AI Coding Agents Raise DeFi Security Stakes as Ex-OpenZeppelin CTO Warns Investors
DeFiSecurity
Bearish

AI Coding Agents Raise DeFi Security Stakes as Ex-OpenZeppelin CTO Warns Investors

Manuel Aráoz, former CTO of OpenZeppelin, advised investors on May 27 to exit DeFi positions, arguing AI coding agents make smart contract attacks cheaper to mount. OpenZeppelin and Aave's founder pushed back, saying recent losses stemmed mainly from operational failures rather than audited code.

Nov 7, 2026, 06:02 PM3 min read

Written by CoinArticle’s AI Newsroom · from 2 cited sources. How we work

The Warning and Its Basis

On May 27, Manuel Aráoz, co-founder and former chief technology officer of OpenZeppelin, advised investors to exit DeFi positions, including exposure to lending protocols Aave, MakerDAO and Compound. He wrote that "coding agents are superhuman at finding vulnerabilities, and smart contract security is too asymmetric," adding that defenders must fix every bug while attackers need only one exploit.

The warning arrives during a weak period for the sector. Per the source reporting, DeFi has lost more than $1.1 billion to exploits over the past year, with $635 million lost across 28 reported hacks in April. Total value locked fell from roughly $172 billion in mid-April to $148 billion at press time, marking five consecutive weeks of outflows. Broader market weakness, with Bitcoin approaching $72,000, also contributed to the decline.

Research Supporting the Concern

Venture firm a16z has reported that AI agents consistently identified core vulnerabilities in historical DeFi exploits. Even when agents did not complete an exploit, they often reached the stage that gives an attacker a starting point, which lowers the expertise needed to begin an attack. Anthropic has restricted public access to its unreleased Claude Mythos model for similar reasons, citing its capacity to autonomously discover and weaponize software flaws.

The concern is structural. Many DeFi protocols are public, composable and financially liquid, so their code, governance and integrations can be studied openly. AI can make that study faster and cheaper, placing pressure on teams whose defenses still rely heavily on point-in-time audits, bug bounties and manual review.

Where Industry Figures Disagree

OpenZeppelin has pushed back on the broader framing. The firm argues that many recent incidents stemmed from operational failures rather than flaws in audited contract code. According to OpenZeppelin, most large losses in recent months involved stolen private keys, bridge spoofing, social engineering and access control issues, suggesting attackers have increasingly targeted teams, permissions and infrastructure around protocols. The firm describes the threat from AI as real but manageable.

Aave founder Stani Kulechov made a similar case, saying DeFi now benefits from stronger risk engines, formal verification, audits, bug bounties, cap management, oracle improvements, automated monitoring and circuit breakers. He argued much of the remaining attack surface lies outside the core contract logic.

The sources do not fully agree on the cause of recent losses. Aráoz's framing centers on contract-level vulnerability discovery becoming cheap, while OpenZeppelin and Kulechov point to operational and key-management failures as the dominant source of losses. Neither side's position has been independently quantified in the reporting, and the split in attack categories is drawn from the firms' own characterization rather than a published breakdown.

What the Debate Changes

The discussion has moved beyond individual protocol incidents toward whether AI has lowered the cost of attacking DeFi faster than defenses improve. Both camps appear to agree that point-in-time audits alone are insufficient, a point that increasingly shapes how protocols think about continuous monitoring, circuit breakers and operational security.

Why It Matters

For Traders

Five weeks of TVL outflows and a high-profile exit call may keep pressure on lending protocol positions, though the sources attribute much of the decline to broader market weakness.

For Investors

The debate points to continuous security monitoring and operational controls becoming a durable differentiator for DeFi protocols beyond audit history.

For Builders

Teams should treat key management, bridge permissions and access control as attack surfaces on par with contract logic when hardening systems against AI-assisted discovery.

This article is for information only and is not financial advice. Read the full disclaimer.

Live prices:AaveBitcoin

Related Articles

Latest News