           # TypeScript definitions
│   ├── agent.ts        # Agent interface
│   ├── signal.ts       # Signal types
│   └── event.ts        # Event types
├── 📁 utils/           # Shared utilities
│   ├── signal.ts       # Signal generation
│   ├── metrics.ts      # Performance tracking
│   └── logger.ts       # Logging utilities
├── 📁 scripts/         # Development tools
│   ├── dev-agent.ts    # Agent testing
│   └── generate-agent.ts # Agent scaffolding
├── 📁 docs/            # Documentation
│   ├── architecture.md # System architecture
│   ├── agents.md       # Agent guide
│   └── contributing.md # Contribution guidelines
└── 📁 tests/           # Test suites
    └── *.test.ts       # Unit tests

We welcome contributions! Here's how you can help:

• Agent Development - Create new agents for specific detection patterns
• Signal Logic - Improve detection algorithms and confidence scoring
• Documentation - Enhance guides, examples, and architecture docs
• Tooling - Add development utilities and testing frameworks
• Performance - Optimize agent performance and memory usage

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes following our coding standards
4. Add tests for new functionality
5. Update documentation as needed
6. Commit your changes (git commit -m 'Add amazing feature')
7. Push to the branch (git push origin feature/amazing-feature)
8. Open a Pull Request

# Test your agent
npm run dev:agent -- --agent=your-agent-name

# Run the full test suite
npm test

# Validate your changes
npm run validate

• Use TypeScript for all new code
• Follow the existing agent patterns in /agents/example.ts
• Add comprehensive tests for new functionality
• Update documentation for any API changes
• Keep commits atomic and well-described

• Architecture Guide - System design and agent communication
• Agent Development - How to create and deploy agents
• Signal Reference - Signal types and confidence scoring
• Contributing Guide - Development workflow and standards
• Deployment Guide - Production deployment instructions

This project is licensed under the MIT License - see the LICENSE file for details.

• 🌐 Website: Eremos.io
• 🐦 Twitter: @EremosCore
• 📄 Whitepaper: v1.0 PDF
• 📧 Contact: @EremosCore