Audit Methodology

How the Blockhertz AI
Smart Contract Auditor Works

A transparent explanation of how we analyze smart contracts, detect vulnerabilities, and generate security reports — and where our limitations are.

Try Free Audit

4

Languages supported

10+

Vulnerability classes

60s

Analysis time

Free

Always

How Does the AI Auditor Analyze Smart Contracts?

Six stages, from paste to PDF.

  1. STEP 1

    📥Input Handling

    You paste your smart contract source code directly into the auditor. We accept complete contract files or individual functions. Supported languages: Solidity, Rust, Move and Vyper. The contract is never stored permanently — it exists only for the duration of the analysis session.

  2. STEP 2

    🔍Static Analysis

    The contract code is parsed and analyzed for structural patterns associated with known vulnerability classes. We check function visibility, state variable ordering, external call patterns, access control modifiers, arithmetic operations, and event emissions against a library of known vulnerability signatures.

  3. STEP 3

    🤖AI Reasoning

    An AI model analyzes the contract's logic, identifying vulnerabilities that pattern matching alone misses — including business logic errors, incorrect state machine transitions, and complex multi-function reentrancy paths. The AI reasons about the contract's intended behavior versus its actual implementation.

  4. STEP 4

    📊Severity Classification

    Each finding is classified by severity: Critical (immediate exploit risk), High (significant vulnerability), Medium (moderate risk), and Low (best practice violation). Severity is determined by exploitability, impact on funds, and likelihood of real-world attack.

  5. STEP 5

    🛠️Fix Recommendations

    For each vulnerability found, the auditor generates specific fix recommendations — not generic advice, but contract-specific suggestions referencing the exact functions and line patterns that need changing. OpenZeppelin library alternatives are suggested where applicable.

  6. STEP 6

    📄Risk Score + Report

    A 0-100 risk score summarizes the overall contract security posture. Lower is better. Contracts scoring 80+ qualify for the Blockhertz Security Stamp verification program. The full report is exportable as PDF.

What Vulnerability Classes Does the Auditor Detect?

14 core classes, spanning pattern matching and AI reasoning.

VulnerabilitySeverityDetection Method
ReentrancyCriticalPattern + AI reasoning
Access ControlCriticalPattern matching
Integer OverflowHighArithmetic analysis
Oracle ManipulationCriticalAI reasoning
Flash Loan VectorsHighAI reasoning
Unchecked External CallsHighPattern matching
Delegatecall InjectionCriticalPattern matching
Timestamp DependenceMediumPattern matching
Self-DestructCriticalPattern matching
Signature ReplayHighAI reasoning
Unprotected InitializersCriticalPattern matching
Front-RunningMediumAI reasoning
tx.origin AuthHighPattern matching
Price Oracle ManipulationCriticalAI reasoning

Which Smart Contract Languages Are Supported?

Solidity

Ethereum + EVM chains

Rust

Solana programs

Move

Sui + Aptos contracts

Vyper

Ethereum alternative

What Are the Limitations of AI Smart Contract Auditing?

The Blockhertz AI Auditor is designed as a free first-pass pre-audit tool. It is not a replacement for a professional manual security audit.

Understanding what it does and does not catch:

What it catches well

  • Known vulnerability patterns
  • Common Solidity anti-patterns
  • Missing access control modifiers
  • Reentrancy in standard patterns
  • Integer overflow/underflow
  • Unchecked return values
  • Self-destruct risks
  • tx.origin authentication

What it may miss

  • Complex business logic errors
  • Protocol-specific economic attacks
  • Vulnerabilities requiring deep understanding of external protocol interactions
  • Novel attack vectors not in training data
  • Vulnerabilities introduced by contract upgrade patterns
  • Cross-contract interaction bugs in multi-contract systems

Recommendation

For any contract handling real user funds or assets, use the Blockhertz AI Auditor as a first pass to catch obvious issues, then commission a professional manual audit before mainnet deployment.

AI Audit vs Manual Audit: When to Use Each?

AI Audit (Free)Manual Audit
StageDevelopmentPre-mainnet
Time60 seconds1–4 weeks
CostFree$5K–$200K
Best forCatching obvious issuesProduction deployment
MissesComplex logicAlmost nothing
ReportInstant PDFDetailed document

Run a Free Audit Now

Paste your contract and get a complete security report in under 60 seconds. No signup. No credit card. Always free.

Open AI Auditor

Need a professional manual audit? View our smart contract development and security services →