What is Lyrie Research?

Lyrie Research is an autonomous cybersecurity intelligence platform publishing verified threat intelligence including critical CVEs, active exploitation reports, breach analysis, and original research — every article cross-validated by 3+ primary sources.

How does Lyrie.ai protect against rogue AI threats?

Lyrie.ai uses machine-speed autonomous defense to detect and neutralize rogue AI, prompt injection, and AI supply chain attacks. It responds before human analysts can react.

What cybersecurity topics does Lyrie Research cover?

Lyrie Research covers CVE deep dives, CISA KEV actively exploited vulnerabilities, data breach forensics, original cybersecurity research, and AI threat intelligence including agent-based attacks.

How often is Lyrie Research updated?

Lyrie Research is updated continuously by the autonomous Lyrie Sentinel engine, publishing new threat intelligence multiple times daily as new CVEs, exploits, and breaches are confirmed.

Is Lyrie Research free to access?

Yes, all articles on Lyrie Research are freely accessible. For active protection by the same intelligence engine, visit lyrie.ai.

← Home

#agent-threats

338 stories tagged.

#agent-threats

Claude Code: Sandbox Escape via Git Worktree Path Confusion Allows Unsandboxed Code Execution

IssueTrojanBench: Benchmarking AI Coding Agents Against Malicious Issue Requests

Security Vulnerability Patterns in AI-Generated Code: A Cross-Model Comparative Study

ANSI escape injection in MCP servers: Hidden from humans, visible to AI

An Automated Framework for Extracting Reachable Attack Chains from Cyber Threat Intelligence Reports

Defense Against LLM Backdoors using Critical Neuron Isolation Pruning

LiteLLM: MCP Authentication Bypass via OAuth2 Passthrough Fallback

n8n: Member-Level Users Can Execute Other Users' MCP Server Trigger Workflows via Missing OAuth Authorization

Trusted Credentials, Untrusted Behavior: Benchmarking LLM-Agent Security in High-Performance Computing

They'll Verify. They Just Won't Act. How Authority Framing and Laundered Code Turn a Trusted Agentic CI/CD Pip

Data Leakage Prevention in Agentic Applications via Preemptive Hardening

Measuring and Evaluating the Performance of Generative AI Models for Scam Detection

A Systematic Evaluation of Traditional Privacy Policy Analysis Tools Against LLMs

How Do You Choose Your AI Component? An Interview Study of Secure AI Integration in Practice

SlotGuard: Stop Oversharing Private Local Context in LLM Agent Transcri

Do Agents Dream of False Memories? Black-box Visual Attacks on Long-term Memory in Multimodal AI Agents

oapi-codegen: OpenAPI Server Description Escapes Generated Go Comment and Injects Executable Code

Context Contamination in LLM Analysis of Network Security Logs: Poison with Passive Prompt Injection and Mitig

Routing Ceilings Are Domain-Independent: Structural Prior Injection in Code Security Vulnerability Detection

Beyond Success Rate: Cost-Aware Evaluation of Offensive and Defensive Security Agents

FlowGuard: From Signals to Evidence for MCP Security Detection

Bad Memory: Evaluating Prompt Injection Risks from Memory in Agentic Systems

xChk: Bring Your Own Identity -- Heterogeneous Assurance with Verifier-Determined Sufficiency

DREA: Decoupled Reasoning and Exploration Agents for Repository-Level Vulnerability Detection

ToolHive: SSRF in remote MCP server authentication discovery (host-side, bypasses container isolation)

LangBot: Authenticated RCE Via MCP Configuration

Antiproof: Synthesizing Vulnerability Detectors and Proofs of Exploitability

Bulkhead: Automated Semantic Detection and Remediation of Container Escape Vulnerabilities

yutu: Arbitrary File Write via MCP `caption-download` Tool

n8n-MCP: Cross-tenant access to workflow version backups in multi-tenant HTTP deployments

Understanding Implicit Trust Errors in Core Carrier Networks through Multi-Agent Flaw Discovery and Analysis

Devil in the Lens: Analyzing and Defending Physical Prompt Injection Against Vision-Language Models on Wearabl

NetInjectBench: Benchmarking Indirect Prompt Injection in Tool-Using Large Language Model Agents for Network O

Agent Hacks Agent: Autoresearch for Production-Agent Red-Teaming

Rethinking MCP Security: A Large-Scale Study of Runtime MCP Servers and Security Scanner Reliability

A Study of Microsoft's Early 2026 Rollout of Claude Code and GitHub Copilot CLI

Proof-of-Continuity: A Temporal Model for Authority Propagation in Distributed Systems and AI Agents

Secret Scanner Agent: Extracting Secrets and Access Context from Unstructured Documents

VEXAIoT: Autonomous IoT Vulnerability EXploitation using AI Agents

mcp-atlassian: Arbitrary file read via missing path validation in confluence_upload_attachment

Multimodal Unlearning Across Vision, Language, Video, and Audio: Survey of Methods, Datasets, and Benchmarks

Open Models, Open Risks: Measuring Unsafe Generation in Text-to-Image Models In the Wild

From Legacy Documentation to OSCAL: An MCP-Based Agent Pipeline for Threat-Informed Continuous Compliance in C

Securing Autonomous Vehicle Systems via Twin-Aware Federated Reinforcement Learning

Multi-Agent Firewall Architecture for Privacy Protection of Sensitive Data in Interactions with Language Model

Beware of Agentic Botnets: Scalable Untargeted Promptware Attacks via Universal and Transferable Adversarial H

Large Language Models (LLMs) and Generative AI in Cybersecurity and Privacy: A Survey of Dual-Use Risks, AI-Ge

Mitigating Taint-Style Vulnerabilities in MCP Servers via Security-Aware Tool Descriptions

Serena: Unauthenticated Flask dashboard on fixed port enables DNS rebinding → memory poisoning → RCE

Beyond Refusal: A Same-Lineage Study of Aligned and Abliterated LLMs for Vulnerability Analysis

Unicode TAG-Block Concealment of Tool-Metadata Payloads in the Model Context Protocol: An Approval-View Fideli

The Balkanization of Execution-Security Research for AI Coding Agents: Isolation, Access Control, and Time-of-

@aborruso/ckan-mcp-server: SSRF via base_url allows access to internal networks (Potential fix bypass of CVE-2

Governed Individuation: Cryptographically Decoupling an Agent's Learning from Its Authority

RustMizan: A Compilable, Contamination-Aware Benchmarking Framework for Rust Vulnerabilities

DualView: Preventing Indirect Prompt Injection in Personal AI Agents

Your Agent's Memories Are Not Its Own: Forged Reasoning Attacks on LLM Agent Memory and Defenses

Cloak and Detonate: Scanner Evasion and Dynamic Detection of Agent Skill Malware

JavaVulBench: A Java Vulnerability Benchmark with Realistic Splits, a Unified Multi-Backend Harness, and a Lea

Langroid: Neo4jChatAgent executes LLM-generated Cypher without validation (prompt-to-Cypher injection; config-

The Safari MCP server for web developers

Pmeta-TLA: Backdoor Attacks for Speech Classification Models via Meta-Learning with Timbre Leakage Attack

Knowledge Over Parameters: Evolving Smart Contract Vulnerability Detection

Behind the Refusal: Determining Guardrail Activation via Behavioral Monitoring

Cloak and Detonate: Scanner Evasion and Dynamic Detection of Agent Skill Malware

Grackle: Fail-open authorization in the MCP tool layer lets scoped agents perform cross-task and cross-session

Antaeus: Hunting Repository-Level Logic Vulnerabilities via Context-Grounded LLM Reasoning

Beyond the Prompt: Jailbreaking Function-Calling LLMs via Simulated Moderation Traces

repomix: attach_packed_output can bypass file-read secret scanning for supported local files

Cortex has Untrusted Project Bootstrap Code Execution via `CLAUDE_PROJECT_DIR`

@jshookmcp/jshook: ICMP probe and traceroute skip local-network SSRF authorization

An AI-Based Solution for Secure Service Provisioning in IoT

AI-Generated PowerShell Malware: An Experimental Framework and Dataset

Securing the AI Agent: A Unified Framework for Multi-Layer Agent Red Teaming

Understanding and Evaluating Claw-like Agent Security Through a Computer-Systems Lens

Words Speak Louder Than Code: Investigating Cognitive Heuristics in LLM-Based Code Vulnerability Detection

Forensic Trajectory Signatures for Agent Memory Poisoning Detection

From Tool Connection to Execution Control: Benchmarking Security Invariants in MCP-Style Agent Runtimes

FlipGuard: Defending Large Language Models Against Quantization-Conditioned Backdoor Attacks

A Red-Team Study of Anthropic Fable 5 & Opus 4.8 Models

Smarter Saboteurs, Better Fixers: Scaling & Security in Linear Multi-Agent Workflows