AI Agents Adapt Hacking Tools in Just 55 Seconds During ConnectWise Attack

Attackers abusing compromised ConnectWise application credentials changed their hacking tools just 55 seconds after a failure, according to Huntress.

The rapid adjustment suggests possible AI assistance, although the researchers could not independently confirm that an AI agent controlled the activity.

Huntress first disclosed the credential abuse on October 16, 2026, after detecting unauthorized access to Microsoft 365 environments.

Attackers used trusted ConnectWise applications to authenticate, raise privileges, and plant credentials in other applications for continued access.

Microsoft attributed the campaign to Storm-0940, a China-aligned threat actor active since 2021. Its October 21 report said the group used ARTEX, an open-source framework that lets multiple AI agents automate offensive security tasks. Microsoft recorded attempted sign-ins against more than 2,250 tenants, not 2,250 confirmed breaches.

AI Adapts Hacking Tools

Huntress examined Entra sign-in and directory audit logs covering activity mainly observed on October 15. The records showed large scripted operations alongside smaller sequences that changed after errors.

The clearest example followed a failed certificate upload at 03:48:24 UTC. Microsoft rejected the certificate because its thumbprint did not match. That run ended without retrying.

Just 55 seconds later, another run included retry behavior, waiting approximately 3.5 seconds and then 7.1 seconds between attempts until the operation succeeded. The same retry behavior appeared when the tooling moved to another tenant.

The change suggests an operator adjusted the tool between runs. However, a person editing code, a person using chatbot assistance, or a prepared tool with an unused retry option could also explain it.

Activity on October 15 in Huntress-monitored tenants, grouped by behavior. Operator-driven means a person, or a model acting for one, was steering the tooling between runs (Source: huntress)
Activity on October 15 in Huntress-monitored tenants, grouped by behavior. Operator-driven means a person, or a model acting for one, was steering the tooling between runs (Source: huntress)

Researchers also observed actions tested against one tenant before being repeated in batches. This pilot-then-batch pattern appeared three times across two tools, supporting the possibility of an operator learning from results.

Other activity clearly resembled conventional automation. Around 100 failed token requests hit 100 tenants within 10 seconds. Another run reached 13 to 21 attempts per second, repeatedly using credentials that could not authenticate outside their original tenant.

Huntress cautioned that speed does not prove autonomous AI involvement. After a targeted Teams service principal was deleted, requests continued every four to eight seconds for another 22 minutes without adjustment.

That behavior could reflect unattended scripts or background tasks within a multi-agent framework. Huntress assessed LLM involvement with low-to-medium confidence from its own telemetry, rising to medium-to-high when considering Microsoft’s attribution.

Visibility remained limited. Investigators lacked the attacker’s tools, model-provider traffic, and Microsoft Graph activity logs. Consequently, they could not establish the full scope of Graph reads or confirm data theft.

Indicators of Compromise (IOCs)

TypeIndicatorDescription / Attribution
IPv4 address144.172.64[.]128Application sign-ins and Microsoft Graph activity; reported in the ConnectWise advisory and observed by Huntress.
IPv4 address144.172.64[.]99Cross-tenant password spraying on October 10–14; reported by ConnectWise and Microsoft and corroborated by Huntress.

Note: IP addresses and domains are intentionally defanged (e.g., [.]) to prevent accidental resolution or hyperlinking. Re-fang only within controlled threat intelligence platforms such as MISP, VirusTotal, or your SIEM.

Join 16,000+ SOC teams using ANY.RUN to streamline threat investigations and reduce manual effort. Explore for your team 

Varshini
Varshini
Varshini is a Cyber Security expert in Threat Analysis, Vulnerability Assessment, and Research. Passionate about staying ahead of emerging Threats and Technologies..

Trending News

Related Stories