Anthropic customers report Claude tokens drained by stolen login sessions

Anthropic has told at least one paying customer that hackers used a stolen session key to create unauthorized access tokens on his account and spend his usage allowance without his knowledge, TechCrunch reported on September 8.

Grant De Swardt, an independent AI consultant in East Sussex, U.K., noticed on August 4 that token usage on his Claude Max 20x account was rising even though he was not working. A token is the unit of text an AI model reads or writes, and paid Claude plans cap how many a user gets. The next day he disconnected everything tied to the account and still did no work. Usage kept climbing. “In the clearest controlled interval, it increased from 45% to 55% while I performed no work, scheduled Cowork tasks were paused or completed, Dispatch/cloud execution was disabled, and there was no corresponding active local Claude Code task,” he told TechCrunch.

De Swardt asked Anthropic for an itemized list of what was consuming his allowance. The company did not provide one but agreed something was wrong. It suspended his account, invalidated his sessions and server-side Claude Code tokens, and refunded £44.49 for the unused time on his $200-per-month plan. The suspension disrupted his business, which sets up AI agents for small and mid-size companies and leans on them for its own admin, coding, and website work.

After investigating, Anthropic told him a compromised Claude session key had been used to mint unauthorized Claude Code OAuth tokens. The account, it said, “appeared to have been used by an unauthorized-looking third-party service to handle activity for other people,” and it could not determine how that service got access. Anthropic said the evidence was consistent either with credentials being taken without his knowledge or with the account being connected to an outside service. Because support tracks total usage but not itemized usage, even on request, such theft could run for months before anyone noticed.

Other users report the same pattern

De Swardt posted his account on Reddit and drew about 80 comments, several describing similar events. One person said their account “was auto-upgraded without my consent, my credit card got charged, and the usage shot from 0% to 100% automatically.” Another watched usage go from zero to 49% in 12 minutes after a couple of prompts and a web search. A separate user filed a GitHub report after their allowance drained daily for three days.

Two affected users posted emails in which Anthropic warned them their tokens were being stolen. “We have recently become aware of a bad actor that is using common infostealer malware to steal Claude login sessions from people’s computers, then using those login sessions to access Claude accounts and consume their usage,” the email read. Infostealers are malware that installs on a computer and collects saved passwords, session data, and login credentials. Anthropic said the malware does not come from using Claude and can be picked up from infected downloads or ads. It did not send De Swardt such an email, and he says he found no sign his computer was compromised.

His account was reinstated after about two weeks; he has since cancelled it and moved to Cursor, which lets users pick among several models. Asked how customers can spot this misuse, Anthropic declined to comment. What to watch is whether Anthropic adds tools that show customers which sessions and tasks are spending their tokens.