Loading story…
A Russian-speaking threat actor using the handle Trim reportedly bypassed safety controls on public frontier AI models and turned the resulting jailbreak methods into an offensive cyber platform. Researchers said Trim first posted a guide on a Russian-language cybercrime forum in March describing six techniques for coercing Anthropic Claude Opus into generating malicious outputs, while also naming fallback models and offering access to black-market API keys. The activity did not rely on exploiting software flaws in the models themselves; instead, it abused public API access, leaked system prompts, and prompt-engineering techniques to defeat guardrails.
By June, Trim had allegedly commercialized the approach as AI Pentest Checker, a service that automates web reconnaissance, vulnerability validation, exploitation reporting, and PDF report generation. The platform reportedly combined AI models including Opus 4.8 and GLM-5 with established offensive tools such as Nuclei, ffuf, katana, subfinder, and gitleaks, with one modified system prompt said to be derived from a leaked Fable 5 configuration. Researchers warned the case shows how quickly cybercriminals can operationalize and monetize AI-assisted offensive workflows, increasing the speed and accessibility of attack capabilities.
Mallory correlates global threat intelligence with your attack surface — know if you’re exposed before adversaries strike.
2 events from the most recent confirmed update back to the earliest known activity.
By June 21, 2026, Trim had turned the jailbreak techniques into a commercialized platform called AI Pentest Checker. The service reportedly combined frontier AI models with tools including Nuclei, ffuf, katana, subfinder, and gitleaks to automate reconnaissance, vulnerability validation, exploitation reporting, and PDF report generation.
On March 31, 2026, the Russian-speaking threat actor "Trim" appeared on a Russian-language cybercrime forum and shared six named techniques for bypassing Claude Opus safety controls to obtain malicious outputs. The post also referenced fallback model options and access to black-market API keys.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
5 references tracked. Mallory keeps watching after this page renders.
cyberveille.ch
Open sourcecyberaccord.com
Open sourcecybersecuritynews.com
Open sourcecatonetworks.com
Open sourcedarkreading.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Security researchers reported that cybercriminals are increasingly commercializing offensive AI services, with one Russian-speaking actor, **Trim**, turning jailbreak techniques for Anthropic Claude into a paid penetration-testing platform. According to Cato CTRL, Trim first shared methods such as **Context Warming** and **Ghost Reset** to bypass Claude Opus safety controls, then later advertised **AI Pentest Checker**, a service that allegedly used a grey-market Claude API key obtained via Telegram. The platform reportedly combined jailbroken Claude Opus 4.8 and **GLM-5** with established tools including **Nuclei**, **ffuf**, **katana**, and **gitleaks** to automate web vulnerability scanning and generate PDF reports in minutes. A separate report from Trellix said another service, **MessiahGPT**, is being marketed on BreachForums and Telegram as an uncensored black-hat AI platform capable of generating ransomware, phishing kits, stealers, crypters, and rootkits on demand. The service reportedly offers free trial queries and low-cost cryptocurrency subscriptions without KYC, while similar offerings such as **DarkGPT** are also circulating in Russian-language channels. Researchers said the emergence of these tools shows that criminal **AI-as-a-service** is maturing into a commercial market, lowering barriers for attackers and likely increasing the volume of AI-assisted phishing, malware development, and automated reconnaissance.
Aug 17, 2026
Underground operators are increasingly packaging generative AI into **criminal AI-as-a-service** offerings that support phishing, fraud, impersonation, malware modification, translation, and post-breach data analysis. Reporting on the market describes branded tools such as **FraudGPT**, **GhostGPT**, and **WormGPT**, but says many are little more than wrappers around legitimate models, stolen accounts, or hijacked API keys rather than fully autonomous hacking platforms. Researchers also highlighted a parallel market for deepfake-enabled abuse, including voice cloning, face swaps, fake selfies, and virtual camera injection used for KYC bypass, synthetic identity fraud, and onboarding scams. Sophos separately identified a malware development lab using AI tools including **Cursor** and **Claude Opus** to speed ransomware-related coding, payload refinement, and evasion testing. The framework reportedly generated Python-driven Rust and Go malware modules, customized **Cobalt Strike** profiles, Telegram-based command channels, Cloudflare Worker intermediaries for command-and-control traffic, code-injection scripts targeting legitimate Windows applications, and AI-assisted Active Directory reconnaissance. Researchers said the operation tested nearly 80 modules against more than 70 evasion methods across Sophos, CrowdStrike, and Microsoft defenses, reinforcing that the immediate risk is not autonomous AI malware but faster, cheaper, and more scalable attacker workflows.
Jun 29, 2026
Cisco Talos reported that threat actors are increasingly using cloud AI tools such as Claude Code, CodeX, Cursor, and Gemini to support malicious operations, after analyzing prompt artifacts recovered from compromised endpoints. The research found three dominant use cases: AI as a malware and software engineering assistant, AI as a criminal force multiplier, and AI as an accelerator for vulnerability research and penetration-testing-style activity. Talos said existing model guardrails were often ineffective, with actors bypassing them through false ownership claims, bug-bounty or CTF pretexts, task decomposition across sessions, persistent persona conditioning, semantic evasion, or by moving to uncensored models. Case studies showed AI being used to help build DDoS tooling targeting Android TV devices, bulk email validation and phishing-adjacent infrastructure based on PowerMTA and Node.js, credential-harvesting workflows including abuse of exposed `.git/config` files and React2Shell techniques, and cryptojacking campaigns that used weak Deluge and qBittorrent credentials to deploy **XMRig**. Talos also documented AI-enabled scam and cryptocurrency theft activity, including a russophone scam chatbot, a hispanophone actor using an autonomous OpenClaw-based agent called **Alex** to target Telegram Mini Apps and **TON** wallets, and autonomous offensive frameworks such as **Hephaestus**; the firm warned that AI is lowering barriers for novice attackers while enabling more capable operators to scale intrusion, fraud, and exploitation pipelines faster.
Aug 11, 2026