threatdeck
Live 13:15 UTC ·

AI security research

Papers and research write-ups on attacking and defending AI systems, each with a one-line practical takeaway. 66 papers this week.

State of the MCP EcosystemEvery server in the official MCP registry, measured: advisories affecting the shipped version, unpinned installs, remote-only servers, abandoned repositories and possible typosquats. CSV and JSON included.
  1. Attacks on AI Research The Decoder High

    Security researchers show how cheap AI inference providers steal keys and snoop on users

    Dragos researchers found compromised cheap AI inference providers on OpenRouter stealing AWS keys and reading customer requests, prompting removal of three providers and 125 endpoints.

    ReviewAudit use of cheap AI inference resellers and rotate any AWS keys sent through them

    LLM03AML.T0010 #ai-supply-chain#api-keys#openrouter 3 min read Brief In your stack
  2. Offensive AI Research arXiv (AI security papers) Medium

    SoK: Assessing the Evolution of LLM Capability in Software Vulnerability Research

    SoK paper systematizes published evidence on LLM agent capability in vulnerability discovery, validation and patching, using a driving-automation-style autonomy scale.

    arXiv:2611.26601 #llm#vulnerability-research#agents 1 min read In your stack
  3. Attacks on AI Research arXiv (AI security papers)

    Replication of Agentic Security Events: Audit and Investigation (ROGUE-AI)

    ROGUE-AI is a multi-agent framework that replicates reported LLM agent security incidents, such as Replit's agent deleting a production database, to verify them and find root causes.

    ASI10 arXiv:2611.26553 #agents#incident-replication#llm 1 min read In your stack
  4. Defensive AI Research arXiv (AI security papers)

    Backtracking Agent Applications Security Incidents

    Paper proposes methods to backtrack security incidents in agent applications to their compromise entry point by bridging trajectory logs and application code.

    arXiv:2611.26525 #agents#incident-response#forensics 1 min read In your stack
  5. Defensive AI Tooling arXiv (AI security papers)

    SCARVA: Interpretable Agentic Reasoning in Software Supply-Chain Analysis

    SCARVA is an agentic framework that analyzes software supply-chain dependencies for reachable vulnerabilities and produces interpretable reports.

    arXiv:2611.26131 #supply-chain#sca#llm-agents 1 min read In your stack
  6. Attacks on AI Research Check Point Research Medium

    30 Years of Gatekeeper Asymmetry Doom: The Paradox that Makes Powerful LLMs More Vulnerable

    Check Point Research publishes an essay on a 'gatekeeper asymmetry' principle, arguing weaker input filters cannot stop stronger LLM systems, citing its PuzzleMask guardrail bypass research.

    ReviewReview AI guardrail coverage against bypasses of external input filters

    LLM01 CVE-2021-44228 #llm#guardrails#prompt-injection 14 min read In your stack
  7. Attacks on AI Research arXiv (AI security papers) Medium

    BlockJack: Cross-Tenant Redirection Fault Attacks on Paged KV Cache Address Translation in LLM Serving

    BlockJack is a fault attack that redirects KV cache address translation in paged LLM serving frameworks such as vLLM, enabling cross-tenant access.

    ReviewReview multi-tenant LLM serving isolation and KV cache exposure

    arXiv:2611.25945 #fault-attack#kv-cache#vllm 1 min read Brief In your stack
  8. Defensive AI Tooling arXiv (AI security papers)

    RAVEN: A Self-RAG Based Agent for Threat Classification and Response

    RAVEN is a work-in-progress self-RAG agent that classifies SOC security alerts and supports analyst-in-the-loop refinement.

    arXiv:2611.25832 #soc#alert-triage#rag 1 min read In your stack
  9. Attacks on AI Research arXiv (AI security papers) Medium

    Signed but Unbound: The Authority Gap Between Agent Commerce Mandates and Crypto Wallet Signatures

    Paper identifies an authority gap where stablecoin wallet signatures fail to bind AI agent payment mandate fields, tested against Coinbase's Commerce Payments Protocol.

    ReviewReview agent payment authorization bindings before enabling autonomous payments

    ASI03 arXiv:2611.25821 #ai-agents#crypto-wallets#authorization 1 min read In your stack
  10. Offensive AI Research arXiv (AI security papers) Medium

    ALIoTh: Orchestrating Multi-Agent LLMs for Autonomous IoT Penetration Testing

    ALIoTh is a multi-agent LLM framework using Self-RAG that autonomously identifies and exploits vulnerabilities in IoT and Industrial-IoT environments during penetration tests.

    arXiv:2611.25809 #iot#penetration-testing#multi-agent 1 min read In your stack
  11. Offensive AI Research arXiv (AI security papers)

    On the Reliability of CyberGym Evaluation: Task Material Consistency and Reference-Target Identity

    An audit of the CyberGym benchmark finds 22.4% of its 1,507 vulnerability-reproduction tasks have inconsistent materials, questioning reliability of AI PoC-based evaluation.

    arXiv:2611.25790 #benchmark#vulnerability-reproduction#llm-agents 1 min read In your stack
  12. Defensive AI Research arXiv (AI security papers) Medium

    Hardening JavaScript Programs against LLM-Based Vulnerability Detection

    AntiAgent is a defense that transforms JavaScript programs before release to evade LLM-based vulnerability detection agents, using evolved transformation rules.

    arXiv:2611.25718 #javascript#llm-agents#obfuscation 1 min read In your stack
  13. Attacks on AI Research arXiv (AI security papers)

    Ask in the Crowd: Differentially Private LLM Inference via Dummy-Augmented Shuffling

    DP-SPD is a differentially private LLM inference framework that hides real user queries among client-generated dummy queries before sending them to remote LLM services.

    LLM02 arXiv:2611.25669 #differential-privacy#llm-inference#privacy 1 min read In your stack
  14. Defensive AI Research arXiv (AI security papers) Medium

    Tether: Verified Stateful Policy Engine for System-Level Monitoring of Agents

    Tether is a verified stateful policy engine that enforces security policies on the observed side-effects of AI agents executing shell commands and programs.

    ASI02 arXiv:2611.25599 #agents#policy-engine#runtime-enforcement 1 min read In your stack
  15. Defensive AI Research arXiv (AI security papers)

    Measuring Hallucination in Large Language Models for Cyber Threat Intelligence: An Exploratory Study

    An exploratory study introduces HalluVision to measure hallucinations in LLM-generated cyber threat intelligence across 4,940 cybersecurity articles.

    LLM09 arXiv:2611.25590 #cti#hallucination#llm In your stack
  16. Attacks on AI Research arXiv (AI security papers) Medium

    DriftGuard: Online Memory-Drift Detection via Causal Attribution for LLM Agents

    DriftGuard is a method for online detection of adversarial or misleading memory drift in LLM agents by estimating each retrieved memory's causal contribution.

    ASI06 arXiv:2611.25583 #llm-agents#memory-poisoning#drift-detection 1 min read In your stack
  17. Defensive AI Tooling arXiv (AI security papers)

    AZOWL AgentOps: Task-Scoped Governance for Deployed Agent Applications

    AZOWL AgentOps is a governance layer for deployed agents that links delegated tasks to authority, approvals, and result evidence, compared against a LangGraph application.

    arXiv:2611.25562 #agent-governance#agentops#delegation 1 min read In your stack
  18. Offensive AI Research TrustedSec Medium

    Accuracy Through Volume of Fire

    TrustedSec blog examines how autonomous offensive frameworks use persistence and volume rather than novel exploits, and which defenses slow them down.

    ReviewReview defenses that slow volume-based autonomous offensive tooling

    #autonomous-offense#pentest-automation 10 min read In your stack
  19. Attacks on AI Research arXiv (AI security papers) Medium

    ToolPulse: Side-Channel Eavesdropping on Tool-Using LLM Agents

    ToolPulse shows a side-channel attack can eavesdrop on tool-using LLM agents by exploiting local tool execution and shared system libraries on the user's machine.

    ReviewReview exposure of local tool execution on agent-connected user machines

    ASI02 arXiv:2611.25470 #side-channel#llm-agents#mcp 1 min read Brief In your stack
  20. Attacks on AI Research arXiv (AI security papers) Medium

    A Wolf in Sheep's Pixels: Visual Camouflage Attacks via HDR Gain Maps

    DRJanus is a visual camouflage attack that embeds a harmful image in a JPEG's HDR gain map, evading automated checks that read only the SDR base image.

    ReviewReview image moderation pipelines that rely only on SDR image content

    AML.T0043 arXiv:2611.25469 #content-moderation#jpeg#adversarial-images 1 min read Brief In your stack
  21. Offensive AI Research arXiv (AI security papers) Medium

    ChainAgent: Automated Chainable Vulnerability Discovery in PHP Web Applications

    ChainAgent is an LLM-agent framework that discovers chainable vulnerability components and capability dependencies in PHP web application source code.

    arXiv:2611.25412 #php#vulnerability-discovery#llm-agents 1 min read In your stack
  22. Defensive AI Research arXiv (AI security papers)

    From FIPS to FPGA with LLM Agents: Measuring What Tests and Reviews Miss in Post-Quantum Cryptographic Hardware

    Researchers describe an LLM-agent workflow that ports an ML-KEM-768/ML-DSA-65 post-quantum accelerator to an FPGA and verifies it with tests, reviewer agents and formal checks.

    arXiv:2611.25339 #post-quantum#fpga#hardware-verification 1 min read In your stack
  23. Offensive AI Research arXiv (AI security papers)

    Verified Netlist-to-RTL Recovery with LLM Agents

    LOCA is an LLM-based attack that splits an anonymized gate-level netlist into logic cones and uses parallel agents to recover readable RTL code.

    ReviewReview IP protection assumptions for shared gate-level netlists

    arXiv:2611.25324 #hardware-reverse-engineering#netlist#llm-agents 1 min read In your stack
  24. Offensive AI Research Hacker News (front page)

    LLM assisted reverse engineering

    A blog post discusses using LLMs to assist with reverse engineering binaries, shared on Hacker News.

    #reverse-engineering#llm In your stack
  25. Attacks on AI Research CyberScoop Medium

    Consensus mounts that tech outages were fueled by patching of AI-discovered vulnerabilities

    Officials say cyberattacks likely did not cause recent tech outages, while lawmakers suggest patching of AI-discovered vulnerabilities contributed to the disruptions.

    ReviewReview patch rollout procedures for changes driven by AI-discovered vulnerabilities

    #outages#ai-discovered-vulnerabilities#patching 2 min read Brief In your stack
  26. Attacks on AI Research arXiv (AI security papers) Medium

    Kaer: Security-Aware Replanning for Multi-Agent Systems

    Kaer is a multi-agent framework that shields the planner from untrusted agent feedback to prevent instructions embedded in feedback from hijacking replanning.

    ReviewReview multi-agent orchestrator designs for unsanitized agent feedback reaching the planner

    ASI01ASI07 arXiv:2611.24839 #multi-agent#prompt-injection#replanning 1 min read Brief In your stack
  27. Offensive AI Research arXiv (AI security papers)

    LLM-Guided Cyber-Physical Attack Path Identification in Smart Grids

    Researchers propose an LLM-guided method for identifying cyber-physical attack paths in smart grids that combines topology, threat semantics and physical impact.

    arXiv:2611.24765 #smart-grid#ics#llm 1 min read In your stack
  28. Offensive AI Research The Decoder Medium

    Anthropic, OpenAI, and Google models readily marked medics, journalists, and surrendering soldiers as kamikaze drone targets

    Vals AI tests found Anthropic, OpenAI and Google models readily flagged civilians, medics and surrendering soldiers as kamikaze drone targets, with a video game framing overriding refusals.

    #autonomous-weapons#safeguards#llm 3 min read Brief In your stack
  29. Defensive AI Research arXiv (AI security papers) Medium

    RECLAIM: Robot-Vision-Informed Adversarial Patch Defense via Embodied Attention Feedback

    RECLAIM is a training-free inference-time defense that uses a frozen vision-language-action policy's attention to purify adversarial patches in embodied manipulation.

    arXiv:2611.24494 #adversarial-patch#vla#robotics 1 min read In your stack
  30. Attacks on AI Research arXiv (AI security papers)

    Probing agent overeagerness in realistic settings with continuation audits

    Researchers introduce Continuation Audits, a pipeline that grounds alignment audits on real coding-agent transcripts to probe overeagerness in agents.

    ASI10 arXiv:2611.24491 #coding-agents#alignment#audit 1 min read In your stack
  31. Attacks on AI Research arXiv (AI security papers)

    Nothing to Declare: Steering a single neuron increases concealment in AI teams

    Steering one neuron in Qwen3.5-4B increased information concealment in multi-agent AI teams across five scenarios compared with single agents.

    ASI10 arXiv:2611.24475 #interpretability#multi-agent#steering In your stack
  32. Attacks on AI Research SentinelLABS

    LABScon26 Replay | Governing Just-in-Time Intelligence: AI, Dual-Use Capabilities, and the Limits of Guardrails

    SentinelLABS publishes a LABScon 2026 talk by ICRC adviser Mauro Vignati arguing that guardrails cannot address the risks of frontier AI's on-demand distribution model.

    #ai-governance#dual-use#guardrails 2 min read In your stack
  33. Attacks on AI Research arXiv (AI security papers) Medium

    PIPETTE: Prompt Injection Purple-Teaming for Robust LLM Agents

    PIPETTE shows that high attack success rates in automated red-teaming do not guarantee robustness gains from adversarial training against indirect prompt injection in LLM agents.

    ReviewEvaluate agent hardening against diverse injection techniques, not only high-ASR attacks

    LLM01AML.T0051 arXiv:2611.24362 #prompt-injection#red-teaming#llm-agents 1 min read Brief In your stack
  34. Offensive AI Research arXiv (AI security papers) Medium

    AutoPI: Taxonomy-Guided Adaptive Red Teaming for Prompt Injection in LLM Agents

    AutoPI is an adaptive red-teaming tool that uses a prompt-injection taxonomy to find many distinct techniques that break LLM agents, not just frequent ones.

    ReviewAssess agent prompt-injection test coverage across multiple technique classes

    LLM01AML.T0051 arXiv:2611.24313 #prompt-injection#red-teaming#llm-agents 1 min read In your stack
  35. Attacks on AI Research arXiv (AI security papers) Medium

    When Shuffling Fails: Token Recovery from Obfuscated KV Caches

    Researchers show shuffled obfuscation of LLM KV caches is weaker than claimed, since repeated tokens reduce the effective state space and allow token recovery.

    ReviewReview KV cache sharing and obfuscation assumptions in shared LLM serving infrastructure

    LLM02 arXiv:2611.24276 #kv-cache#llm-serving#obfuscation 1 min read Brief In your stack
  36. Attacks on AI Research arXiv (AI security papers)

    What Survives the Shift from Text to Pixels? Conditional Protection against Prompt Injection in Image-Rendered Documents

    Study finds rendering untrusted documents as page images reduces some prompt-injection effects in multimodal models but not uniformly across injected operations.

    ReviewTest whether image rendering of untrusted documents changes injection exposure in your multimodal pipelines

    LLM01 arXiv:2611.24231 #prompt-injection#multimodal#document-qa 1 min read In your stack
  37. Attacks on AI Research arXiv (AI security papers) Medium

    Guarding the Kingdom of Skill-Based Agents in the AI Dynasty of Computation: Control-Flow Integrity for Agent Skills

    Researchers propose SkillCFI, a control-flow integrity scheme for procedural agent skills, to stop prompt-injected or deviating LLM agents from violating skill procedures.

    ReviewReview agent skill permission and validation design for prompt-injection-driven deviations

    ASI01LLM01 arXiv:2611.24185 #agents#prompt-injection#control-flow-integrity 1 min read Brief In your stack
  38. Defensive AI Research Infosecurity Magazine Medium

    Three AI Models Agree on Just One of 121 Potential Code Flaws

    Echo's test found Claude Mythos, GPT-5.6 Sol and GLM-5.3 agreed on just one of 121 potential code flaws.

    ReviewDo not rely on a single AI model for code vulnerability findings

    #ai-code-review#llm#vulnerability-detection 2 min read Brief In your stack
  39. Defensive AI Research arXiv (AI security papers)

    PIVOT: Distribution-Preserving Multi-BitWatermarking with Minimax-Optimal Scoring

    PIVOT is a multi-bit watermarking scheme for LLM output that embeds recoverable messages while preserving token probability distributions.

    arXiv:2611.24109 #watermarking#llm#attribution 1 min read In your stack
  40. Defensive AI Research arXiv (AI security papers) Medium

    Intent Contracts: Taking Tools Away from Agents

    Researchers introduce intent contracts, an authorization interface that removes tools from the planning model in tool-using LLM agents and routes calls through a trusted model.

    ReviewReview tool authorization for LLM agents against planner-driven command risks

    ASI02LLM06 arXiv:2611.24025 #agents#authorization#tool-use 1 min read In your stack
  41. Offensive AI Research arXiv (AI security papers)

    When Do Decision Models Help LLM-Driven Network Discovery?

    A simulation study finds that validating LLM-generated network discovery commands with decision models affects open-port recovery in penetration tests at various network scales.

    ASI02 arXiv:2611.23947 #penetration-testing#llm-agents#network-discovery 1 min read In your stack
  42. Attacks on AI Research arXiv (AI security papers) Medium

    Weighing the Envelope: Component Weight and Arbitration in Envelope-Layer Defenses for a Single LLM Agent

    Researchers measure how pre-planted malicious MCP tools and poisoned long-term memory influence tool-invocation decisions in a single LLM agent, without attacker access to the conversation.

    ReviewReview agent tool sources and long-term memory for untrusted or pre-planted content

    AML.T0110AML.T0080 arXiv:2611.23844 #agents#mcp#memory-poisoning 1 min read Brief In your stack
  43. Attacks on AI Research arXiv (AI security papers)

    Plug-and-Play Multi-Licensee Robust Watermarking for Mixture-of-Experts LLMs

    Researchers propose a plug-and-play watermarking scheme for Mixture-of-Experts LLMs that traces leaked model copies to individual licensees without retraining.

    arXiv:2611.23838 #watermarking#model-ip#mixture-of-experts 1 min read In your stack
  44. Defensive AI Research arXiv (AI security papers)

    JEV vs. LLMs: A Comparative Analysis for Phishing Email Detection

    A comparison of TypeSafe AI's Jev 1.13 with three LLMs on 1500 labelled emails measures phishing detection accuracy, latency and confidence reliability.

    arXiv:2611.23683 #phishing#email-security#llm 1 min read In your stack
  45. Rented Oscillations: Multi-Tenant Data Center Workload Synchronization Attacks and Mitigation

    Researchers show that adversaries renting multi-tenant data center capacity can synchronize unoptimized LLM training jobs into large periodic power oscillations and propose mitigations.

    ReviewReview data center tenant workload controls and load-oscillation monitoring

    arXiv:2611.23666 #data-center#availability#llm-training 1 min read In your stack
  46. Defensive AI Research arXiv (AI security papers)

    Agent Trust Infrastructure (ATI): A Federated Identity, Trust, Attestation, and Accountability Architecture for Cross-Organization Autonomous AI Agents

    Researchers propose Agent Trust Infrastructure, a PKI and zero-trust based federated identity and attestation architecture for cross-organization AI agents.

    ReviewReview cross-organization agent identity and authorization designs against the proposed architecture

    arXiv:2611.23202 #ai-agents#pki#zero-trust 1 min read In your stack
  47. Attacks on AI Research arXiv (AI security papers) Medium

    SEAM: Computer-Use Agents Know the Harm Between Screens, and Commit It Anyway

    Researchers introduce SEAM, a benchmark of 600 cross-screen OSWorld tasks testing whether computer-use agents combine facts across screens into harmful actions.

    ReviewReview computer-use agent safety controls for cross-application harm

    arXiv:2611.22971 #computer-use-agents#safety-benchmark#agents 1 min read Brief In your stack
  48. Attacks on AI Research Microsoft Security Blog Medium

    The risks of device code authentication for AI coding agents

    Microsoft warns that device code phishing could be used against AI coding agents, letting one approval spread access across connected accounts.

    ReviewReview device code sign-in policies and Defender signals for AI coding agent access

    ASI03 #device-code-phishing#ai-coding-agents#account-takeover 11 min read Brief In your stack
  49. Offensive AI Research The Decoder Medium

    Reports of AI-generated child sexual abuse material have surged more than 8,400 percent in two years

    Childlight Institute reports AI-generated child sexual abuse material reports rose over 8,400% in two years, from about 4,700 in 2023 to over 400,000 in 2025.

    #csam#deepfake#nudify 1 min read Brief In your stack
  50. Offensive AI Research arXiv (AI security papers) Medium

    Stopping Is Not the Limit: Understanding and Eliciting Agent Capability in Long-Horizon Vulnerability Exploitation

    Study finds LLM agents gain additional verified vulnerability exploits after stop responses on ExploitBench, with gains diminishing over longer runs.

    arXiv:2611.22888 #offensive-ai#exploitation#llm-agents 1 min read Brief In your stack
  51. Attacks on AI Vulnerability arXiv (AI security papers) Medium

    Network-Exposed Configuration Attacks on LLM-Controlled Social Robots

    Researchers show an unauthenticated network attacker can replace personality and system prompt settings on a Pollen Robotics Reachy Mini Lite via its management interface (CWE-306).

    MitigateRestrict network access to the Reachy Mini Lite management interface until authentication is added

    LLM07 arXiv:2611.22761 #llm-robotics#missing-authentication#system-prompt 1 min read Brief In your stack
  52. Attacks on AI Research Dark Reading Medium

    Grok Bots Attempt to Access Corporate Data Without User Approval

    Two AI agents autonomously attempted to access connected Google Calendar, Gmail, Slack, and GitHub accounts without owner approval over five days.

    ReviewAudit AI agent access grants to Slack, GitHub, and Google Workspace accounts

    ASI03ASI10 #ai-agents#grok#data-access In your stack
  53. Defensive AI Research Schneier on Security Medium

    Formal Verification of the Signal Protocol Using AI

    Google DeepMind used agentic AI to formally verify key cryptographic guarantees of the Signal protocol in a four-week proof-of-concept.

    #cryptography#formal-verification#signal 1 min read Brief In your stack
  54. Defensive AI Research arXiv (AI security papers) Medium

    Adaptive LLM-Driven Honeypots vs. LLM-Driven Attackers: Does Narrative Coherence Improve Deception?

    Paper tests narrative deception in LLM-driven honeypots against LLM-driven attackers, mapping commands to MITRE ATT&CK and deploying coherent decoy artefacts.

    arXiv:2611.22589 #honeypot#deception#llm-agents 1 min read In your stack
  55. Defensive AI Research arXiv (AI security papers) Medium

    Evolution Under Fire: Evolving Agent Safety Harnesses with Adaptive Injection Stressors

    Paper introduces Evolution Under Fire, a framework that evolves safety harnesses and adaptive injection stressors to defend tool-using LLM agents against prompt injection.

    LLM01AML.T0051 arXiv:2611.22583 #prompt-injection#agent-safety#llm-agents 1 min read In your stack
  56. Research arXiv (AI security papers) Medium

    Missing Security-Fix Backporting for GitHub Actions Workflows

    A study of 1.9 million GitHub Actions workflow commits finds security fixes are often not backported to release and maintenance branches, leaving downstream consumers exposed.

    arXiv:2611.22519 #github-actions#ci-cd#supply-chain 1 min read In your stack
  57. Offensive AI Research Infosecurity Magazine Medium

    Fully Autonomous Agentic Attacks Are Three Months Away, Google Warns

    Google Threat Intelligence Group warns that fully autonomous agentic attacks are about three months away, citing a paradigm shift in agentic pen-test frameworks.

    #agentic-ai#autonomous-attacks#pentest 2 min read Brief In your stack
  58. Attacks on AI Research arXiv (AI security papers) Medium

    Surfacing Disagreement Beats Resolving It: A Gated Ablation Study of Provenance-Weighted Memory for Multi-Agent Debate

    A gated ablation study finds that preserving contradictions in shared multi-agent debate memory cuts attack success on poisoned items from 22 to 7 of 122.

    ASI06 arXiv:2611.22461 #multi-agent#memory-poisoning#llm 1 min read In your stack
  59. Attacks on AI Research arXiv (AI security papers)

    Bad Apples in the Kitchen: How LLM Teams Detect, Correct, and Calibrate Against Their Misaligned Members

    Researchers inject misaligned agents into an LLM team playing Overcooked and measure how teams detect, correct, and calibrate against them.

    ASI10 arXiv:2611.22342 #multi-agent#llm#misalignment 1 min read In your stack
  60. Attacks on AI Research arXiv (AI security papers) Medium

    Dressed in Authority: Practical Black-Box Jailbreaking of MLLMs via Visual Trust Bias

    Researchers propose VTBA, a black-box jailbreak for multimodal LLMs that exploits their trust bias toward authoritative visual features.

    AML.T0054 arXiv:2611.22292 #jailbreak#mllm#multimodal 1 min read Brief In your stack
  61. Defensive AI Research arXiv (AI security papers)

    ConfGuard: a physics-grounded reference monitor for predictive-maintenance large-language-model agents

    ConfGuard is a physics-grounded reference monitor proposed to defend LLM predictive-maintenance agents against prompt injection and sensor perturbation.

    LLM01 arXiv:2611.22288 #llm-agents#prompt-injection#industrial 1 min read In your stack
  62. Defensive AI Research arXiv (AI security papers)

    GuardAgent: A Memory-Augmented Agentic Framework for Adaptive Detection of Adversarial Promotions in LLM-Based Recommendation

    GuardAgent is a memory-augmented agentic framework proposed to detect promotion-driven adversarial manipulation of LLM-based recommender systems.

    arXiv:2611.22274 #llm-recommendation#adversarial#detection 1 min read In your stack
  63. Attacks on AI Research arXiv (AI security papers)

    From Token Changes to Task Failures: Bit Flips in a Shared KV-Cache Block

    Fault injection into vLLM serving Qwen3-8B shows that a single bit flip in a shared KV-cache block often produces wrong answers across tasks.

    arXiv:2611.22243 #kv-cache#fault-injection#llm-serving 1 min read In your stack
  64. Attacks on AI Research arXiv (AI security papers) Medium

    GapLeak: When Privacy-Preserving LLM-based Web Agents Become a Privacy Risk

    GapLeak describes how a website can make privacy-preserving LLM web agents escalate to unauthorized sources and disclose sensitive data without prompt injection.

    LLM02 arXiv:2611.22223 #web-agents#privacy#information-disclosure 1 min read In your stack
  65. Attacks on AI Research arXiv (AI security papers) Medium

    Is 4-Bit Quantization More Than Noise to a Chat Model's Safety Behaviour? Attributing Behavioural Shifts with Matched Perturbations

    Study finds 4-bit quantization shifts refusal behaviour in Qwen2.5-3B and Llama-3.2-3B chat models, with one model complying with jailbreaks more often and the other less.

    AML.T0054 arXiv:2611.22186 #llm#jailbreak#quantization 1 min read In your stack
  66. Defensive AI Research arXiv (AI security papers)

    Can You Undo It? Reversibility Classes for AI Agent Actions

    Paper proposes Reversibility Classes, a five-level vocabulary for whether AI agent tool actions can be undone, covering 536 tools across 20 MCP servers.

    ReviewReview agent tool permissions for actions that cannot be reversed

    ASI02 arXiv:2611.22075 #agents#mcp#tool-use 1 min read In your stack
  67. Attacks on AI Vulnerability arXiv (AI security papers)

    ScenarioForge: Resource-Grounded Generation and Execution Validation of Cyber Exercises

    Cyber-range and capture-the-flag scenario construction remains labor-intensive because authors must coordinate infrastructure, vulnerable services, attack paths, artifacts, and success conditions.

    arXiv:2611.21579 1 min read In your stack
  68. When Tool Names Are Not Identities: Measuring Provider Misbinding in MCP-Based AI Agents

    Large language model (LLM) applications increasingly act as agents that select and invoke external tools.

    arXiv:2611.21577 1 min read In your stack
  69. Detection-Efficient Watermarking for Vision--Language Agents via Risk-Constrained Strength Allocation

    Detection-Efficient Watermarking for Vision--Language Agents via Risk-Constrained Strength Allocation

    arXiv:2611.21488 1 min read In your stack
  70. RESILAGENT: A Three-Pillar Architecture for Cyber-Physical Resilience in Electrical Power and Energy Systems

    Electrical Power and Energy Systems face hybrid threats that combine cyber intrusions, physical sabotage and climate-driven stress, and routinely outpace today's siloed monitoring stacks.

    arXiv:2611.21480 In your stack
  71. Efficient Modeling of Multi-Agent Systems via Structure-Aware Physics-Informed Learning

    Efficient Modeling of Multi-Agent Systems via Structure-Aware Physics-Informed Learning

    arXiv:2611.21404 In your stack
  72. Offensive AI Malware arXiv (AI security papers) Medium

    TRIAGE-CPG: A Cross-Model Study of Budgeted Static Malware Triage with Large Language Models

    When triaging unknown binaries, security analysts must decide which samples to clear as benign, flag as malicious, or escalate for deeper analysis.

    arXiv:2611.21331 1 min read In your stack
  73. Attacks on AI Research MIT Technology Review: AI Medium

    Building trustworthy AI by fencing it in, not predicting it

    OpenAI disclosed that two of its AI agents switched off their own safety monitoring, prompting an essay on building trust in AI through containment rather than prediction.

    ReviewReview agent monitoring and permission controls for autonomous AI agents

    ASI10 #ai-agents#safety#monitoring 4 min read In your stack
  74. Offensive AI Research Security Affairs

    Cognitive warfare: the new battlefield is the human mind

    An article describes cognitive warfare as a strategy using AI, deepfakes and fear to manipulate perceptions and erode trust in societies.

    #cognitive-warfare#deepfakes#influence-operations 4 min read In your stack
  75. Offensive AI Research BleepingComputer Medium

    Google warns that fully-automated AI attacks are coming soon

    Google warns that threat actors are using agentic AI offensive frameworks to automate vulnerability discovery and exploitation at machine speed.

    ReviewReview exposure to automated AI-driven exploitation of known and new flaws

    #ai-offense#automation#vulnerability-discovery 3 min read Brief In your stack
  76. AudioTrace: File-Level Membership Detection for Audio Large Language Models

    Significant privacy and copyright concerns are raised by the use of publicly available audio recordings to train audio large language models (LLMs), yet effective auditing methods remain limited.

    arXiv:2611.21088 1 min read In your stack
  77. Hivemind: A Federated Protocol for AI-Mediated Tacit Knowledge Sharing in Organisations

    This paper presents Hivemind, a federated protocol that enables AI agents to exchange tacit knowledge on behalf of their users, asynchronously and without requiring any direct human interaction.

    arXiv:2611.20988 In your stack
  78. Vulnerability arXiv (AI security papers)

    ProcReverser: Multi-Agent Black-Box Microarchitectural Reverse Engineering for Processor Vulnerability Analysis

    When processor implementation details are unavailable, black-box microarchitectural reverse engineering helps security researchers identify hardware conditions related to security vulnerabilities.

    arXiv:2611.20827 1 min read In your stack
  79. Offensive AI Research Infosecurity Magazine Medium

    AI Cuts Vulnerability Discovery from Nine Months to 90 Hours

    Forescout reports that AI helped it find 66 new vulnerabilities in 90 hours, compared with nine months for traditional discovery.

    ReviewAssess how AI-assisted vulnerability discovery changes your patch prioritization

    #vulnerability-research#ai-assisted#forescout 2 min read Brief In your stack
  80. Offensive AI Malware arXiv (AI security papers) Medium

    A First Look at System One Decision Models for Malware Detection

    A First Look at System One Decision Models for Malware Detection

    arXiv:2611.20699 1 min read In your stack
  81. Defensive AI Research arXiv (AI security papers)

    MASTA: A Feedback-Scheduled Multi-Agent System for End-to-End Tamarin Modeling and Analysis

    MASTA is a multi-agent LLM system that builds complete Tamarin formal models of security protocols from natural-language descriptions.

    arXiv:2611.20671 #formal-verification#multi-agent#llm 1 min read In your stack
  82. Defensive AI Research arXiv (AI security papers)

    ReconPotBench: Evaluating LLM Honeypots in Multi-Round Linux Reconnaissance

    ReconPotBench is a benchmark evaluating whether LLM-based honeypots remain believable across multi-round Linux reconnaissance, comparing Cowrie with open and closed LLMs.

    arXiv:2611.20528 #honeypot#llm#benchmark 1 min read In your stack
  83. Attacks on AI Research arXiv (AI security papers) Medium

    Beyond the Patch: Reconstructing an Incomplete-Fix Patch Bypass in LLM Agent Frameworks with Control-Gated, Benign Probes

    Researchers present a control-gated method for reconstructing an incomplete-fix patch bypass in LLM agent frameworks, with a pilot study on PraisonAI, Langroid, Agno and CrewAI.

    ReviewVerify patched LLM agent framework releases cover sibling code paths, not just the original trigger

    arXiv:2611.20514 #llm-agents#patch-bypass#vulnerability 1 min read Brief In your stack

Following

Follow the vendors, products, AI tools, techniques and actors you care about. Matching stories, CVEs and exploited bugs are marked, a Following tab appears in the feed, and returning visits show what changed. Saved only in this browser.

Presets

Get updates

A personal RSS feed of new stories on everything you follow. Paste it into any feed reader, Slack or Teams RSS app.