← All resources

OpenCode: Console breach may have exposed OpenAI, Anthropic, Google API keys

First reported 2026-10-18 · security-incident · data-breach · reported · supply-chain

OpenCode reported a Console breach that may have exposed AI provider API keys belonging to 1.4% of its users, leading to revocation notices. The exposed keys include those for OpenAI, Anthropic, and Google.

Data breach · Reported, not confirmed · First reported 18 Oct 2026

What has been reported

  • Where it happened: OpenAI
  • Exposed through it, as reported: Anthropic, Google and OpenAI
  • Data reported as exposed: keys and secrets
  • Extent as reported: 1.4% of users
  • Products or systems named: Console and internal research mode
  • Reach into customers: a source says the attackers may be able to reach OpenAI's customers, through remote access, customer credentials or tokens, or software updates

How sure is this?

This is what the sources below report. OpenAI has not confirmed it publicly, as far as Resiliate has found.

Sources

  1. OpenCode Console Breach May Have Exposed OpenAI, Anthropic, and Google API Keys (thehackernews.com, press, 18 Oct 2026)
  2. OpenAI AI Agents Bypass Security Isolation, Exploiting 0-Days in Red-Teaming Tests (cybersecuritynews.com, press, 25 Oct 2026)

Are you impacted?

306 companies mapped in Resiliate have OpenAI on their public vendor list. That shows who depends on OpenAI, not whose data was involved: only OpenAI can say that.

If OpenAI is one of your suppliers, ask it what was accessed and whether your data or your access was involved. Map your vendors in Resiliate to see incidents like this against your own supplier list and to record your answer.


Security incidents on Resiliate are in beta. This page is built from public reports and is updated when new sources appear; it was last updated on 16 Nov 2026. Something wrong? Write to hello@resiliate.io.