ResearchTasksLLM Agent DefenseFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedAgentDojo OverallRepeat Prompt84.54Clean Utility12Feb 26, 2026AgentDojo SlackNo Defense80.95Clean Utility12Feb 26, 2026AgentDojo WorkspaceTask Shield85Clean Utility12Feb 26, 2026