Poison Claude Sells Discounted Claude Access While Its Operator Sees Every Customer Prompt
Illegal AI model access services like Poison Claude are selling discounted access while exposing customer prompts.
Cybersecurity researchers have identified several underground services offering illicit access to AI models, including Anthropic's Claude. One such service, Poison Claude, leverages free bonus credits and charges a fraction of the official price, accepting cryptocurrency payments. A configuration error exposed an API status endpoint, revealing over 800 users, though this has since been fixed. The service provider has full visibility into customer prompts, posing a significant privacy risk.
AI summary of the linked source. Not reviewed by an editor. How this works
Run Anthropic?
Get an email when a new story names Anthropic, usually within the hour.
Free. Your list stays private and never appears in a subject line. One click stops it. How Vendor Watch worksPrivacy
The lesson from this incident
Full lessonPoison Claude and similar underground services exploit stolen or misused API credentials to resell AI model access at steep discounts, attracting users who unknowingly surrender full visibility of their prompts to a malicious third party. A configuration error left an API status endpoint publicly accessible, leaking data on over 800 users before it was patched. This matters because users interacting with unofficial AI proxies have no assurance of data confidentiality — every query, including sensitive business or personal information, is exposed to the service operator. Beyond privacy risks, using such services may violate terms of service and expose organizations to legal and compliance liability. It underscores the danger of prioritizing cost savings over verified, trustworthy service providers.
What helps against this
How we list vendors
- Identity threat detection and response (ITDR)
Attackers who log in with real, stolen credentials get noticed.
- Threat intelligence
Attack campaigns reach your sector, suppliers or software before your team hears about them.
- AI application security
Prompt injection, data leaks through AI tools and unapproved AI use get caught.
Indicators of Compromise
Extracted by AI from the linked source. Not reviewed by an editor. How this works
- domain — api.claudeopus[.]shop
- domain — poison-claude.bitsender[.]top
- domain — Ecomagent.in
- domain — dakaka[.]org
- domain — emailinbo[.]live
- domain — ratixq[.]com