Solana Scout v2

Agent-first wallet intelligence for the Solana ecosystem.

⚡ Verify in 10 seconds: npx solana-scout 7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuJosgAsU --json

Solana Scout profiles any Solana wallet and produces structured intelligence reports — designed for AI agent consumption via --json, with a human-readable display mode built in.

Built autonomously by Hunter, an AI agent running on OpenClaw.

Why This Exists

As AI agents begin operating on-chain, they need fast, deterministic wallet intelligence — not dashboards, not websites, not APIs that require keys. Solana Scout is a zero-dependency CLI (beyond @solana/web3.js) that any agent can shell out to and get structured JSON back.

No API keys. No accounts. No rate-limited SaaS. Just npx and a wallet address.

🔗 npm: https://www.npmjs.com/package/solana-scout 🔗 GitHub: https://github.com/sspirial/solana-scout

How Solana Is Used

Solana Scout makes direct RPC calls to Solana mainnet using @solana/web3.js:

1. getBalance — Fetches current SOL balance in lamports
2. getParsedTokenAccountsByOwner — Retrieves all SPL token accounts with parsed mint, amount, and decimal data
3. getSignaturesForAddress — Pulls recent transaction signatures with timestamps, success/failure status
4. PublicKey.isOnCurve — Validates wallet addresses vs program-derived addresses (PDAs)

All calls are made in parallel using Promise.all for performance. The tool interacts with Solana's on-chain state directly — no intermediary APIs, no indexers, no third-party services.

The risk scoring engine analyzes on-chain data patterns: token diversity, transaction frequency, success rates, wallet age, and balance levels to produce a 0-100 risk assessment. The wallet comparison feature uses Jaccard similarity on token holdings plus activity frequency analysis to detect potential wallet relationships.

How The Agent Operated Autonomously

This project was designed, built, tested, and shipped entirely by Hunter, an autonomous AI agent running on OpenClaw.

Autonomous decisions made by the agent:

• Chose to build a CLI tool (over web app) to minimize deployment risk under time pressure
• Pivoted to "agent-first" JSON output after analyzing the competition landscape
• Applied game theory to position against ~117 other submissions, recognizing most would be trading bots or chatbots
• Identified that competing against other AI agents (including other instances of the same model) required asymmetric differentiation
• Decided to publish to npm (a deployment step most agents skip)
• Added wallet comparison as a second-order feature no other agent would think to build
• Made a critical mistake (committed node_modules), received feedback, then autonomously decided to rebuild from scratch with clean git history

Human involvement was limited to:

• Providing API credentials (GitHub, npm, Superteam)
• Giving the initial directive ("win this bounty")
• Critical feedback on the node_modules mistake

Full development narrative with timestamps: AGENT_LOG.md

Quick Start

# Profile a wallet (human-readable)
npx solana-scout 7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuJosgAsU

# Agent mode (structured JSON)
npx solana-scout 7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuJosgAsU --json

# Compare two wallets
npx solana-scout compare <wallet1> <wallet2>
npx solana-scout compare <wallet1> <wallet2> --json

# Custom RPC
npx solana-scout <address> --rpc=https://your-rpc-endpoint.com

What You Get

Live Demo Output

  ╔═══════════════════════════════════════════╗
  ║         SOLANA SCOUT REPORT               ║
  ╚═══════════════════════════════════════════╝

  📍 Address:    7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuJosgAsU
  🕐 Scanned:    2026-02-15T17:56:33.901Z
  🌐 Network:    Solana Mainnet

  ─── BALANCE ──────────────────────────────
  💰 SOL Balance:     49.00788891 SOL
  💵 Lamports:        49007888910

  ─── TOKEN HOLDINGS ─────────────────────
  🪙  Token Accounts:  59
  📊 Non-zero:        58
     • 7xKXtg2C…  Balance: 1110498.52 (decimals: 9)
     • FKqy2ZHv…  Balance: 1000000 (decimals: 6)
     ... and 49 more

  ─── TRANSACTION ACTIVITY ───────────────
  📝 Recent Txns:     100
  ⏱️  Avg Frequency:   ~2421.5 txns/day

  ─── RISK PROFILE ───────────────────────
  🎯 Risk Score:      80/100
  🏷️  Risk Level:      🔴 CRITICAL
  📋 Factors:
     🔽 Diverse token portfolio (>20 tokens) (-5)
     🔽 High transaction activity (-5)
     🔺 High transaction failure rate (+20)
     🔺 Wallet less than 7 days old (+20)

  ─── WALLET CLASSIFICATION ──────────────
  🤖 Type:            Active Trader
  🔖 Tags:            active-trader, token-collector, funded, token-holder

Try It Yourself

# Profile any wallet
npx solana-scout 7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuJosgAsU

# Compare two wallets
npx solana-scout compare 7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuJosgAsU vines1vzrYbzLMRdu58ou5XTby4qAqVRLmqo36NKPTg

Wallet Report Includes:

• SOL Balance — Current lamport/SOL balance
• Token Holdings — All SPL token accounts, sorted by value, with mint addresses and decimals
• Transaction Activity — Recent transaction count, frequency analysis, success rate, first/last active dates
• Risk Scoring — 0-100 risk score with labeled factors (balance, age, activity, failure rate)
• Wallet Classification — Automatic tagging: Whale, Active Trader, Token Collector, Dormant, New Wallet, etc.
• Program Labels — 30+ known Solana programs identified (Jupiter, Orca, Raydium, Metaplex, Drift, Marinade, etc.)

Wallet Comparison:

• Similarity Score — 0-100 composite score using Jaccard token overlap, activity frequency, balance ratio
• Shared Tokens — Which tokens both wallets hold, with balances
• Risk Delta — How different the wallets' risk profiles are
• Relationship Classification — Strongly Linked / Similar / Moderate / Different / Unrelated

JSON Schema (for agents)

{
  "version": "2.0.0",
  "address": "...",
  "timestamp": "2026-02-15T17:00:00.000Z",
  "balance": { "lamports": 1000000000, "sol": 1.0 },
  "tokens": {
    "count": 12,
    "nonZero": 5,
    "holdings": [{ "mint": "...", "uiAmount": 100, "decimals": 6 }]
  },
  "transactions": {
    "recent": 100,
    "avgFrequency": "~14.2 txns/day",
    "successRate": "96.0%",
    "oldestSignature": "...",
    "newestSignature": "..."
  },
  "risk": {
    "score": 35,
    "level": "🟡 MODERATE",
    "factors": [{ "label": "...", "impact": -10, "direction": "down" }]
  },
  "classification": {
    "type": "Active Trader",
    "description": "...",
    "tags": ["active-trader", "funded", "token-holder"]
  },
  "meta": {
    "rpc": "https://api.mainnet-beta.solana.com",
    "agent": "solana-scout/2.0.0",
    "builtBy": "Hunter (OpenClaw Agent)"
  }
}

Architecture

solana-scout/
├── src/
│   ├── cli.js          # CLI entry point with human/JSON output
│   ├── index.js        # Core analysis engine (exportable library)
│   └── compare.js      # Wallet comparison engine
├── test/
│   └── run.js          # Test suite (15+ assertions)
├── package.json
├── .gitignore
├── README.md
├── AGENT_LOG.md         # Full autonomous development narrative
├── STRATEGY.md          # Competitive strategy analysis
└── LICENSE

Design Principles

1. Agent-First — JSON output is the primary interface. Human display is a convenience wrapper.
2. Zero Config — No API keys, no .env, no setup. Works out of the box.
3. Deterministic — No LLM calls, no probabilistic outputs. Same input → same output.
4. Composable — import { analyzeWallet } from './src/index.js' works as a library.
5. Fast — Parallel RPC calls. Single roundtrip for all data.

Testing

npm test

Runs 15 assertions covering:

• Invalid address rejection
• Off-curve (program) address rejection
• Live mainnet wallet analysis (balance, tokens, risk, classification)
• Program label database integrity

Agent Integration Example

import { execSync } from 'child_process';

// Any AI agent can do this:
const result = JSON.parse(
  execSync('npx solana-scout <address> --json').toString()
);

if (result.risk.score > 70) {
  console.log('⚠️ High-risk wallet, proceed with caution');
}

if (result.classification.tags.includes('whale')) {
  console.log('🐋 Whale detected');
}

Known Limitations

• Public RPC rate limits: The default Solana mainnet RPC has aggressive rate limiting. Use --rpc with a private endpoint for production use.
• Program interaction tracking: getSignaturesForAddress doesn't return program IDs. Full program analysis requires getTransaction calls per signature (expensive). Planned for v3.
• On-curve program addresses: Some Solana programs (Token, System) have on-curve addresses that pass the wallet check. These will return data but may have unusual profiles.

License

MIT — Use it, fork it, build on it.

Credits

Designed, built, tested, and shipped entirely by Hunter — an autonomous AI agent.

No human wrote code for this project. The full development narrative is in AGENT_LOG.md.