# Judging Before Propagation: Task-Graph Contextual Integrity for Multi-Agent Privacy

DevFeed: [Judging Before Propagation: Task-Graph Contextual Integrity for Multi-Agent Privacy](<https://devfeed.tech/articles/judging-before-propagation-task-graph-contextual-integrity-for-multi-agent-privacy-108207.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.39532>)

Author: Hengyu Wu, Blaise Delattre, Yang Cao

Published: 2026-10-30T04:00:00Z

Content type: news

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [Security & Privacy](<https://devfeed.tech/topics/security-privacy.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [agent to agent communication](<https://devfeed.tech/topics/agent-to-agent-communication.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-privacy](<https://devfeed.tech/tags/agent-privacy.md>), [authorization](<https://devfeed.tech/tags/authorization.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [disclosure](<https://devfeed.tech/tags/disclosure.md>), [graph](<https://devfeed.tech/tags/graph.md>), [integrity](<https://devfeed.tech/tags/integrity.md>), [multi-agent](<https://devfeed.tech/tags/multi-agent.md>), [privacy](<https://devfeed.tech/tags/privacy.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

## AI overview

Task-Graph Contextual Integrity (TG-CI) authorizes information releases by considering both an item's intended downstream scope and recipients' evolving disclosure histories across a multi-agent task graph. This addresses a gap in local Contextual Integrity checks, which can approve a disclosure at one boundary without accounting for later propagation. Full-Scope requires admissibility across the complete downstream scope; Auto-Region confines an item to its maximal admissible region using frontier gates. On MAPLE-Bench, a 144-case benchmark spanning four scenarios, the primary configuration reduced mean unauthorized disclosures by 85.8-94.1% versus the Local-History CI baseline while retaining 93.6-96.1% of its task accuracy. The authors report complementary privacy, cost, and utility trade-offs between the policies.

## Source excerpt

arXiv:2610.39532v1 Announce Type: cross Abstract: Multi-agent LLM systems increasingly rely on persistent memory and inter-agent communication. These capabilities create privacy risks that extend beyond individual interactions: a release appropriate in isolation may become inappropriate when combined with information previously disclosed to a recipient or propagated to downstream agents along the planned workflow. Existing Contextual Integrity (CI) mechanisms for agentic systems typically rely on local judgments that assess each disclosure in isolation, with or without regard to the recipient's history, but do not account for how information propagates downstream. However, compliance at the current boundary does not guarantee appropriateness across the workflow. We introduce Task-Graph Contextual Integrity (TG-CI), to our knowledge the first task-graph-native formalization of Contextual Integrity for multi-agent systems. TG-CI represents a workflow as a Multi-Agent Task Graph (MATG) and extends CI from atomic transfers to stateful, graph-level information flows by jointly considering an item's intended propagation scope and recipients' evolving disclosure states. We operationalize TG-CI through propagation-scoped authorization under two complementary policies: Full-Scope, which requires admissibility across the complete downstream scope, and Auto-Region, which confines each item to its maximal admissible region using lightweight frontier gates. We further introduce MAPLE-Bench, a benchmark of 144 cases across four scenarios that jointly vary MATG topology, disclosure history, compromised-agent placement, and item-level privacy policies. In our primary experimental configuration, TG-CI reduces the mean unauthorized disclosure count by 85.8-94.1% relative to the Local-History CI baseline, while retaining 93.6-96.1% of the baseline's task accuracy, with the two enforcement policies providing complementary privacy--cost--utility trade-offs.