@inproceedings{zhang-etal-2026-e2a,
title = "e2a: A Secure Agent-Native Email Gateway with Phishing and Prompt-Injection Defense",
author = "Zhang, Josh and
Jia, Hongyuan and
Zhang, Xueyan and
Tan, Hanzhi and
Hu, Jingzhuo",
editor = "Agrawal, Sweta and
Abad, Alberto",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-demos.19/",
pages = "199--207",
ISBN = "979-8-89176-461-3",
abstract = "Email is becoming the universal interface for AI agents, but an ingested message contains un- trusted, malicious data the agent may execute as instructions. The indirect prompt-injection failure underlies real-world incidents such as EchoLeak (M365 Copilot, CVE-2025-32711) and the Gemini-in-Gmail hidden-HTML hijack. We present e2a, a deployable, open-source authenticated-email gateway that gives an agent a verified email identity and exposes two agent- agnostic defense hook points. The (G)ateway authenticates senders via SPF/DKIM/DMARC and parses MIME structure; the (L)LM-layer screens the normalized prompt for injection. On top of the gateway, we construct a bench- mark that isolates email structure as the first- class injection surface: one payload rendered across eight transport surfaces. We measure the per-surface evasion delta for open-source classifiers, three commercial security APIs, and frontier LLM judges. We find that structural hiding defeats detection. A naive text/plain- first parse costs every detector substantial AUC, dropping most classifiers below chance, and even on the gateway{'}s structure-aware view no non-LLM detector survives. A frontier LLM judge on the same view recovers most of the loss, and conditioning the threshold on crypto- graphic sender authentication (DMARC) lifts held-out recall further. No single layer suf- fices. e2a is live and open-source (Apache-2.0) at https://e2a.dev, released with the bench- mark and harness."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-e2a">
<titleInfo>
<title>e2a: A Secure Agent-Native Email Gateway with Phishing and Prompt-Injection Defense</title>
</titleInfo>
<name type="personal">
<namePart type="given">Josh</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongyuan</namePart>
<namePart type="family">Jia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xueyan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hanzhi</namePart>
<namePart type="family">Tan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jingzhuo</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sweta</namePart>
<namePart type="family">Agrawal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alberto</namePart>
<namePart type="family">Abad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-461-3</identifier>
</relatedItem>
<abstract>Email is becoming the universal interface for AI agents, but an ingested message contains un- trusted, malicious data the agent may execute as instructions. The indirect prompt-injection failure underlies real-world incidents such as EchoLeak (M365 Copilot, CVE-2025-32711) and the Gemini-in-Gmail hidden-HTML hijack. We present e2a, a deployable, open-source authenticated-email gateway that gives an agent a verified email identity and exposes two agent- agnostic defense hook points. The (G)ateway authenticates senders via SPF/DKIM/DMARC and parses MIME structure; the (L)LM-layer screens the normalized prompt for injection. On top of the gateway, we construct a bench- mark that isolates email structure as the first- class injection surface: one payload rendered across eight transport surfaces. We measure the per-surface evasion delta for open-source classifiers, three commercial security APIs, and frontier LLM judges. We find that structural hiding defeats detection. A naive text/plain- first parse costs every detector substantial AUC, dropping most classifiers below chance, and even on the gateway’s structure-aware view no non-LLM detector survives. A frontier LLM judge on the same view recovers most of the loss, and conditioning the threshold on crypto- graphic sender authentication (DMARC) lifts held-out recall further. No single layer suf- fices. e2a is live and open-source (Apache-2.0) at https://e2a.dev, released with the bench- mark and harness.</abstract>
<identifier type="citekey">zhang-etal-2026-e2a</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-demos.19/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>199</start>
<end>207</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T e2a: A Secure Agent-Native Email Gateway with Phishing and Prompt-Injection Defense
%A Zhang, Josh
%A Jia, Hongyuan
%A Zhang, Xueyan
%A Tan, Hanzhi
%A Hu, Jingzhuo
%Y Agrawal, Sweta
%Y Abad, Alberto
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-461-3
%F zhang-etal-2026-e2a
%X Email is becoming the universal interface for AI agents, but an ingested message contains un- trusted, malicious data the agent may execute as instructions. The indirect prompt-injection failure underlies real-world incidents such as EchoLeak (M365 Copilot, CVE-2025-32711) and the Gemini-in-Gmail hidden-HTML hijack. We present e2a, a deployable, open-source authenticated-email gateway that gives an agent a verified email identity and exposes two agent- agnostic defense hook points. The (G)ateway authenticates senders via SPF/DKIM/DMARC and parses MIME structure; the (L)LM-layer screens the normalized prompt for injection. On top of the gateway, we construct a bench- mark that isolates email structure as the first- class injection surface: one payload rendered across eight transport surfaces. We measure the per-surface evasion delta for open-source classifiers, three commercial security APIs, and frontier LLM judges. We find that structural hiding defeats detection. A naive text/plain- first parse costs every detector substantial AUC, dropping most classifiers below chance, and even on the gateway’s structure-aware view no non-LLM detector survives. A frontier LLM judge on the same view recovers most of the loss, and conditioning the threshold on crypto- graphic sender authentication (DMARC) lifts held-out recall further. No single layer suf- fices. e2a is live and open-source (Apache-2.0) at https://e2a.dev, released with the bench- mark and harness.
%U https://aclanthology.org/2026.emnlp-demos.19/
%P 199-207
Markdown (Informal)
[e2a: A Secure Agent-Native Email Gateway with Phishing and Prompt-Injection Defense](https://aclanthology.org/2026.emnlp-demos.19/) (Zhang et al., EMNLP 2026)
ACL