SourcesPyVision-RL Solving Agent Interaction CollapsearXiv·medium signalAddresses RL-trained agents learning to reduce tool usage. Oversampling-filtering-ranking rollout strategy sustains interaction.SourceSource pagearXiv↳ Follow the threadStack layer'Who Broke the System?' AgentLocate Pinpoints the Failing Agent and Step in Multi-Agent RunsarXivPolicy dependency / Stack layerReason Less, Verify More: Deterministic Gates Catch a Silent Policy-Violation Failure in Tool-Using AgentsarXivStack layer / Threat patternMitigating Taint-Style MCP Server Vulnerabilities Without Touching Server Code (SPELLSMITH)arXivStack layer / Update threadSLBench Finds Up to 70% Unsafe Execution When Codex and Claude Code Follow Logical Relations in Skill FilesarXivSame source domain / Semantic neighborS²AE Makes Sparse Autoencoders Learn Modality-Consistent Concepts in VLMsarXivPolicy dependency / ContrastMake context compaction a trained policy, not an inference-time heuristic — CompactionRL lifts SWE-bench up to +7 pointsarXiv 2607.05378Stack layer / ContrastLong-Horizon-Terminal-Bench: Best Frontier Agent Clears Just 15% of Real Terminal Tasks, Mean 4.3%arXiv (Tencent Hunyuan) / HuggingFace Daily PapersStack layer / Update threadDominoTree Improves Speculative Decoding With Path-Dependent Tree DraftingarXiv