$ 45.09 € 50.6 zł 11.57
+17° Kyiv +12° Warsaw +11° Washington

Incident at OpenCode Console affected 1.4% of users — TechCrunch

Fedir Kryshtovskyi 20 October 2026 03:24
Incident at OpenCode Console affected 1.4% of users — TechCrunch

Attackers infiltrated an old system of the OpenCode Console online platform, where some users stored API keys for access to AI services. The incident affected 1.4% of Console users, TechCrunch reports, citing a company statement. OpenCode said it had notified all affected users by email.

Unauthorized charges

After the attack, developers reported on Reddit and X receiving spending alerts and seeing unauthorized charges on accounts with OpenAI, Anthropic, and Google AI services. Some users said the charges amounted to hundreds of dollars, while several reported amounts exceeding $1,000.

Anthropic and Google blocked API keys they considered compromised. OpenAI said it worked with OpenCode to restore access and return credits to its API customers that were used during the attack. OpenCode also said reimbursement for unauthorized expenses is ongoing.

Attack on Console

According to OpenCode co-founder Dax Raad, the breach was limited to OpenCode Console, a web platform where users could store API keys in workspaces. The open-source OpenCode application and command-line tool were not affected.

More current news is available on the UA.News Telegram channel Telegram.

The company described the incident as a large-scale model distillation attack, but did not explain its mechanism. Some experts cited by TechCrunch questioned this characterization, noting that the use of compromised keys specifically for distillation has not been proven.

Limitations of defensive AI tools

Raad said that over 48 hours, the attackers continuously attempted to penetrate OpenCode's systems. The team used AI agents for defense, but models involved in analyzing malicious code began refusing to work after security filters were triggered.

OpenCode later gained access to OpenAI's Daybreak Blue program for verified users, whose models are less likely to refuse requests related to defensive cybersecurity. According to Raad, this improved the team's ability to investigate the attack. The company advised affected users to revoke and replace their keys directly with the relevant service providers; the incident is ongoing.

Read us on
Download our app