2026/10/26
Imagine buying the most expensive lock for your front door.
Then someone discovers that the door used by your cleaning service has a much weaker lock.
That's a useful way to think about the latest cybersecurity news involving Anthropic, the company behind Claude.
On October 26, The Washington Post reported that researchers at a cybersecurity startup called Cathedral had found ways into some of Anthropic's internal company systems.
They used AI agents to help discover weaknesses.
But they didn't need to break Claude itself.
The vulnerabilities were in third-party software used for everyday business tasks.
According to a draft research report, those weaknesses opened paths to sensitive workplace information, including salary and job-offer details and meeting records.
Anthropic confirmed access to a limited set of human resources, sales and IT systems.
To be clear, the researchers were investigating vulnerabilities under the company's security-testing program. This wasn't evidence of criminals stealing everyone's Claude conversations.
Anthropic says customer conversations, model code and model weights were not accessed. It blocked the affected paths and says it found no evidence of exploitation by anyone else.
What makes this story important is how ordinary the entry point was.
Almost every company uses outside software for hiring, payments, customer management and internal communication.
And the more systems are connected, the more important those connections become.
Here's the question:
**Could someone reach sensitive information in your company through a tool nobody thinks of as security-critical?**
Original report:
https://www.washingtonpost.com/technology/2026/10/26/researchers-cybersecurity-start-up-cathedral-got-inside-anthropic/
More thoughts on technology, privacy and work: inclinq.com