Skip to content
Chain Defenders

Chain Defenders · Security audits

We audit the codeyour moneydepends on.

Smart contracts that hold it. AI systems that spend it. Two security researchers read every line, and you get back what is wrong, ranked by what it would cost you.

H-1HighExample

Ticket triage runs on the largest model. The smallest passes the same tests.

Input, per million tokens$4.00 → $0.10

Claude Opus 5.5 vs Claude Haiku 5.5, published prices on 2026-10-09

Example finding, AI cost audit

One method

The same audit,
whatever the code.

A smart contract and an AI system fail in different ways. We look for both the same way: scope it, read it by hand, report what it would cost, and check the fixes.

01

Scope

What is in, what good looks like, and a fixed quote before anything starts.

Smart contracts
The commit, the threat model, the money at stake.
AI systems
Last month's invoice and the features behind it.
02

Read

Both of us, by hand, assuming something is wrong until we can show it is not.

Smart contracts
Every line of Solidity or Rust, plus fuzzing where it pays.
AI systems
Every prompt, model choice and call path that costs money.
03

Report

Findings ranked by what they would cost you, written for a founder to act on.

Smart contracts
Severity by funds at risk, with a fix for each.
AI systems
Euros a month for each leak, with the change that stops it.
04

Verify

We check the fixes, because a report only matters once they are in.

Smart contracts
A fix review of the patched code.
AI systems
The saving read off your next invoice.

The record

What we have found so far.

Every finding in our public audit portfolio, read live from GitHub, so you can check any number against the report behind it. Counting private work that is not listed, that is over 100 critical or high findings. The same reading is what we bring to AI systems.

0

findings in the public portfolio

0
private audits
0
contests
0
top-five finishes
  • 89Critical or high
  • 172Medium
  • 372Low
  • 262Informational
  • 56Gas
Having seen their excellent track record on Code4rena, I engaged their audit service and they delivered exactly as expected. The team provided swift responses with clear findings, demonstrating both technical expertise and professional communication through efficient yet thorough analysis. Their excellent balance of quality and speed makes them particularly recommendable for teams seeking rapid turnaround without compromising on audit quality.

@Zak3939.eth

Phi

Worked with them twice now and honestly, they're probably the best auditors we've tried. Super flexible, gave really clear feedback, and caught everything we needed to fix. Shoutout to @PeterSRWeb3 and @1337web3 - you guys made this process way smoother than expected.

@lib_defi

Pandemic Labs

Who reads your code

Two people. You talk to the ones doing the work.

0x539.eth and PeterSR run every audit themselves. There is no account manager and no hand-off.

0x539.eth

0x539.eth

Co-founder, Security and AI

Web3 security sleuth who thrives on outsmarting blackhat hackers, with a knack for spotting critical bugs and a belief that competition among auditors is the secret sauce to a safer decentralised world.

Audits · AI integration · AI cost reduction

PeterSR

PeterSR

Co-founder, Security and AI

Web3 security enthusiast with a Web2 background at SAP and VMware, obsessed with auditing every spare minute and convinced that fuzz testing and pen-and-paper notes are the keys to crushing vulnerabilities.

Audits · AI integration · AI cost reduction

Tell us what you are working on.

A contract, a protocol or an AI bill. The first conversation is free, and a person replies.

Message us on Telegram