The Cognitive Firewall:Securing Browser Based AI Agents Against Indirect Prompt Injection Via Hybrid Edge Cloud Defense

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lan, Qianlong, Kaul, Anuj
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918407555776512
author Lan, Qianlong
Kaul, Anuj
author_facet Lan, Qianlong
Kaul, Anuj
contents Deploying large language models (LLMs) as autonomous browser agents exposes a significant attack surface in the form of Indirect Prompt Injection (IPI). Cloud-based defenses can provide strong semantic analysis, but they introduce latency and raise privacy concerns. We present the Cognitive Firewall, a three-stage split-compute architecture that distributes security checks across the client and the cloud. The system consists of a local visual Sentinel, a cloud-based Deep Planner, and a deterministic Guard that enforces execution-time policies. Across 1,000 adversarial samples, edge-only defenses fail to detect 86.9% of semantic attacks. In contrast, the full hybrid architecture reduces the overall attack success rate (ASR) to below 1% (0.88% under static evaluation and 0.67% under adaptive evaluation), while maintaining deterministic constraints on side-effecting actions. By filtering presentation-layer attacks locally, the system avoids unnecessary cloud inference and achieves an approximately 17,000x latency advantage over cloud-only baselines. These results indicate that deterministic enforcement at the execution boundary can complement probabilistic language models, and that split-compute provides a practical foundation for securing interactive LLM agents.
format Preprint
id arxiv_https___arxiv_org_abs_2603_23791
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle The Cognitive Firewall:Securing Browser Based AI Agents Against Indirect Prompt Injection Via Hybrid Edge Cloud Defense
Lan, Qianlong
Kaul, Anuj
Cryptography and Security
Artificial Intelligence
Deploying large language models (LLMs) as autonomous browser agents exposes a significant attack surface in the form of Indirect Prompt Injection (IPI). Cloud-based defenses can provide strong semantic analysis, but they introduce latency and raise privacy concerns. We present the Cognitive Firewall, a three-stage split-compute architecture that distributes security checks across the client and the cloud. The system consists of a local visual Sentinel, a cloud-based Deep Planner, and a deterministic Guard that enforces execution-time policies. Across 1,000 adversarial samples, edge-only defenses fail to detect 86.9% of semantic attacks. In contrast, the full hybrid architecture reduces the overall attack success rate (ASR) to below 1% (0.88% under static evaluation and 0.67% under adaptive evaluation), while maintaining deterministic constraints on side-effecting actions. By filtering presentation-layer attacks locally, the system avoids unnecessary cloud inference and achieves an approximately 17,000x latency advantage over cloud-only baselines. These results indicate that deterministic enforcement at the execution boundary can complement probabilistic language models, and that split-compute provides a practical foundation for securing interactive LLM agents.
title The Cognitive Firewall:Securing Browser Based AI Agents Against Indirect Prompt Injection Via Hybrid Edge Cloud Defense
topic Cryptography and Security
Artificial Intelligence
url https://arxiv.org/abs/2603.23791