Close-up of colourful programming code on a computer monitorFile photo: Programming code on a computer monitor. Photo: Markus Spiske via Wikimedia Commons (CC0)

An OpenCode breach exposed AI provider keys saved by 1.4% of users of the coding tool’s paid Console. Developers say strangers ran up hundreds, in some cases thousands, of dollars on their OpenAI, Anthropic and Google accounts overnight. Anthropic and Google have revoked or blocked keys, and OpenCode says refunds are underway.

By HahuPrime News Desk | October 18, 2026

At 2 a.m. New York time on Saturday, someone started using Colin Parsons’s OpenAI key. The key was “pounded” for $281 of usage, and his Anthropic key for another $168, he wrote on X. He said he had only ever used both keys in one place: OpenCode, a popular open-source AI coding assistant.

Hours later the company confirmed the OpenCode breach that Parsons and others had already guessed at. It had suffered “a security breach last night related to a large scale distillation attack that impacted 1.4% of OpenCode Console users”, it said on X, all of them people who had linked their own accounts with outside AI providers. “The incident is ongoing so we cannot share more details at this point but we are working with Anthropic, OpenAI, and Google to remediate,” the project’s official account posted.

Key points

  • OpenCode says attackers reached a “legacy system” and exposed provider API keys saved by 1.4% of its Console users.
  • Users report unauthorised charges from about $65 to roughly $3,000; these figures have not been independently verified.
  • Anthropic says it revoked affected Claude keys; a Google staffer says exposed Gemini keys were blocked.
  • OpenCode says refunds are underway. It has not said how many people were hit or who was behind the attack.

What OpenCode has confirmed

OpenCode lets programmers hand coding tasks to AI models from more than 75 providers. Its paid Console product can also store a customer’s own keys, the long passwords that let software use a service such as OpenAI’s and bill the owner for every request. The project’s GitHub page has about 215,000 stars, and the company says 16 million developers use it every month, The Latent reported.

The email sent to affected users on Saturday morning carried the subject line “Action Required: Potential provider key breach”. It said someone had gained “unauthorized access to a legacy system that we were in the process of decommissioning” and that keys “saved in your workspace may have been exposed.” Users were told to revoke them and watch for unexpected usage, according to screenshots shared by security writer Matt Johansen.

Others posted their own losses. Preston Zeller put his at $1,700. Francisco Giancarelli said he saw “over 700 in openai usage before I pulled the plug.” On Reddit, one user said Google charges reached about €200 before his project was suspended for abuse, Notebookcheck reported. None of these figures has been independently verified.

The AI companies moved quickly. “We’ve been working with OpenCode on yesterday’s breach and have revoked Claude API keys we believe were affected,” wrote Lydia Hallie, who works on Anthropic’s Claude Code team. Evan Otero, who works on Google’s AI Studio and Gemini API, said Google had blocked exposed keys and was working with affected customers. OpenAI had not said publicly whether it had revoked keys as of Sunday, according to a running incident tracker kept by the security site ctxsec.

Dax Raad, who builds OpenCode at the company Anomaly, said the break-in hit the hosted Console, not the app itself. Keys kept on a developer’s own machine “never come to us,” he wrote. Late on Saturday he added: “we already contacted everyone, refunds are already underway.”

Why a stolen key is worth so much

The thieves did not appear to want passwords or card numbers. In a distillation attack, someone sends huge numbers of questions to a powerful AI model and uses the answers to train a cheaper copy of it. Every answer costs money, and with stolen keys the bill goes to someone else. OpenCode has not explained how it linked the break-in to distillation, and it has not named a suspect.

Less than a week earlier, Anthropic said in a blog post that new defences had cut the data harvested by distillers on its platform by around 90 to 99%. It described networks of fraudulent accounts, some built with false identities and stolen payment details, and warned that distillers spread traffic across “third-party cloud platforms, and other inference providers.” Analysis: the keys of real, paying customers are harder to spot and ban than fake accounts, which makes any service that stores them in one place a tempting target. HahuPrime reported this week on AI models forging access tokens during OpenAI’s own tests and on spyware delivered through online ads.

“They’ve been non stop trying to attack our systems for the past 48 hours. A side effect is everything is getting hardened but damn we basically need agents going 24/7, because that’s what they have.” Dax Raad, OpenCode

Raad posted that on Sunday. He did not say whether any other systems or data had been reached.

What the OpenCode breach means for developers in Africa

OpenCode has not said where affected users live, and there is no public sign that East African developers were among them. But the risk is the same for anyone who pastes a key into a hosted tool. A freelancer in Addis Ababa or a startup team in Nairobi pays for these models in dollars. With the birr’s sharp slide against the dollar this year, a $400 overnight bill is a serious loss for an Ethiopian developer, even if a refund follows.

The providers’ advice is simple: create a new key, delete the old one and check usage since October 16. Security trackers such as ctxsec also recommend a hard monthly spending cap on every key, so a stolen one can only do limited damage.

OpenCode has not said how the legacy system was reached, how long keys were exposed, how many people were affected or what the refunds will total. It has promised further updates as its investigation with the three AI companies continues.

Sources

85 Views

By HahuPrime News Desk

The HahuPrime News Desk reports on Ethiopia, Eritrea, the Horn of Africa, East Africa, the Red Sea and the Middle East. Our reports separate confirmed facts from claims and analysis, attribute every claim, and list their sources. Corrections: info@hahuprime.com.