01
Botpress says an attacker read bot configurations in 16 customer workspaces, including stored API keys
In a disclosure published on 6 November, the chatbot platform Botpress said an attacker exploited a flaw in an internal authentication proxy and had access to internal tools from 13 to 21 October. On 21 October, using an employee's API credential with admin rights, the attacker read bot configurations in 16 customer workspaces and changed one integration's settings in two more. Such configurations can hold AI provider keys, messaging tokens and CRM credentials. Botpress says the apparent aim was to obtain AI usage for resale, that it found no access to end-user conversations or changes to its code, and that it has contacted the affected customers.
What it means for you
If you were contacted, rotate every credential stored in the listed bots, as Botpress advises. For any chatbot or agent platform you use, keep a list of the keys stored there, give them only the permissions they need and set spending limits with your AI provider so a leaked key cannot run up large costs.