https://vvnblog.com/en/tags/llm/Recent content in LLM on Vivian@SecMomBagHugo -- gohugo.ioenCopyright © 2026 Vivian All Rights Reserved.Sat, 15 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-14/Sat, 15 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-14/Tools, arguments, execution identity, permission scope. Work through them one at a time and you can usually see where the excessive agency is hiding.https://vvnblog.com/en/posts/ai-dark-arts-13/Fri, 14 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-13/When a model’s answer is not meant for a human but goes straight into a browser, a database or a shell, it is no different from text submitted through a form.https://vvnblog.com/en/posts/ai-dark-arts-12/Thu, 13 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-12/Levels four and five of Break The Prompt: how a full set of system instructions gets translated out in the open, and how the filter in front of the output gets walked around.https://vvnblog.com/en/posts/ai-dark-arts-11/Wed, 12 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-11/Levels one to three of Break The Prompt, and how the rules written into a system prompt come apart one at a time.https://vvnblog.com/en/posts/ai-dark-arts-10/Tue, 11 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-10/Once a model starts reading images and documents, the spell no longer has to be typed into the chat window.https://vvnblog.com/en/posts/ai-dark-arts-09/Mon, 10 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-09/How indirect prompt injection uses RAG poisoning to make an AI read malicious documents, leak data, and even fire off tool calls.https://vvnblog.com/en/posts/ai-dark-arts-08/Sun, 09 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-08/Someone asked ChatGPT to play their late grandmother, and it offered condolences while reading out five Windows keys. Jailbreaking goes after the model’s own prison.https://vvnblog.com/en/posts/ai-dark-arts-07/Sat, 08 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-07/Why does a carefully written system prompt still lose to one sentence? To an LLM, instructions and data look exactly the same.https://vvnblog.com/en/posts/ai-dark-arts-02/Mon, 03 Aug 2026 00:00:00 +0800https://vvnblog.com/en/posts/ai-dark-arts-02/AI hallucination is not a bug. It is what you get when a machine writes by chaining probabilities.