--- title: Doors of (AI)pportunity: The Front and Backdoors of LLMs... description: Security researchers from Hidden Layer present a comprehensive breakdown of LLM attack surfaces, covering both 'front doors' (prompt injection techniques) and... canonical: https://daily.dev/posts/doors-of-ai-pportunity-the-front-and-backdoors-of-llms---kasimir-schulz-kenneth-yeung-524z4aok1 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Doors of (AI)pportunity: The Front and Backdoors of LLMs - Kasimir Schulz & Kenneth Yeung | daily.dev og:description: Security researchers from Hidden Layer present a comprehensive breakdown of LLM attack surfaces, covering both 'front doors' (prompt injection techniques) and... og:url: https://daily.dev/posts/doors-of-ai-pportunity-the-front-and-backdoors-of-llms---kasimir-schulz-kenneth-yeung-524z4aok1 og:image: https://api.daily.dev/og/posts/524Z4AOK1.png og:image:alt: Doors of (AI)pportunity: The Front and Backdoors of LLMs - Kasimir Schulz & Kenneth Yeung og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Doors of (AI)pportunity: The Front and Backdoors of LLMs - Kasimir Schulz & Kenneth Yeung **[NDC Conferences](https://daily.dev/sources/ndc)** · 53 min read · 0 upvotes · 0 comments ## Summary Security researchers from Hidden Layer present a comprehensive breakdown of LLM attack surfaces, covering both 'front doors' (prompt injection techniques) and 'backdoors' (model tampering). The talk demonstrates prompt injection via direct and indirect methods, introduces CROP (Knowledge Return Oriented Prompting) for bypassing content filters using cultural references, and explains Policy Puppetry — a universal alignment bypass that works across all major LLMs including ChatGPT, Claude, and Gemini. The second half covers supply chain risks when loading third-party ML models from HuggingFace, including arbitrary code execution via Python pickle, trust_remote_code exploitation, and repository sideloading. A live demo shows how to backdoor an ONNX model's computational graph to insert a trigger-based payload (e.g., misclassifying a bear as a dog when red pixels are detected) that evades antivirus and is invisible to standard code review. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Similar posts on daily.dev - [Backdoors in LLMs: Why model scanning isn't enough](https://daily.dev/posts/backdoors-in-llms-why-model-scanning-isn-t-enough-bbk64raco) · Red Hat Developer · 0 upvotes · 0 comments - [The security paradox of local LLMs](https://daily.dev/posts/the-security-paradox-of-local-llms-v2cxahmuu) · Hacker News · 1 upvotes · 0 comments - [How Your AI Chatbot Can Become a Backdoor](https://daily.dev/posts/how-your-ai-chatbot-can-become-a-backdoor-e8icglhaz) · Trend Micro · 0 upvotes · 0 comments --- Tags: [#security](https://daily.dev/tags/security), [#machine-learning](https://daily.dev/tags/machine-learning), [#prompt-injection](https://daily.dev/tags/prompt-injection) [View this post on daily.dev](https://daily.dev/posts/doors-of-ai-pportunity-the-front-and-backdoors-of-llms---kasimir-schulz-kenneth-yeung-524z4aok1) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Doors of (AI)pportunity: The Front and Backdoors of LLMs - Kasimir Schulz & Kenneth Yeung","url":"https://daily.dev/posts/doors-of-ai-pportunity-the-front-and-backdoors-of-llms---kasimir-schulz-kenneth-yeung-524z4aok1","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/doors-of-ai-pportunity-the-front-and-backdoors-of-llms---kasimir-schulz-kenneth-yeung-524z4aok1"},"datePublished":"2026-07-28T13:25:52.140Z","dateModified":"2026-07-28T13:26:12.553Z","description":"Security researchers from Hidden Layer present a comprehensive breakdown of LLM attack surfaces, covering both 'front doors' (prompt injection techniques) and...","image":"https://i.ytimg.com/vi/5XlD2gn8iIA/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/5XlD2gn8iIA/sddefault.jpg","isAccessibleForFree":true,"articleSection":"NDC Conferences","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"NDC Conferences","logo":"https://media.daily.dev/image/upload/s--SiIyojEj--/f_auto/v1748760369/logos/ndc","url":"https://daily.dev/sources/ndc"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/doors-of-ai-pportunity-the-front-and-backdoors-of-llms---kasimir-schulz-kenneth-yeung-524z4aok1","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"security,machine-learning,prompt-injection","timeRequired":"PT53M","video":{"@type":"VideoObject","name":"Doors of (AI)pportunity: The Front and Backdoors of LLMs - Kasimir Schulz & Kenneth Yeung","description":"Security researchers from Hidden Layer present a comprehensive breakdown of LLM attack surfaces, covering both 'front doors' (prompt injection techniques) and...","thumbnailUrl":"https://i.ytimg.com/vi/5XlD2gn8iIA/sddefault.jpg","uploadDate":"2026-07-28T13:25:52.140Z","duration":"PT53M","url":"https://api.daily.dev/r/524Z4AOK1","embedUrl":"https://www.youtube.com/embed/5XlD2gn8iIA"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"NDC Conferences","item":"https://daily.dev/sources/ndc"},{"@type":"ListItem","position":3,"name":"Doors of (AI)pportunity: The Front and Backdoors of LLMs - Kasimir Schulz & Kenneth Yeung"}]} ```