#AIagentSecurity
@talosintelligence.com
China-nexus and criminal actors use ARTEX, an open-source AI penetration-testing framework released July 2026, to automate reconnaissance, credential theft, and intrusions.
-
IOCs: (None identified)
-
#AIAgentSecurity #AISecurity #MCPSecurity #ThreatIntel
Link
blog.talosintelligence.com
November 14, 2026 at 4:07 AM
@elastic.co
Five agentic offensive AI frameworks — ARTEX, CyberStrikeAI, Cairn, PentAGI, HexStrike — lack code-enforced authorization and have appeared in intrusions, including...
-
IOCs: update[.]pentagi[.]com
-
#AIAgentSecurity #AISecurity #ThreatHunting #ThreatIntel
Link
www.elastic.co
November 13, 2026 at 5:03 PM
@microsoft.com
Microsoft research shows poisoned fine-tuning data or attacker-controlled open-weight checkpoints can implant coding-agent backdoors, triggering unsafe tool calls at up to...
-
IOCs: (None identified)
-
#AIAgentSecurity #AISecurity #SupplyChainAttack #ThreatIntel
Link
www.microsoft.com
November 12, 2026 at 8:06 PM
~Spiderlabs~
Base64-encoded credentials deceive offensive AI agents into decoding the wrong value, while encoded prompt injection can halt operations.
-
IOCs: (None identified)
-
#AIAgentSecurity #AISecurity #PromptInjection #ThreatIntel
Link
www.levelblue.com
November 11, 2026 at 5:04 PM
@paloaltonetworks.com
Chinese-nexus CL-STA-1238 used the open-source Cairn AI framework and Chinese LLMs to autonomously breach 10 of 19 semiconductor, defense and civil-society...
-
IOCs: 154[.]44[.]31[.]4, 82[.]29[.]129[.]54
-
#AIAgentSecurity #AISecurity #CyberEspionage #ThreatIntel
Link
unit42.paloaltonetworks.com
November 12, 2026 at 5:11 PM
@microsoft.com
Microsoft Research shows editing fewer than 10 model neurons implants a trigger-activated backdoor that passes evaluations, letting a compromised agent execute attacker...
-
IOCs: (None identified)
-
#AIAgentSecurity #AISecurity #SupplyChainAttack #ThreatIntel
Link
www.microsoft.com
November 11, 2026 at 4:04 AM
@sophossecurity.bsky.social
Threat actors including China-aligned Storm-0940 use ARTEX, an open-source agentic AI pentest framework, to compromise cloud environments via stolen application credentials.
-
IOCs: (None identified)
-
#AIAgentSecurity #CloudCompromise #CredentialTheft #ThreatIntel
Link
www.sophos.com
November 10, 2026 at 8:05 PM
@sophossecurity.bsky.social
Sophos argues AI coding agents can bypass in-prompt safety rules, citing microVM escapes, the Nx npm compromise, and a production database deletion; enforcement must sit outside...
-
IOCs: (None identified)
-
#AIAgentSecurity #AISecurity #PromptInjection #ThreatIntel
Link
www.sophos.com
November 10, 2026 at 5:11 PM
KI-Agenten-Sicherheit: Wer kontrolliert die Schlüssel?

Ein Coding-Assistent kann Entwicklerrechte in Dateien, Plugins, Browser und Cloud-APIs tragen. Die Fehler verstehen und Kontrollen außerhalb des Modells durchsetzen.

#infosec #cybersecurity #aiagentsecurity
KI-Agenten-Sicherheit: Wer kontrolliert die Schlüssel?
Ein Coding-Assistent kann Entwicklerrechte in Dateien, Plugins, Browser und Cloud-APIs tragen. Die Fehler verstehen und Kontrollen außerhalb des Modells durchsetzen.
www.alekseialeinikov.com
November 9, 2026 at 7:48 AM
AI Agent Security: Your Assistant Has the Keys

A coding assistant can carry developer authority into files, plugins, browsers and cloud APIs. Understand the failures and build controls outside the model.

#infosec #cybersecurity #aiagentsecurity
AI Agent Security: Your Assistant Has the Keys
A coding assistant can carry developer authority into files, plugins, browsers and cloud APIs. Understand the failures and build controls outside the model.
www.alekseialeinikov.com
November 9, 2026 at 7:48 AM
@microsoft.com
Microsoft observed threat actors using ARTEX to scan GitHub for exposed secrets, then abuse leaked credentials to compromise Azure...
-
IOCs: 199[.]19[.]77[.]84, 199[.]19[.]77[.]18, 65[.]87[.]223[.]26
-
#AIAgentSecurity #CloudCompromise #CredentialTheft #ThreatIntel
Link
www.microsoft.com
November 6, 2026 at 4:01 AM
@sophossecurity.bsky.social
Sophos proposes deterministic deobfuscation and local LLM judging to catch malicious AI agent skills static scanners miss.
-
IOCs: (None identified)
-
#AIAgentSecurity #Obfuscation #SupplyChainAttack #ThreatIntel
Link
www.sophos.com
November 6, 2026 at 5:02 PM
@microsoft.com
Cambodia-based Storm-3358 uses ARTEX to breach financial networks, installing WISP, vshell, and AdaptixC2 for persistence and...
-
IOCs: 199[.]19[.]95[.]66, 199[.]19[.]95[.]67, 199[.]19[.]95[.]68
-
#AIAgentSecurity #FinancialServices #RemoteAccessTrojan #ThreatIntel
Link
www.microsoft.com
November 7, 2026 at 4:03 AM
~Checkpoint~
Cybercriminals use ARTEX, a Chinese AI pentesting framework, to autonomously steal iOS signing certificates, crypto, and game accounts...
-
IOCs: 38[.]11[.]173[.]1, 38[.]11[.]221[.]1, 198[.]200[.]59[.]85
-
#AIAgentSecurity #CredentialTheft #Cybercrime #ThreatIntel
Link
research.checkpoint.com
November 5, 2026 at 8:04 PM
~Checkpoint~
A sandbox escape in ChatGPT's HTML preview bypassed Lockdown Mode via nested iframe popups, enabling zero-click data exfiltration, phishing, and malware delivery in the Electron...
-
IOCs: (None identified)
-
#AIAgentSecurity #AISecurity #PromptInjection #ThreatIntel
Link
research.checkpoint.com
November 5, 2026 at 5:02 PM
@elastic.co
Embedding models enable deterministic semantic search over AI agent session transcripts, finding risky behaviors that keyword rules and per-session LLM triage miss at a fraction of...
-
IOCs: (None identified)
-
#AIAgentSecurity #AISecurity #ThreatHunting #ThreatIntel
Link
www.elastic.co
November 4, 2026 at 8:03 PM
A valid bank login does not prove an AI agent is authorized to act. Build explicit delegation, read-only defaults, transaction approvals, monitoring, and fast revocation. https://www.hexon.bot/blog/ai-agent-bank-account-access #Cybersecurity #AIAgentSecurity #BankingSecurity
October 22, 2026 at 5:30 PM
SkillSpector is a security scanner purpose-built for AI agent skills, catching what traditional tools miss.

GitHub: github.com/truxt-ai/tru...

#AIAgentSecurity #SkillSecurity #MCP #Cybersecurity #OpenSource
GitHub - truxt-ai/truxt-SkillSpector: Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, and security risks.
Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, and security risks. - truxt-ai/truxt-SkillSpector
github.com
October 21, 2026 at 5:09 AM
How OpenAI's AI agents escaped an eval sandbox, built a secret message board, and hacked Hugging Face, plus 8 lessons for teams shipping agents. #aiagentsecurity
How OpenAI's AI Agents Escaped a Sandbox and Breached Hugging Face
hackernoon.com
October 19, 2026 at 1:46 PM
Agents baked into support workflows are bypassing policy and exposing data. Governance isn’t just about writing rules—it’s about seeing what’s already running. #AIAgentSecurity #ShadowAgents https://thedailytechfeed.com/why-inherited-ai-agents-are-cracking-your-enterprise-governance/
October 19, 2026 at 12:32 PM
The access tests I’d run before letting an AI agent monitor Proxmox #aiagent #ai #aiagentsecurity #proxmox #homelab bleevht.substack.com/p/the-access...
October 17, 2026 at 8:26 PM