Search: “prompt-injection”

arXiv:2608.10260v1 Announce Type: new Abstract: Lens methods interpret large language models (LLMs) by mapping intermediate activations to the output vocabulary, revealing how next-token predictions d

arXiv:2608.09124v1 Announce Type: new Abstract: Temporal decisions in language-model systems often depend on both symbolic task state and elapsed wall-clock time, such as cache expiration, job complet

arXiv:2608.08100v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) enables large language models to classify network flows and generate human-readable incident reports by retrievin

Stealing Reasoning Traces from Proprietary LLM APIs A vanity domain name (stolen-thoughts.com) for a neat paper: Anthropic, OpenAI, and Google return encrypted chain-of-thought blocks to clients that

arXiv:2608.07808v1 Announce Type: cross Abstract: Four years after prompt injection was first identified in 2022, attacks are still predominantly documented as verbatim strings rather than structured

Auto mode is now the default in Claude Code for Pro, Max, and Team plans Anthropic are really confident in Claude Code's auto mode, to the point that they are making it the default setting for new ses

Incident Report: unsanctioned agent behaviour during cyber testing It happened again. This time it was the UK government's AI Security Institute who accidentally attacked other companies while running

arXiv:2608.00005v1 Announce Type: new Abstract: Peer review at major venues is under unprecedented submission pressure, motivating the use of large language models (LLMs) as review assistants. Existin

arXiv:2607.26639v1 Announce Type: cross Abstract: A self-check defense asks the target model to assess a request before answering it; SAGE, the strongest published instance, reports an average 99% def

After a few more hours, I think I've figured out Opus 5. Opus 5 is trained to be more agentic than anything I've used. All Claude 5 models are like that. So what changes? The way to interact with Opus

AI Worming through Word Neat new prompt injection variant by Håkon Måløy, who found a way to upgrade prompt injection attacks against Microsoft Word to full self-replicating worms: An attacker places

More than any of these eval scores, what is most exciting to me is something else: Opus 5 is our least prompt injectable model yet. It is a bit buried in the system card, but across PI evals and red t

arXiv:2603.13026v2 Announce Type: replace Abstract: Prompt injection poses serious security risks to real-world LLM applications, particularly autonomous agents. Although many defenses have been propo

arXiv:2607.20090v1 Announce Type: cross Abstract: Retrieval-augmented large language models frequently face contexts that interleave useful evidence with misleading statements or instruction-like cont

How I tricked Claude into leaking your deepest, darkest secrets I've been impressed by the way the Claude web_fetch tool is designed to avoid data exfiltration attacks. Ayush Paul found a hole in that

arXiv:2607.04222v1 Announce Type: new Abstract: Interpretability methods aim to reveal the features represented inside large language models (LLMs). Many existing methods begin with labeled examples o

arXiv:2607.02514v1 Announce Type: new Abstract: As AI coding agents become more autonomous, they increasingly ship code iteratively, with the codebase persisting across sessions. This persistence crea

arXiv:2607.01153v1 Announce Type: cross Abstract: Safety evaluations for language models increasingly depend on judgments about ambiguous natural-language behaviour: whether a model has followed an in

arXiv:2606.20470v2 Announce Type: replace-cross Abstract: Agentic AI systems increasingly rely on language-model components to interpret instructions, process external data, invoke tools, and coordina

arXiv:2606.27567v1 Announce Type: cross Abstract: Prompt injection is the top security risk for LLM-integrated applications, yet every defense proposed so far has been broken. We prove this is not a c

Incident Report: CVE-2026-LGTM Spectacular hypothetical incident report by Andrew Nesbitt. Day 2, 16:00 UTC --- Two AI review agents from competing vendors, both attached to a downstream pull request

What happened after 2,000 people tried to hack my AI assistant Fernando Irarrázaval ran a challenge on hackmyclaw.com to see if anyone could leak secrets held by his OpenClaw test instance by sending

arXiv:2606.09563v1 Announce Type: new Abstract: As LLMs are deployed as agents, reliable monitoring requires knowing not only what they output, but which instructions are steering their behavior. This

arXiv:2606.05449v1 Announce Type: new Abstract: Agentic artificial intelligence (AI) systems are transforming the risk landscape by extending beyond information generation to autonomous planning, tool

OpenAI Help: Lockdown Mode OpenAI first teased this in February, but now it's live and 'rolling out to eligible personal accounts, including Free, Go, Plus, and Pro, and self-serve ChatGPT Business ac

arXiv:2605.31042v1 Announce Type: cross Abstract: LLM agents are evolving from conversational chatbots to operational tools in real-world workspaces. In local agentic harnesses, an LLM can read and wr

Hackers Simply Asked Meta AI to Give Them Access to High-Profile Instagram Accounts. It Worked I had trouble believing this story was true, but I've seen it verified from multiple sources now: One vid

arXiv:2605.28999v1 Announce Type: cross Abstract: LLMs are vulnerable to prompt injection attacks. However, this vulnerability has been primarily demonstrated conceptually in academic studies or throu

arXiv:2605.30189v1 Announce Type: cross Abstract: We show that LoRA adapters, the dominant distribution format for fine-tuned LLMs, can be reliably backdoored through training data poisoning while pre

arXiv:2605.28116v1 Announce Type: cross Abstract: Mobile graphical user interface (GUI) agents driven by vision-language models (VLMs) perceive the screen as rendered pixels and choose actions from wh

Microsoft Copilot Cowork Exfiltrates Files The biggest challenge in designing agentic systems continues to be preventing them from enabling attackers to exfiltrate data. In this case Microsoft Copilot

It's hard to find much to write about Google I/O this year because I have a policy of not writing about anything that I can't try out myself, and a lot of the big announcements are 'coming soon'. I ac

arXiv:2605.18133v1 Announce Type: cross Abstract: LLM-based chatbot agents increasingly process user requests by combining natural-language reasoning with external tools such as web browsing. These ca

arXiv:2602.14211v2 Announce Type: replace-cross Abstract: Agent skills are increasingly used to extend LLM agents with task-specific instructions, executable scripts, and auxiliary resources. While im

arXiv:2605.16233v1 Announce Type: new Abstract: Can LLM agents improve decision-making through self-generated memory without gradient updates? We propose FORGE (Failure-Optimized Reflective Graduation

arXiv:2605.13044v1 Announce Type: cross Abstract: LLM-powered agents can silently delete documents, leak credentials, or transfer funds on a routine user request, not because the agent was attacked, b

arXiv:2605.12233v1 Announce Type: new Abstract: Terminal agents are increasingly capable of executing complex, long-horizon tasks autonomously from a single user prompt. To do so, they must interpret

arXiv:2605.06669v1 Announce Type: cross Abstract: Educational LLM tutors face a core AI alignment challenge: they must follow user intent while preserving pedagogical constraints and safety policies.

arXiv:2605.01449v1 Announce Type: cross Abstract: Universal adversarial attacks on aligned multimodal large language models are increasingly reported with attack success rates in the 60-80% range, sug

arXiv:2605.02187v1 Announce Type: cross Abstract: Bring-Your-Own-Key (BYOK) agent architectures let users route LLM traffic through third-party relays, creating a critical integrity gap: a malicious r

arXiv:2604.24920v2 Announce Type: replace-cross Abstract: Agentic systems increasingly act with user secrets for APIs, messaging platforms, and cloud services. Today's bearer-secret interfaces impleme

arXiv:2604.23270v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting has emerged as a simple and effective way to elicit step-by-step solutions from large language models (LLMs). However,

arXiv:2603.11331v2 Announce Type: replace-cross Abstract: Adversarial attacks can reliably steer safety-aligned large language models toward unsafe behavior. Empirically, we find that strong adversari

Thoth v3.14.0: multi-provider cloud and multi-provider image generation. Five cloud providers now work as first-class citizens: OpenAI, Anthropic, Google, xAI, and OpenRouter. Image generation spans t