Multichain Scam Deployer Graph is an advanced on-chain intelligence toolkit designed to identify, analyze, and visualize malicious smart-contract deployers across multiple blockchains. Using block-explorer API data, it builds a full relationship graph linking:

• Deployer addresses
• Contracts they created
• Chains they operate on
• Graph-level metrics (degree, centrality, expansion)
• Risk-based features suitable for ML models

This system enables deep blockchain forensics, including:

Detecting scammer clusters Identifying deployers shared across chains Quantifying deployer influence Profiling contract factories and rugpull creators Generating ML-ready datasets to detect high-risk deployers

The project outputs NetworkX graphs, GEXF files, CSV edge + node tables, and a full feature matrix with risk scores.

Designed for security analysts, data scientists, threat researchers, and anti-scam detection teams.

multichain-scam-deployer-graph/
│
├── README.md                        # You are reading this
├── requirements.txt                 # Dependencies
│
├── data/
│   ├── raw/                         # Raw JSON from block explorers
│   ├── graphs/                      # Output graph files (.gexf, nodes.csv, edges.csv)
│   └── features/                    # ML-ready deployer features
│
└── src/
    ├── config.py                    # Chain configuration + API key handlers
    ├── explorers.py                 # Unified Etherscan V2 fetcher
    ├── graph_builder.py             # Graph construction + metrics
    ├── features.py                  # Feature engineering for ML
    ├── risk_rules.py                # Heuristic risk scoring
    ├── cli_fetch.py                 # CLI: fetch raw deployer history
    ├── cli_build_graph.py           # CLI: build graph + export
    └── cli_features.py              # CLI: compute deployer features

Install dependencies:

pip install -r requirements.txt

Set block-explorer API keys:

set ETHERSCAN_API_KEY=YOUR_KEY
set BSCSCAN_API_KEY=YOUR_KEY
set ARBISCAN_API_KEY=YOUR_KEY
set BASESCAN_API_KEY=YOUR_KEY

export ETHERSCAN_API_KEY="YOUR_KEY"
export BSCSCAN_API_KEY="YOUR_KEY"
export ARBISCAN_API_KEY="YOUR_KEY"
export BASESCAN_API_KEY="YOUR_KEY"

Fetch the "normal" transactions for any deployer:

python -m src.cli_fetch --chain ethereum \
    --deployer 0xDEADBEEF... \
    --out data/raw/eth.json

This downloads:

• all transactions by the deployer
• filters ones that created contracts
• stores them as structured JSON

Example output:

{
  "chain": "ethereum",
  "deployer": "0xDEADBEEF...",
  "txs": [ ... ]
}

Repeat for:

bsc
arbitrum
base

Combine any number of raw JSON fetches:

python -m src.cli_build_graph \
  --inputs data/raw/*.json \
  --out_graph data/graphs/multichain.gexf \
  --out_nodes data/graphs/nodes.csv \
  --out_edges data/graphs/edges.csv

Graph for Gephi / Cytoscape visual analytics.

Node-level attributes:

Edges between nodes:

python -m src.cli_features \
  --nodes data/graphs/nodes.csv \
  --edges data/graphs/edges.csv \
  --out data/features/deployer_features.csv

Use this file for:

• clustering
• anomaly detection
• training ML models
• scam fingerprinting
• statistical analysis

flowchart LR
    A[Deployer Address<br/>0xABC...] -->|txlist v2| B[Block Explorer API<br/>Etherscan/BscScan/etc.]
    B --> C[data/raw/*.json<br/>Raw tx history]
    C --> D[Graph Builder<br/>NetworkX]
    D -->|export| E[data/graphs/multichain.gexf]
    D --> F[data/graphs/nodes.csv]
    D --> G[data/graphs/edges.csv]
    F & G --> H[Feature Engineering<br/>Risk Scoring]
    H --> I[data/features/deployer_features.csv<br/>ML-ready dataset]

The GEXF file represents the full deployer network:

• Deployer nodes are squares
• Contracts are circles
• Edges represent DEPLOYED relationships
• Degree indicates activity level
• Cross-chain deployers form bridges between subgraphs
• Clusters reveal scam families / contract factories

You can view this file using Gephi, which enables:

• Force Atlas 2 layout
• Degree-based sizing
• Chain-based coloring
• Risk-level overlays

Identify deployers who create many unrelated contracts across multiple chains.

Track scammers migrating from ETH → BSC → Base.

Generate deployer risk profiles (Low / Medium / High).

Spot deployers with unusually high out-degree patterns.

Use the feature CSV to train a classifier that detects high-risk deployers.

Trace entire scam ecosystems from a single deployer.

A deployer may be dangerous if:

• deploys 30+ contracts in short timeframe
• uses the same address across ETH/BSC
• creates mostly ERC-20 tokens with no verified source code
• deploys honeypots or high-tax tokens
• exhibits “clustered” behavior in graph space
• has high betweenness centrality (connects many scam subgraphs)

Your project reveals all of this automatically.

Used for graph construction and metric computation.

Your fetcher uses:

https://api.etherscan.io/v2/api

with chainid parameter.

Provides:

• Degree
• In-degree
• Out-degree
• Betweenness (optional)

ML analysts can integrate directly into:

• pandas
• scikit-learn
• PyTorch Geometric
• Graph ML pipelines

Planned enhancements:

• Multi-deployer batch scanning
• Automatic multi-chain merging
• Dynamic risk model (SciKit-Learn / XGBoost)
• Real-time Etherscan polling
• Scam cluster detection via graph embeddings (Node2Vec)
• PyVis interactive browser graph
• Solidity contract content analysis
• On-chain event analysis (Logs → graph edges)