SourcesTAPE Tool-Guided Adaptive Planning for Robust AgentsarXiv·medium signalXBlueskyLinkedInCopy linkGraph-based multi-plan aggregation with constrained decoding and adaptive re-planning. For agents in costly real-world environments.SourceSource pagearXiv↳ Follow the threadPolicy dependency / Stack layerSCA-Agent reconstructs a dependency's trace across Code, Build, Release, Deploy and Runtime and beats the best traditional SCA tool by 18.76 points of F1arXiv 2609.18391Stack layer / Update threadTwo Tool-Level Defenses Drive Prompt Injection and Memory Poisoning to 0% Attack Success in Many SettingsarXiv 2609.16098Threat pattern / ContrastTracing C/C++ Reuse Through Intermediate Projects Found 48 Unpatched Vulnerabilities, 23 Now FixedarXiv 2609.17018Threat patternThreat taxonomy for autonomous pentest agents argues conversational guardrails do not transfer to agents with persistent memory and real-world actionsarXivPolicy dependency / Stack layerConverting GUI Trajectories Into Replayable MCP-Style Calls Instead of Unstructured MemoriesarXiv 2609.16635Stack layer / Threat pattern37,623 provenance-labeled agent PRs: Codex code was reverted half as often as human code, Devin's 31% more, and Claude Code PRs waited 12.6 hours for first reviewarXiv 2609.17598Stack layer / ContrastAFL++ fuzzed agent-written reimplementations of ten Linux utilities: fewer memory errors than the shipped versions, but more infinite loopsarXiv 2609.18298Stack layer / ContrastEmergence World ran 10 agents per world for 16 days and found no frontier model contained an injected attack — one acted on poisoned memory 46 hours laterarXiv