The study evaluates safety properties of a controlled MCP-to-A2A agent configuration by measuring verbatim field egress across ten record scenarios under three labeling conditions (CONFIDENTIAL, no header, PUBLIC – OK TO SHARE). Using four models repeated four times each, 480 trials were conducted, and the results show that adding a PUBLIC header is descriptively linked to higher verbatim egress, with the effect varying strongly by model. The study releases code, byte‑pinned traces, and an offline analysis pipeline as a public artifact.
arXiv:2608.29028v1 Announce Type: new Abstract: Multi-agent LLM systems often coordinate by compressing an upstream interaction into a handoff artifact that downstream agents treat as shared state. W...
arXiv:2605. 27784v2 Announce Type: replace Abstract: LLM agents are governed by long-lived prompt policies, where individually reasonable stand- ing rules can jointly govern the same pre- generation state.
The paper evaluates CASCADE, a fully local layered defense for Model Context Protocol (MCP)-based systems, by conducting a component ablation and corpus audit on a fixed 5,000-sample dataset. It demonstrates that the choice of aggregation convention heavily influences reported metrics, that detection performance varies with provenance, and that the released configuration does not fully disclose the operating point. The study also shows that a local review model invoked for a third of requests does not alter classification outcomes, highlighting the importance of reproducibility and transparency in defense evaluations.
BodhiPromptShield is a policy‑aware mediation layer for LLM agent pipelines that detects sensitive text spans before they propagate, replacing them with typed placeholders, semantic abstractions, or secure tokens and restoring them only at authorized execution boundaries. In evaluations on AI4Privacy, PrivacyLens, and AgentDojo datasets, the system reduces identifier exposure to 7.4% and 1.8% respectively, and limits exact identifier leakage in final actions to 2.1–3.1%. While mediation preserves factual content according to automated metrics, human annotations show a significant drop in inferability from 100% to 24–53%, indicating the need for human validation of semantic‑leakage measures.
The paper investigates a critical flaw in AI coding-agent systems such as Claude Code, Codex CLI, and Cursor, where the action approved by a human is not the same as the action executed by the harness. It introduces the concept of Approval Laundering, categorizing six systematic failure modes—Scope, Argument, Temporal, Tool, Delegation, and Semantic laundering—and demonstrates these failures through controlled experiments. The authors propose an Approval Token mechanism that mitigates some laundering types but leaves others unaffected, highlighting the limitations of current enforcement strategies.
arXiv:2608. 02665v1 Announce Type: cross Abstract: A benchmark score is a measurement instrument, yet most benchmarks read each item at a single canonical surface form.
The paper introduces Task-Graph Contextual Integrity (TG‑CI), a formalism that extends traditional Contextual Integrity to multi‑agent systems by modeling workflows as Multi‑Agent Task Graphs (MATGs). TG‑CI evaluates privacy not just at individual disclosures but across the entire downstream propagation, offering two enforcement policies—Full‑Scope and Auto‑Region—that balance privacy protection with task performance. A new benchmark, MAPLE‑Bench, demonstrates that TG‑CI can cut unauthorized disclosures by up to 94% while preserving over 94% of task accuracy compared to existing local‑history CI methods.
arXiv:2608. 04618v1 Announce Type: new Abstract: Heterogeneous language-model ensembles expand the space of candidate responses, yet they lack a principled criterion for when a newly generated answer should supersede an already supported one.
The paper investigates how large language model agents can inadvertently leak confidential information by combining data from multiple documents—a phenomenon termed the "mosaic effect." Using 30 synthetic cybersecurity, medical, and military scenarios, the authors generate 10,800 simulated and 1,440 live answers from three LLM agents, finding that even the best model discloses forbidden composites in over 7% of simulated responses. While visible classification headers and aggregation rules reduce disclosure rates, unsafe disclosures remain in roughly one third of simulated answers, and increased refusals do not fully prevent leakage of prohibited conclusions.
Modern AI coding-agent harnesses (Claude Code, Codex CLI, Cursor) rest their security boundary on a largely unexamined assumption: that the action A a human approves is the same action A' the harness...
arXiv:2606. 07150v1 Announce Type: cross Abstract: Agent-interoperability protocols such as A2A and MCP standardize what agents say to one another, but assume address-based transport over HTTP(S).
The paper presents a framework for causal attribution in agentic AI systems, outlining estimators and conditions where they fail. It distinguishes between marginal total effects and common‑random‑number total effects, introduces a natural direct effect under pinned downstreams, and derives a coupling method to keep direct effects estimable. The authors also propose a traceability specification to meet upcoming regulatory requirements for high‑risk AI systems.