Cybersecurity experts have identified six undisclosed services promoting unauthorized access to artificial intelligence models across dark web forums and chat platforms This article explores cloudflare added phishing. . The service claims it adds targeted accounts to its pool, routing requests through a specific account without revealing the identity, while charging 5-15% of the official per-token price based on model type.
Poison Claude accepts cryptocurrency payments. An identity security firm reported that an oversight in configuration revealed the "api.claudeopus.shop/api/status" endpoint for the API, displaying a count of 881 total users and 872 active users. Following the responsible disclosure process, Cloudflare has added a phishing alert to the site but appears to have declined to take any action regarding the API domain that utilizes Cloudflare Turnstile for bot protection.
A similar service operating in gray market is Ecomagent.in, which is estimated to have nearly 970 users and claims to offer discounted access to Anthropic's Opus 4.8, Opus 4.6, Sonnet 4.6, and OpenAI's GPT Codex 5.5 via a custom API endpoint. Furthermore, evidence suggests that bad actors are abusing free trials offered by AI services to facilitate synthetic identity creation at scale through disposable domains such as dakaka[. ]org, emailinbo[.
]live, and ratixq[.]com. Residential proxies enable malicious traffic to originate from seemingly innocent consumer IP addresses that may lack a history of malicious behavior, making it risky to block them.
