[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f-QMCtQJNWMpu-x_TsQ5LA3SUfkYvSSy6jg7k4cRZN6E":3},{"article":4,"iocs":51,"watch_terms":55,"solutions":56,"lesson":144},{"id":5,"title":6,"slug":7,"summary":8,"ai_summary":9,"brief":10,"url":11,"image_url":12,"published_at":13,"ingested_at":14,"relevance_score":15,"entities":16,"category_id":28,"category":29,"article_tags":33},"9f007acd-7b35-46e1-96f1-6a4313c84f6c","AI Guardrails Aren’t Stopping Criminal Hackers","ai-guardrails-aren-t-stopping-criminal-hackers-d16d77","Hackers around the world are all-in on using agentic AI to wage attacks—and manipulating tools like Anthropic’s Claude to help them carry out their crimes.","Threat actors are increasingly leveraging agentic AI tools, including Anthropic's Claude and Deepseek models, to automate and scale cyberattacks. Four distinct threat groups were observed using AI to bypass model safeguards, often by framing malicious activities as legitimate penetration testing. These attacks exploit common security hygiene issues rather than novel vulnerabilities, highlighting the growing challenge of AI-assisted cybercrime.","Hackers are using agentic AI, like Anthropic's Claude, to bypass safeguards and conduct cyberattacks.","https:\u002F\u002Fwww.wired.com\u002Fstory\u002Fai-guardrails-arent-stopping-criminal-hackers\u002F","https:\u002F\u002Fmedia.wired.com\u002Fphotos\u002F6ad7cd65322aff8e01d3ff70\u002Fmaster\u002Fpass\u002FSecurity_Claude%20Code's%20Guardrails%20Aren't%20Stopping%20Real-Life%20Attacks_v2.jpg","2026-10-21T14:00:00+00:00","2026-10-21T16:00:07.882962+00:00",8,[17,20,23,25],{"name":18,"type":19},"Anthropic","vendor",{"name":21,"type":22},"Claude","product",{"name":24,"type":19},"CrowdStrike",{"name":26,"type":27},"Deepseek","technology","839da5c1-3c34-47e2-9499-f7201640e3ac",{"id":28,"icon":30,"name":31,"slug":32},null,"AI Security","ai-security",[34,39,41,46],{"category":35},{"id":36,"icon":30,"name":37,"slug":38},"80544778-fabb-4dcd-aa35-17492e5dcf4f","Vulnerabilities","vulnerabilities",{"category":40},{"id":28,"icon":30,"name":31,"slug":32},{"category":42},{"id":43,"icon":30,"name":44,"slug":45},"89f78b1c-3503-45a1-9fc7-e23d2ce1c6d5","Malware","malware",{"category":47},{"id":48,"icon":30,"name":49,"slug":50},"e7b231c8-5f79-4465-8d38-1ef13aea5a14","Threat Intelligence","threat-intelligence",[52],{"type":45,"value":53,"context":54},"ClawGod","A tool used by a Chinese-speaking threat actor to modify Anthropic's Claude Code system prompt.",[24,18,21],{"controls":57,"vendors":140},[58,86,114],{"slug":59,"name":60,"problem":61,"products":62},"endpoint-detection-response","Endpoint detection and response (EDR and XDR)","Malware and hands-on-keyboard attackers on laptops and servers get spotted and stopped.",[63,70,78],{"vendor_slug":64,"vendor_name":24,"vendor_logo":65,"product_slug":66,"product_name":67,"url":68,"description":69},"crowdstrike","\u002Fvendor-logos\u002Fcrowdstrike.png","falcon-insight-xdr","Falcon Insight XDR","https:\u002F\u002Fwww.crowdstrike.com\u002Fen-us\u002Fplatform\u002Fendpoint-security\u002Ffalcon-insight-xdr\u002F","Endpoint detection and response with extended detection across other security tools.",{"vendor_slug":71,"vendor_name":72,"vendor_logo":73,"product_slug":74,"product_name":75,"url":76,"description":77},"microsoft","Microsoft","\u002Fvendor-logos\u002Fmicrosoft.png","defender-for-endpoint","Microsoft Defender for Endpoint","https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fsecurity\u002Fbusiness\u002Fendpoint-security\u002Fmicrosoft-defender-endpoint","Endpoint detection and response for Windows, macOS, Linux and mobile devices.",{"vendor_slug":79,"vendor_name":80,"vendor_logo":81,"product_slug":82,"product_name":83,"url":84,"description":85},"eset","ESET","\u002Fvendor-logos\u002Feset.png","eset-protect","ESET PROTECT","https:\u002F\u002Fwww.eset.com\u002Fint\u002Fbusiness\u002Fsolutions\u002Fendpoint-protection\u002F","Endpoint protection platform with detection and response add-ons.",{"slug":32,"name":87,"problem":88,"products":89},"AI application security","Prompt injection, data leaks through AI tools and unapproved AI use get caught.",[90,98,106],{"vendor_slug":91,"vendor_name":92,"vendor_logo":93,"product_slug":94,"product_name":95,"url":96,"description":97},"boomi","Boomi","\u002Fvendor-logos\u002Fboomi.png","boomi-ai-gateway","Boomi AI Gateway","https:\u002F\u002Fboomi.com\u002Fplatform\u002Fai-gateway\u002F","Gateway for AI model and MCP tool traffic, built on Lunar.dev technology. It limits which tools each agent can call. Its enterprise version adds an approval step for new MCP servers and logs each call with the user or agent behind it.",{"vendor_slug":99,"vendor_name":100,"vendor_logo":101,"product_slug":102,"product_name":103,"url":104,"description":105},"hiddenlayer","HiddenLayer","\u002Fvendor-logos\u002Fhiddenlayer.png","hiddenlayer-ai-security","HiddenLayer AI Security Platform","https:\u002F\u002Fhiddenlayer.com\u002F","Model scanning, AI red teaming and runtime protection for AI applications.",{"vendor_slug":107,"vendor_name":108,"vendor_logo":109,"product_slug":110,"product_name":111,"url":112,"description":113},"cisco","Cisco","\u002Fvendor-logos\u002Fcisco.png","ai-defense","Cisco AI Defense","https:\u002F\u002Fwww.cisco.com\u002Fsite\u002Fus\u002Fen\u002Fproducts\u002Fsecurity\u002Fai-defense\u002Findex.html","Discovery of AI use, testing of models and guardrails for AI applications.",{"slug":115,"name":116,"problem":117,"products":118},"security-awareness-training","Security awareness training and phishing simulation","People recognise and report attacks instead of falling for them.",[119,127,134],{"vendor_slug":120,"vendor_name":121,"vendor_logo":122,"product_slug":123,"product_name":124,"url":125,"description":126},"huntress","Huntress","\u002Fvendor-logos\u002Fhuntress.png","managed-sat","Huntress Managed Security Awareness Training","https:\u002F\u002Fwww.huntress.com\u002Fplatform\u002Fsecurity-awareness-training","Security awareness training and phishing simulation, managed for you.",{"vendor_slug":128,"vendor_name":129,"vendor_logo":130,"product_slug":115,"product_name":131,"url":132,"description":133},"knowbe4","KnowBe4","\u002Fvendor-logos\u002Fknowbe4.png","KnowBe4 Security Awareness Training","https:\u002F\u002Fwww.knowbe4.com\u002Fproducts\u002Fsecurity-awareness-training","Training library and simulated phishing with reporting.",{"vendor_slug":135,"vendor_name":136,"vendor_logo":137,"product_slug":135,"product_name":136,"url":138,"description":139},"hoxhunt","Hoxhunt","\u002Fvendor-logos\u002Fhoxhunt.png","https:\u002F\u002Fhoxhunt.com\u002F","Adaptive phishing simulations and micro-training, with a focus on reporting.",[141],{"slug":64,"name":24,"logo_path":65,"roles":142},[143],"reported_by",{"slug":145,"title":146,"body":147,"framework_refs":148,"framework_notes":156},"ai-guardrails-fail-to-stop-criminal-exploitation-of-agentic-ai-tools","AI Guardrails Fail to Stop Criminal Exploitation of Agentic AI Tools","Threat actors are actively weaponizing AI models like Claude and DeepSeek by framing malicious requests as legitimate penetration testing, effectively bypassing built-in safety guardrails. This exploitation does not rely on novel vulnerabilities but instead targets fundamental security hygiene weaknesses, meaning organizations are being compromised by well-understood problems at unprecedented AI-driven scale. The incident reveals that AI providers' safety mechanisms are insufficient as a standalone defense layer. As agentic AI becomes more accessible, the attack surface for automated, scalable cybercrime grows dramatically, demanding that defenders evolve their controls to account for AI-assisted adversaries.",[149,150,151,152,153,154,155],"CIS Control 14: Security Awareness and Skills Training","CIS Control 17: Incident Response Management","NIST CSF 2.0 PR.AA: Identity Management, Authentication, and Access Control","NIST SP 800-53 SI-2: Flaw Remediation","ISO\u002FIEC 27001:2022 Clause 7.2: Competence","MITRE ATT&CK T1566: Phishing","NIS2 Article 21(2)(b)",{"NIS2 Article 21(2)(b)":157,"MITRE ATT&CK T1566: Phishing":158,"NIST SP 800-53 SI-2: Flaw Remediation":159,"ISO\u002FIEC 27001:2022 Clause 7.2: Competence":160,"CIS Control 17: Incident Response Management":161,"CIS Control 14: Security Awareness and Skills Training":162,"NIST CSF 2.0 PR.AA: Identity Management, Authentication, and Access Control":163},"NIS2 requires incident handling capabilities commensurate with evolving threats; the emergence of agentic AI-assisted cyberattacks constitutes exactly the kind of evolving threat landscape that organizations subject to NIS2 must actively address in their risk management frameworks.","AI tools are being used to scale and personalize phishing and social engineering at rates impossible for human attackers alone, making this a directly applicable technique that defenders must tune detections against in the context of AI-augmented adversaries.","AI-assisted attackers are targeting known, unpatched vulnerabilities at scale; a disciplined flaw remediation process would close the security hygiene gaps that make these automated attacks so effective.","Security teams must develop competency in understanding AI-assisted threat vectors, as the novel framing of attacks as 'penetration testing' to bypass AI guardrails requires informed human oversight that only trained personnel can provide.","Because AI-assisted attacks operate at machine speed and scale, organizations need a mature, practiced incident response capability that can detect and contain automated attack chains before they cause widespread damage.","Threat actors succeeded partly because defenders and AI models alike lacked sufficient awareness of how malicious prompts disguised as pen-testing can bypass guardrails; structured training would help security teams recognize and respond to AI-assisted social engineering tactics.","The article notes that AI-driven attacks exploit common security hygiene failures including weak access controls, and enforcing least-privilege access and strong authentication would directly limit the blast radius of these automated attacks."]