The short answer
- OpenCode said on October 17, 2026 that a security breach impacted 1.4% of OpenCode Console users.
- OpenCode said it emailed all affected users and that users who did not get an email were not impacted.
- According to OfficeChai, OpenCode told users that third-party provider keys, such as OpenAI and Anthropic keys, saved in their workspace may have been exposed.
- OpenCode asked affected users to revoke and replace their keys through each provider's dashboard and review recent usage, OfficeChai reported.
- OpenCode is working with Anthropic, OpenAI and Google on remediation, according to Startup Fortune and OfficeChai.
OpenCode, the open-source coding agent, disclosed a security breach on October 17, 2026 that it said impacted 1.4% of OpenCode Console users. In a post on its verified X account, OpenCode said affected users had been notified by email and that only "a subset of users who had external providers" were affected. Provider API keys saved in those users' workspaces may have been exposed, according to the company's notice as reported by OfficeChai.
#What happened?
OpenCode said the breach was "related to a large scale distillation attack" and affected 1.4% of OpenCode Console users. Its post, first published at 17:56 UTC on October 17, 2026 and edited at 18:08 UTC (23:38 IST), reads: "We had a security breach last night related to a large scale distillation attack that impacted 1.4% of OpenCode Console users."
According to OfficeChai, OpenCode said it had "identified unauthorized access to a legacy system" that was in the process of being decommissioned, and that "third party provider keys (eg OpenAI, Anthropic) saved in your workspace may have been exposed." OfficeChai describes OpenCode Console as "the hosted dashboard where users manage accounts, subscriptions and, crucially, these linked provider credentials".
OpenCode has not said what it means by "distillation attack" in this case, how the attackers got access, or how many users the 1.4% figure represents.
Startup Fortune reported that the company told affected users it could not confirm whether the exposed keys had been used for anything malicious, and that its investigation "to date" had found no evidence other data had leaked.
#Who is affected?
OpenCode said only users who received its notification email were affected. "We have notified them all this morning via email. If you did not get an email you were not impacted, it was a subset of users who had external providers," the company wrote in its October 17 post.
The affected group is limited to OpenCode Console users who had connected external model providers, based on OpenCode's post.
#What should affected users do?
OpenCode asked affected users to replace their keys at the provider, OfficeChai reported: "Affected users have been asked to revoke and replace their keys through the relevant provider’s dashboard, update any applications using them, and review recent usage for anything unexpected."
Startup Fortune described the company's advice in more detail:
- Revoke and reissue every provider key through the OpenAI, Anthropic and Google dashboards directly, rather than clicking links in the breach email.
- Check billing and usage logs going back to October 17, 2026.
- Set spending limits so a stolen key cannot run up an open-ended bill.
Replacing a key also means updating every place that uses it, such as scripts, CI jobs and teammates' local setups, as Startup Fortune noted. Our guide to controlling what AI coding agents can reach on a Mac covers related local controls for coding agents.
#Who is OpenCode working with?
OpenCode is working with Anthropic, OpenAI and Google to remediate the incident, according to both Startup Fortune and OfficeChai. An earlier version of OpenCode's X post said: "The incident is ongoing so we cannot share more details at this point but we are working with". That text was cut from the edited post.
Startup Fortune identifies the company behind OpenCode as Anomaly Innovations, "the company formerly known as SST". OpenCode co-founder Dax Raad posted on X that "my personal accounts are being attacked now too whole new world with ai powered hacking," according to OfficeChai.
#Is this linked to the earlier OpenCode vulnerability?
OfficeChai reported that the breach "appears to be unrelated" to an earlier flaw. That flaw, a critical remote-code-execution bug tracked as CVE-2026-22812, was patched by OpenCode in version 1.1.10, according to OfficeChai.
More coverage of agent security sits in our guide to Dots privacy, security and admin controls, the Privacy & Security hub and the Coding Agents hub.
#What is still unclear?
OpenCode has not published a blog post or technical report on the incident. Open questions:
- How access happened. OpenCode has not explained the attack or the "legacy system" beyond the notice OfficeChai quoted.
- Scale. The company gave a percentage, 1.4%, not a number of users.
- Misuse of keys. Startup Fortune reported that OpenCode could not confirm whether exposed keys were used maliciously.
- Timeline. OfficeChai dated the emails to October 18, but OpenCode's own post on October 17 said the emails went out "this morning". This article follows OpenCode's post.
OfficeChai reported that OpenCode said it will share further updates as its investigation continues. DotsBot will update this article when it does.
Frequently asked questions
Was I affected by the OpenCode breach?
OpenCode said it emailed every affected user and that anyone who did not get an email was not impacted. It said the affected group was a subset of OpenCode Console users who had external providers.
What data was exposed in the OpenCode breach?
According to OfficeChai, OpenCode told users that third-party provider keys, such as OpenAI and Anthropic keys, saved in their workspace may have been exposed. Startup Fortune reported that the company's investigation to date found no evidence other data leaked.
What should affected OpenCode users do?
OfficeChai reported that OpenCode asked affected users to revoke and replace their keys through the relevant provider's dashboard, update any applications that use them, and review recent usage for anything unexpected.
Is this related to the earlier OpenCode CVE-2026-22812 flaw?
OfficeChai reported that the breach appears to be unrelated to CVE-2026-22812, a remote-code-execution flaw that OpenCode patched in version 1.1.10.
Sources
Every factual claim in this article traces back to one of these.
One email a week on AI agents and bots.
Launches, pricing and rollout changes and the guides worth reading. At most once a week, and you can leave with one click.


