Claim: Anthropic hack allegedly gave attacker full Slack admin access
Former DOGE staffers claim they hacked Anthropic's internal systems within 12 hours; Anthropic has not confirmed the breach in the sources. An X post from 2026-10-28 alleges, without a verifiable source, that the intruder gained full administrator access to the Slack workspace, able to read every message and log in as arbitrary users. No fix applies to third parties; watch for an official Anthropic statement.
What happened
Former DOGE staffers claim they hacked Anthropic's internal systems within 12 hours. A post on X alleges, without a verifiable source, that the intruder gained full administrator access to the company's Slack, able to read messages and log in as other users. Anthropic has confirmed nothing in the sources.
Why it matters
If confirmed, Slack administrator access would expose internal conversations and, per the post, the company's research secrets. Businesses that share sensitive data with Anthropic may want to review that exposure. For now, everything is an unverified claim.
What to do
Wait for an official Anthropic statement before acting. Review what sensitive data your company shares with AI vendors and where that data is stored. No fix applies to third parties in this case.
For the technical team
- Affected
- Anthropic Sistemas internos da Anthropic; workspace do Slack da empresa, segundo nova alegação · Anthropic's internal systems and the company's Slack workspace, per the allegations. No specific version or product cited.
- Exploitation
- Breach alleged by former DOGE staffers, amplified by ZeroHedge and Kalshi on X (2026-10-26). @willdepue (2026-10-28) says he is hearing that 'cathedral' achieved full administrator access to the Slack, able to see every message and log in as arbitrary users, and claims Anthropic is underplaying the incident. No Anthropic confirmation, no PoC, and no public evidence in the sources.
- Fix
- No fix cited; alleged incident inside Anthropic's internal systems, without official confirmation.
- Vector
- Alleged: breach of Anthropic's internal systems, with no confirmed vector. @willdepue's post compares the case to the 2023 OpenAI incident, in which an intruder obtained an access token from an external app service, and cites admin API tokens spread across the company. This is the author's analogy, not a verified vector.
- More details
- Timeline: 2026-10-26, ZeroHedge and Kalshi amplify the claim of a 12-hour breach, without Anthropic confirmation; 2026-10-28, @willdepue alleges full Slack administrator access, where every research secret would sit, and compares it to the 2023 OpenAI incident. @TiffMoodNukes speculates whether they are hackers or paid government operatives. No independent verification in the sources.
Timeline
X post claims, without a verifiable source, that the breach yielded full administrator access to Anthropic's Slack, where research secrets would be kept.
New post echoes the claim by former DOGE staffers that they hacked Anthropic's internal systems within 12 hours.
First item received on the case, with no Anthropic confirmation yet.
Sources 4 sources
- @zerohedge @zerohedge
- @Kalshi @Kalshi
- @willdepue @willdepue
- @TiffMoodNukes @TiffMoodNukes
