ResearchSleeper Cell Temporal Backdoors in Tool-Using LLMs via SFT-GRPOarXiv·high signalPoisoned models pass all benchmarks while harboring trigger-activated malicious tool calls. Supply-chain risk for LoRA adopters.SourceSource pagearXiv↳ Follow the threadStack layer / Threat patternMitigating Taint-Style MCP Server Vulnerabilities Without Touching Server Code (SPELLSMITH)arXivPolicy dependency / Stack layerReason Less, Verify More: Deterministic Gates Catch a Silent Policy-Violation Failure in Tool-Using AgentsarXivStack layer / ContrastLong-Horizon-Terminal-Bench: Best Frontier Agent Clears Just 15% of Real Terminal Tasks, Mean 4.3%arXiv (Tencent Hunyuan) / HuggingFace Daily PapersThreat patternMalaika Grounds Agentic Malware Analysis in Domain, Program Semantics, and Threat IntelarXivPolicy dependency / Stack layerCodenotary ships AgentMon 3 for agent runtime policy and compliance loggingAI Agent StoreStack layerCache and reuse agent-generated tools instead of regenerating code every request (arXiv 2607.08010, July 9)arXiv 2607.08010Update thread / Follow-up threadEvoAgentBench: A Benchmark That Isolates Whether Agents Can Turn Past Sessions Into Reusable ProceduresarXivThreat patternDual-Graph Verification Framework Audits Legacy IT Security Concepts for NIS-2 / OSCAL MigrationarXiv