Blockchain investigator ZachXBT said on Monday that the $250,000 Ripple paid the security firm Veria Labs for finding a bug that could have minted trillions of XRP was far too small.
The situation has reignited ZachXBT’s long-running criticism of how Ripple treats researchers.
How much did Ripple pay Veria Labs?
ZachXBT criticized the $250,000 awarded to Veria Labs by Ripple in a post on X on October 12. He wrote that the finding “should be a 7 fig pay day at minimum” for Veria Labs and co-founder Cayden Liao.
ZachXBT said that the payout is part of a trend of “well capitalized organizations” shortchanging white-hat security researchers.
He argued that $250,000 was too little when the paying organization’s “treasury has billions,” and warned that thin rewards could tempt the next person who finds such a flaw to exploit it instead of disclosing it.
Liao, for his part, said that Ripple paid Veria the program’s maximum $250,000 bounty and called it the largest ever amount paid for a vulnerability found entirely by an AI agent.
Veria Labs found a critical integer overflow in the XRP Ledger’s payment engine, the code that settles trades against the network’s built-in order book.
When one payment filled many offers at once, the engine added up the amounts without checking for overflow. A big enough total wrapped around to a tiny number. Sellers were paid in full, but the buyer was charged only the tiny number.
This difference created new XRP that could be spent the same as any other.
The ledger has a safety check, “XRPNotCreated,” meant to block any transaction that creates new coins, but it had a similar bug, so it missed the imbalance.
Veria said its AI agent chained the two bugs, written in November 2015 and February 2017, into a working proof of concept. It credited 18.45 trillion XRP across 256 accounts in a single transaction. That is about 184 times XRP’s fixed supply of 100 billion. Veria estimated that the amount at risk was nearly the token’s entire $94 billion market value, based on a price of $1.59.
Notably, developers found no sign the bug was ever used. Ripple’s report on the issue explained that for the attack to occur, hundreds of offers would need to be priced in ways no real trader would use.
Veria reported the flaw on September 22 through the XRPL bug bounty program and received a $250,000 reward for its efforts.
ZachXBT referred to the $1.5 million that Zcash awarded to researcher Taylor Hornby in his argument against Ripple’s payment.
Hornby disclosed a counterfeiting bug in Zcash’s Orchard shielded pool, a flaw that put roughly $2.3 billion of ZEC at risk, and received the $750,000 bounty plus a matching $750,000 bonus grant that was approved by coinholders in a third-quarter retroactive round that distributed about $8.39 million.
Shielded Labs executive director Jason McGee, who filed the bonus nomination, argued the original amount was only about 0.033% of the funds at risk against a lowest precedent of 0.24%. He wrote that a higher reward was “cheap insurance” for the protocol. Like Veria’s find, Hornby’s discovery was AI-assisted.
Has the bug been fixed?
RippleX, Ripple’s engineering team, confirmed the bug the same day Veria reported it. After finding that the XRP created by the bug could actually be spent, they rated it critical. A fix was merged on September 23 and released in xrpld 3.4.1 on September 25.
Typically, a fix like this would go through the amendment process, which requires more than 80% of trusted validators to support it for two weeks straight, but this time the patch took effect as each server upgraded instead.
By the day of release, over 80% of validators on the default Unique Node List were already running it.
The company’s report stated that this was the first transaction-processing change in more than ten years that did not go through the amendment system due to the fact that the bug would have been exposed for weeks before the fix was released.
Still, delaying the public source release drew criticism. Justin Bons, founder of Cyber Capital, said XRP had been “running closed-source code” until the full 3.4.1 code was published.
RippleX engineer Mayukha Vadari responded to the criticism, saying that the old approach no longer works. She explained that a critical patch can’t be included into a routine public release because attackers would spot it and reverse-engineer it right away.
The smartest crypto minds already read our newsletter. Want in? Join them.