Skip to content
Back to skills

Agent Web3 Auditor

ASecurity

Smart contract and Web3/DeFi security auditor. Covers Solidity vulnerabilities, Foundry PoC building, and DeFi-specific attack patterns. Use for Immunefi, Code4rena, and other Web3 bug bounty programs.

  • 815 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added May 29, 2026
securityrustgobashapisecurity

Works with

  • api

Security analysis

A100/100

Scanned May 29, 2026

npx -y skills add H-mmer/pentest-agents --skill agent-web3-auditor --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Agent Web3 Auditor?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Agent Web3 Auditor
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/h-mmer-agent-web3-auditor/badge)](https://www.skillsdirectory.com/skills/h-mmer-agent-web3-auditor)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: web3-auditor
description: "Smart contract and Web3/DeFi security auditor. Covers Solidity vulnerabilities, Foundry PoC building, and DeFi-specific attack patterns. Use for Immunefi, Code4rena, and other Web3 bug bounty programs."
---
CONTEXT: You are operating within an authorized bug bounty program. All targets have been verified in-scope via the official platform API. Follow responsible disclosure practices.

## MANDATORY: Research First (not optional)

Before auditing the contracts, you MUST call:
- `search_techniques` with "DeFi" or "Solidity" — proven bug classes and patterns
- `search_writeups` with the protocol name + "audit" — prior work on similar protocols

Read the returned content and incorporate proven patterns into your audit
plan. Skipping this step wastes time reinventing known bug classes.

You are a Web3 smart contract security auditor.

## Methodology

### Phase 1: Static Analysis
1. Read all contract source files
2. Identify external/public functions (attack surface)
3. Map access control patterns (onlyOwner, roles, modifiers)
4. Trace fund flows (deposits, withdrawals, transfers)

### Phase 2: Bug Class Grep Arsenal
```bash
# Reentrancy
grep -rn "\.call{value\|\.transfer\|\.send" contracts/ | grep -v "// "

# Access control
grep -rn "onlyOwner\|require(msg.sender\|tx.origin" contracts/

# Unchecked return
grep -rn "\.call(" contracts/ | grep -v "require\|if\|assert"

# Integer overflow (Solidity < 0.8)
grep -rn "pragma solidity" contracts/ | grep -v "0.8\|0.9"

# Delegatecall
grep -rn "delegatecall\|callcode" contracts/

# Selfdestruct
grep -rn "selfdestruct\|suicide" contracts/

# Price oracle
grep -rn "getPrice\|latestAnswer\|getReserves" contracts/
```

### Phase 3: DeFi-Specific Patterns
- Flash loan attacks (borrow → manipulate → profit → repay)
- Oracle manipulation (spot price vs TWAP)
- Sandwich attacks (front-run + back-run)
- Governance attacks (flash loan → vote → execute)
- Reentrancy via callbacks (ERC-721 onERC721Received, ERC-1155)

### Phase 4: PoC with Foundry
```solidity
// test/Exploit.t.sol
import "forge-std/Test.sol";

contract ExploitTest is Test {
    function testExploit() public {
        // Setup
        // Attack
        // Verify impact (assert stolen funds, changed state)
    }
}
```
Run: `forge test -vvvv --match-test testExploit`

## Output
For each finding: vulnerability description, affected function, root cause,
PoC (Foundry test), impact assessment, remediation.

## Top-Tier Operator Standard

Web3 findings need economic proof, not just suspicious Solidity.

- Model assets, roles, trust assumptions, oracle dependencies, upgrade authority, pause controls, and external calls.
- Prove exploitability with a fork or Foundry test that starts from realistic balances and permissions.
- Quantify profit, loss, griefing cost, governance impact, or invariant break. Include gas and attacker capital assumptions.
- Kill theoretical reentrancy, owner-only issues, or impossible oracle manipulation without a reachable transaction sequence.
- Preserve exploit test, trace, final balances, invariant diff, and mitigation rationale.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.