[{"data":1,"prerenderedAt":278},["ShallowReactive",2],{"article-\u002Fblog\u002Fbyodt-available":3,"related-\u002Fblog\u002Fbyodt-available":168},{"id":4,"title":5,"author":6,"body":8,"date":154,"description":155,"extension":156,"image":157,"meta":158,"navigation":159,"path":160,"seo":161,"stem":162,"tags":163,"video":166,"__hash__":167},"blog\u002Fblog\u002Fbyodt-available.md","BYODt is available",{"name":7},"dether.net",{"type":9,"value":10,"toc":147},"minimark",[11,15,20,87,90,93,97,100,103,106,110,126],[12,13,14],"p",{},"You can now subscribe to BYODt (Bring Your Own Dethernety). You run the platform on your own infrastructure, and we serve specialized module packages to it from our cloud.",[16,17,19],"h2",{"id":18},"editions-and-prices","Editions and prices",[21,22,23,42],"table",{},[24,25,26],"thead",{},[27,28,29,33,36,39],"tr",{},[30,31,32],"th",{},"Edition",[30,34,35],{},"Seats",[30,37,38],{},"Includes",[30,40,41],{},"Price per month",[43,44,45,60,74],"tbody",{},[27,46,47,51,54,57],{},[48,49,50],"td",{},"BYODt",[48,52,53],{},"1",[48,55,56],{},"The packages you choose",[48,58,59],{},"EUR 25 per package",[27,61,62,65,68,71],{},[48,63,64],{},"BYODt Team 5",[48,66,67],{},"5",[48,69,70],{},"Every technology package",[48,72,73],{},"EUR 199",[27,75,76,79,82,84],{},[48,77,78],{},"BYODt Team 20",[48,80,81],{},"20",[48,83,70],{},[48,85,86],{},"EUR 499",[12,88,89],{},"Five packages are available: AWS, Azure, Containers & Kubernetes, Self-Managed Infrastructure, and Agentic AI. Every package also includes Risk Assessment, Dethereal Triage and the Threat Patterns engine, and Agentic AI adds the patterns it runs. Team editions include packages published later at no extra cost.",[12,91,92],{},"Prices exclude VAT. Billing is monthly, and you can cancel at any time. We sell to businesses only, and we check a team's tax number before its first purchase where a register can be asked.",[16,94,96],{"id":95},"what-leaves-your-network","What leaves your network",[12,98,99],{},"Your models and findings stay in your platform. When you ask for an evaluation, your deployment sends the configured attributes of a single element, meaning the answers to that class's form. Most answers are yes\u002Fno or a fixed choice, but some are free-form values, such as a role ARN or a policy document. We run the policy over those values in memory, return the result, and keep neither the input nor the output.",[12,101,102],{},"We do not record which class you asked about, and no request carries a deployment identifier. We do know who you are and when your deployment asked for something, because users sign in against our identity service, and we meter request volume per team. Accounts and content are hosted in the EU.",[12,104,105],{},"AI clients are separate. Dethereal and Triage run in your coding agent, which sends what they read to the model provider you configure.",[16,107,109],{"id":108},"how-to-start","How to start",[12,111,112,113,120,121,125],{},"The platform stays free and open source, so you can ",[114,115,119],"a",{"href":116,"rel":117},"https:\u002F\u002Fdocs.dether.net\u002Fplatform\u002Fquick-start\u002F",[118],"nofollow","run it first",". There is no self-serve trial. If you would like an evaluation, ",[114,122,124],{"href":123},"\u002Fcontact","contact us"," and we can arrange access.",[12,127,128,129,134,135,140,141,146],{},"To subscribe, ",[114,130,133],{"href":131,"rel":132},"https:\u002F\u002Fbyodt.dethernety.io\u002Fofferings",[118],"see the plans"," or read the ",[114,136,139],{"href":137,"rel":138},"https:\u002F\u002Fdocs.dether.net\u002Fbyodt-subscription\u002F",[118],"subscription guide",". Our ",[114,142,145],{"href":143,"rel":144},"https:\u002F\u002Fbyodt.dethernety.io\u002Fdpa",[118],"data processing terms"," are published on the portal.",{"title":148,"searchDepth":149,"depth":149,"links":150},"",2,[151,152,153],{"id":18,"depth":149,"text":19},{"id":95,"depth":149,"text":96},{"id":108,"depth":149,"text":109},"2026-11-04","BYODt (Bring Your Own Dethernety): host the platform yourself and get module packages served from our cloud. From EUR 25 per package per month.","md","\u002Fimages\u002Fhero-graph.jpeg",{},true,"\u002Fblog\u002Fbyodt-available",{"title":5,"description":155},"blog\u002Fbyodt-available",[164,165],"announcement","editions",null,"0pl5wM45wf4HqdM7miwYeVFnxrOuymxcuYyH2n18uhE",[169,208],{"id":170,"title":171,"author":172,"body":173,"date":196,"description":197,"extension":156,"image":198,"meta":199,"navigation":159,"path":200,"seo":201,"stem":202,"tags":203,"video":206,"__hash__":207},"blog\u002Fblog\u002Fthreat-modeling-agentic-architectures.md","Threat modeling agentic architectures: two new videos",{"name":7},{"type":9,"value":174,"toc":194},[175,178,186],[12,176,177],{},"Two new videos cover threat modeling for AI agents. Both use a fictional support copilot whose tickets arrive from the internet.",[12,179,180,185],{},[114,181,184],{"href":182,"rel":183},"https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=pzeDZ8YaLAU",[118],"Threat modeling of agentic architectures"," (6:56) is for people working on AI agents, copilots and their tools. It starts with threats inside one component, such as an agent acting without human approval, then moves to threats across components. It draws on two cases from MITRE ATLAS: a package posing as an e-mail service's MCP server, which began copying every e-mail it sent to the attacker (AML.CS0053), and the PromptArmor demonstration against Slack AI (AML.CS0035). The second half models the copilot in Dethernety and goes through its findings, open questions, fixes and the summary page for the CISO.",[12,187,188,193],{},[114,189,192],{"href":190,"rel":191},"https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=BF7OMnhj6qg",[118],"How a support AI can leak customer data with zero clicks"," (1:34) is the short version, with captions. It follows one attack on the copilot. A ticket hides an instruction to add customer details to an image link in the reply. When a staff member opens the draft, the console loads the image by itself, which sends the data to the attacker's server. Nobody has to click the link.",{"title":148,"searchDepth":149,"depth":149,"links":195},[],"2026-10-16","A seven-minute video and a 90-second short on threat modeling AI agents, using a fictional support copilot.","\u002Fimages\u002Fblog\u002Fpost-agentic.jpg",{},"\u002Fblog\u002Fthreat-modeling-agentic-architectures",{"title":171,"description":197},"blog\u002Fthreat-modeling-agentic-architectures",[204,205],"video","agentic AI","pzeDZ8YaLAU","szjbpIFycX9vuzKXN9ZWLnw8vR0RJAsbsbHhG1RjDa8",{"id":209,"title":210,"author":211,"body":212,"date":268,"description":269,"extension":156,"image":270,"meta":271,"navigation":159,"path":272,"seo":273,"stem":274,"tags":275,"video":166,"__hash__":277},"blog\u002Fblog\u002Fagentic-modules-threat-patterns.md","Modeling AI agents: new classes and Threat Patterns",{"name":7},{"type":9,"value":213,"toc":264},[214,217,221,224,227,231,234,237,250,253,256],[12,215,216],{},"Dethernety can now model AI agents and the systems around them. We have added a vendor-neutral class library and a new analysis called Threat Patterns.",[16,218,220],{"id":219},"classes-for-agentic-systems","Classes for agentic systems",[12,222,223],{},"Five modules add 50 classes. They cover agents and their runtimes, MCP servers and gateways, hosted and self-hosted models, vector stores and memory, self-hosted platforms such as n8n and Dify, and controls such as human approval and tool authorization. Separate AWS and Azure modules cover services including Bedrock AgentCore, Foundry agents, Copilot Studio and Microsoft 365 Copilot.",[12,225,226],{},"MITRE ATT&CK has no technique for prompt injection or tool poisoning. MITRE ATLAS does, so ATLAS joined the platform's reference data in release 0.10.0, and the new classes cite it.",[16,228,230],{"id":229},"threat-patterns","Threat Patterns",[12,232,233],{},"Some threats appear only when several elements line up. Take an agent that reads content anyone can write, can reach private data, and can send data where an outsider can read it. Simon Willison has called this combination the lethal trifecta.",[12,235,236],{},"Threat Patterns reads the whole model and looks for combinations like this one. Others include:",[238,239,240,244,247],"ul",{},[241,242,243],"li",{},"Credential chain: a foothold in the agent stack reaches a credential that unlocks sensitive data or actions.",[241,245,246],{},"Injection to action: untrusted content can trigger a write or code execution that nobody approves.",[241,248,249],{},"Delegation hop: an agent exposed to untrusted content hands the work to an agent that holds the other conditions.",[12,251,252],{},"The analysis runs without AI. A finding has the pattern's fixed severity band but no risk score. It is confirmed in the model when every required condition is present. If the model does not say enough to decide, the finding reads Not ruled out, and the analysis names the fact it needs.",[12,254,255],{},"It reads the model, not the running system, so a confirmed finding describes the design and is not proof of an exploit.",[12,257,258,259,263],{},"The classes and the patterns will be sold as an Agentic AI package in BYODt, our subscription edition, which is in development. ",[114,260,262],{"href":261},"\u002Fthreat-patterns","Read more about Threat Patterns",".",{"title":148,"searchDepth":149,"depth":149,"links":265},[266,267],{"id":219,"depth":149,"text":220},{"id":229,"depth":149,"text":230},"2026-10-12","New modules for AI agents and their tools, and an analysis that finds threats spanning several elements of a model.","\u002Fimages\u002Ffeatures\u002Fthreat-patterns-diagram.jpg",{},"\u002Fblog\u002Fagentic-modules-threat-patterns",{"title":210,"description":269},"blog\u002Fagentic-modules-threat-patterns",[164,276,205],"modules","DbbJ4alNUWiY4O7rQ4NYwzRWhUNFdGqX099gsqRz5RE",1793814258327]