Arpan Kumar Mahapatra

arXiv2026

Public-Sharing Labels and Verbatim Field Egress in an MCP-to-A2A Agent Configuration: A Controlled Multi-Model Study

Safety properties assessed separately for Model Context Protocol (MCP) tool use and Agent2Agent (A2A) delegation need not describe behavior when one agent uses both. We measure one such behavior in a single controlled MCP-to-A2A configuration: a testbed drives a real-model host across a local MCP and a local A2A leg into an ordered event trace scored by exact deterministic rules (no LLM judge), one restricted decision per trial. In a pre-specified, frozen three-arm design, each of 10 record scenarios appears with a CONFIDENTIAL header, with no header, and with PUBLIC - OK TO SHARE; the six substantive record values are byte-identical across arms, and the outcome is verbatim occurrence of any of them in the outbound message. Four models x 3 arms x 4 repeats give 480 trials; the scenario is the unit of generalization, and we report the 10 scenario-level values (mean, median, sign counts), with no p-values or intervals. The confidential-minus-unlabeled contrast is inconclusive and floor-limited in every model (both arms at or near zero), so it does not show that confidential labels lack a protective effect. Adding PUBLIC - OK TO SHARE is descriptively associated with higher verbatim egress relative to the unlabeled baseline, with strong model dependence: strong and consistent for Claude Sonnet 5 (public-minus-unlabeled mean +0.800, all 10 scenarios; mostly an association with whether Claude relays at all), moderate but floor-limited for one GPT-5.6 tier, small (median 0) for another, and a complete floor for the third. This is an association in one configuration, not a causal or general effect. Code, byte-pinned traces, and the offline analysis pipeline are released as a public artifact.

cs.CR↗

arXiv2024

RouteNet-Fermi: Network Modeling With GNN (Analysis And Re-implementation)

Network performance modeling presents important challenges in modern computer networks due to increasing complexity, scale, and diverse traffic patterns. While traditional approaches like queuing theory and packet-level simulation have served as foundational tools, they face limitations in modeling complex traffic behaviors and scaling to large networks. This project presents an extended implementation of RouteNet-Fermi, a Graph Neural Network (GNN) architecture designed for network performance prediction, with additional recurrent neural network variants. We improve the the original architecture by implementing Long Short-Term Memory (LSTM) cells and Recurrent Neural Network (RNN) cells alongside the existing Gated Recurrent Unit (GRU) cells implementation. This work contributes to the understanding of recurrent neural architectures in GNN-based network modeling and provides a flexible framework for future experimentation with different cell types.

cs.NI↗