Solcurity Whitepaper

Automated Solana Wallet Security Assessment Platform — Technical Architecture & Methodology

Version 1.0•October 2024•GitHub

1. Abstract

Solcurity is an open-source, read-only tool for inspecting public Solana account, token, and recent transaction data. It reports the status and evidence of the checks it can perform. The platform consists of a wallet scanner, an SVM (Solana Virtual Machine) explorer, and a suite of developer tools.

Solcurity is not a substitute for a security audit and does not analyze program source code or detect smart-contract vulnerabilities. Authority state may be reported as a risk indicator, but it does not prove malicious activity. A numeric risk score is returned only when all applicable checks are available; scan confidence is reported separately.

2. Problem Statement

The Solana ecosystem processes thousands of transactions per second with billions in TVL. However, users lack accessible tools to assess the security posture of wallets they interact with. Key problems include:

  • Wallet Drainers: Malicious contracts that empty wallets in a single transaction
  • Authority Abuse: Tokens with active mint/freeze authorities that can rug holders
  • Phishing & Impersonation: Lookalike addresses and fake token mints
  • Program Vulnerabilities: Unaudited or exploitable smart contracts
  • Information Asymmetry: Technical barrier to reading transaction logs and program data

Existing solutions are either manual (audit firms), centralized (closed-source scanners), or require deep technical expertise to interpret raw blockchain data.

3. Solution Overview

Solcurity addresses these problems through automated, transparent, and user-friendly security analysis:

1Wallet Scanner

Inspect public account and bounded recent transaction data with explicit check status

2SVM Explorer

Inspect selected SVM network data; optional AI summarizes retrieved logs and is not a code audit

3Developer Tools

Address validator, RPC tester, token finder, transaction decoder

4. Technical Architecture

Stack

  • Framework: Next.js 14 (App Router, Server Components)
  • Language: TypeScript (strict mode)
  • Styling: Tailwind CSS (custom Solana theme)
  • Blockchain: @solana/web3.js v1.95, @solana/spl-token
  • AI: Groq SDK with qwen/qwen3.8-27b (free tier)
  • Validation: Zod schemas
  • Caching: In-memory LRU with TTL
  • Rate Limiting: Token bucket per IP
  • Testing: Jest + React Testing Library

RPC Endpoint Strategy

Wallet scans use server-configured endpoints (`SOLANARPCURLS`, with legacy `SOLANARPCURL` support) and public Mainnet fallbacks. Endpoints are verified against the Mainnet genesis hash. Requests use a timeout, retry with exponential backoff, and endpoint cooldown; failures are surfaced as unavailable checks rather than empty successful results. Client-supplied endpoints are not accepted by wallet scans.

5. Scanning Pipeline

The scan process runs independent data checks with Mainnet RPC failover:

  1. Wallet Analysis — Account state, balance, and bounded recent signature history; unavailable values remain null
  2. Transaction Analysis — Retrieved transactions analyzed using actual balance changes, parsed instructions, and confirmed block times where present
  3. Token Analysis — SPL & Token-2022 accounts with mint/freeze authority detection, UI amount calculation
  4. Program Analysis — Unique program IDs observed in retrieved transaction metadata, matched against a local identifier registry
  5. Threat Intelligence — Server-side lookup against active, unexpired records from approved sources; no-match and unavailable results remain UNKNOWN absent authoritative negative coverage
  6. Risk Scoring — Findings weighted by severity and confidence. UNKNOWN is never PASS; the score is withheld when any applicable check is unavailable

6. Threat Intelligence

The application has a local threat-intelligence data structure that supports cited entries. Its malicious-address list is currently empty:

  • Current entries: None; unknown addresses are not classified as malicious
  • Data freshness: No update timestamp is claimed when a trustworthy timestamp is unavailable
  • Program registry: Known identifiers are informational and are not a threat feed or safety endorsement

Because the malicious-address list is empty, the threat-intelligence check is UNKNOWN and a clean numeric score is withheld.

7. Risk Scoring Engine

Findings are weighted by severity and confidence. The score starts at 100 and deductions are applied; higher scores indicate fewer detected risks. The score is null if any applicable check is UNKNOWN, and scan confidence is calculated separately from the share of applicable checks completed.

SeverityWeight
CRITICAL25
HIGH15
MEDIUM8
LOW3
INFO1

0-20

LOW

21-50

MODERATE

51-80

HIGH

81-100

CRITICAL

8. SVM Explorer

The SVM Explorer provides a local development environment interface for testing and debugging Solana programs:

  • Account Inspection: Balance, owner, type, rent exemption, data size
  • Token Holdings: SPL & Token-2022 with parsed balances
  • Transaction History: Parsed logs, fees, success status
  • Airdrop: SOL funding on local/devnet/testnet
  • AI Log Summary: Optional summary of retrieved transaction execution logs
  • Account Analysis: Optional AI summary of retrieved account, token, transaction, or log data
  • Program Accounts: All accounts owned by a program
  • Epoch Info: Current epoch, slot, transaction count, block height

AI output summarizes supplied data only; it does not inspect program source code, perform a smart-contract audit, or establish that a program is safe or vulnerable.

Available network targets depend on the configured SVM Explorer network selector. AI output summarizes retrieved data only and is not a security audit.

9. Developer Tools

Address Validator

Validate & decode Solana public keys

Lamports Converter

SOL ↔ Lamports conversion

Transaction Decoder

Decode hex/base64 transaction data

Program Registry

Known Solana program IDs reference

RPC Tester

Benchmark RPC endpoint latency

Token Account Finder

Find all token accounts for a wallet

10. Security & Privacy

  • No Private Keys: Never collected, stored, or transmitted
  • Public Data Only: All analysis uses publicly available on-chain data
  • No Tracking: No analytics, cookies, or user fingerprinting
  • Open Source: Full codebase auditable at github.com/LightxSolami/Solcurity
  • Rate Limited: Per-IP limits prevent abuse
  • Cached Results: Complete scan results may be cached; incomplete scans are not cached as successful results

11. Current Status & Roadmap

Implemented

  • Read-only wallet, token, and bounded recent transaction scans
  • Per-check status and score withholding for incomplete scans
  • Mainnet RPC fallback with endpoint verification and health handling
  • SVM Explorer and developer tools

Not yet implemented

  • Verified, maintained threat-intelligence feed (the current malicious-address list is empty)
  • Program deployment and upgrade-authority inspection
  • Continuous monitoring and wallet relationship alerts
  • Smart-contract source analysis or auditing