ProofPoint Email Protection Evasion via Surrogate Model

ProofPoint Email Protection

Researchers built a surrogate machine learning model to study and bypass Proofpoint's email protection system. The insights allowed them to craft malicious emails that evaded spam detection.

CVSS: 3.7
EPSS: 0.02 · top 24% · 2026-11-13
EPSS scores by FIRST
first report: 23 Oct 2020
last update: 31 Mar 2023
CVE ids: CVE-2019-20634 other records

Timeline · 6

Sources · 6

Advisories and databases

2020-10-23
National Vulnerability Database entry for CVE-2019-20634
NVD · nvd.nist.gov/vuln/detail/CVE-2019-20634

News and write-ups

2020-10-23
2019 DerbyCon presentation "42: The answer to life, the universe, and everything offensive security"
GitHub · github.com/moohax/Talks/blob/master/slides/DerbyCon19.pdf
2020-10-23
Proof Pudding (CVE-2019-20634) Implementation on GitHub
GitHub · github.com/moohax/Proof-Pudding
2020-10-23
2019 DerbyCon video presentation "42: The answer to life, the universe, and everything offensive security"
youtube.com · youtube.com/watch?ab-channel=AdrianCrenshaw&v=CsvkYoxtexQ

Research and catalogues

2020-10-23
ProofPoint Evasion
MITRE ATLAS · atlas.mitre.org/studies/AML.CS0008
2023-03-31
AVID-2023-V009: Proof Pudding (CVE-2019-20634) is a code repository that describes how ML researchers evaded ProofPoint's email protection system by first building a copy-cat email protection ML model, and using the insights to bypass the live system.
AVID · github.com/avidml/avid-db/blob/main/vulnerabilities/2023/AVID-2023-V009.json

Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.

Classification

lane: Research
AI role: Attack on AI
categories: Adversarial evasion
severity: 1 · Research / proof of concept
targets: Enterprise deployment
owasp llm top 10: no entry for these categories
mitre atlas: AML.T0015 Evade AI Model AML.T0043 Craft Adversarial Data
tags: evasion adversarial-ml email-security
dates: happened 9 Sep 2019 · first reported 23 Oct 2020
record: confidence 0.95 · created 2026-10-18 06:55Z · updated 2026-10-18 18:20Z

Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned

Copy a brief

This record as a short Markdown brief, for a ticket, a chat or a coding agent.

the brief as text

# ProofPoint Email Protection Evasion via Surrogate Model

https://gcdradar.vercel.app/incidents/proofpoint-email-protection-evasion-via-surrogate-model · AIR-2020-0025 · closed 2020-11-06

- Lane: Research
- Severity: 1 (Research / proof of concept)
- Fix status: Fix status unknown
- First reported: 2020-10-23 (happened 2019-09-09)
- AI role: Attack on AI
- Products: ProofPoint Email Protection
- CVE ids: CVE-2019-20634
- CVSS: 3.7

## Summary

Researchers built a surrogate machine learning model to study and bypass Proofpoint's email protection system. The insights allowed them to craft malicious emails that evaded spam detection.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2023-03-31: AVID-2023-V009: Proof Pudding (CVE-2019-20634) is a code repository that describes how ML researchers evaded ProofPoint's email protection system by first building a copy-cat email protection ML model, and using the insights to bypass the … (https://github.com/avidml/avid-db/blob/main/vulnerabilities/2023/AVID-2023-V009.json)
- 2020-10-23: 2019 DerbyCon video presentation "42: The answer to life, the universe, and everything offensive security" (https://youtube.com/watch?ab-channel=AdrianCrenshaw&v=CsvkYoxtexQ)
- 2020-10-23: Proof Pudding (CVE-2019-20634) Implementation on GitHub (https://github.com/moohax/Proof-Pudding)
- 2020-10-23: 2019 DerbyCon presentation "42: The answer to life, the universe, and everything offensive security" (https://github.com/moohax/Talks/blob/master/slides/DerbyCon19.pdf)
- 2020-10-23: National Vulnerability Database entry for CVE-2019-20634 (https://nvd.nist.gov/vuln/detail/CVE-2019-20634)

## Sources

- 2020-10-23 MITRE ATLAS: ProofPoint Evasion https://atlas.mitre.org/studies/AML.CS0008
- 2020-10-23 NVD: National Vulnerability Database entry for CVE-2019-20634 https://nvd.nist.gov/vuln/detail/CVE-2019-20634
- 2020-10-23 GitHub: 2019 DerbyCon presentation "42: The answer to life, the universe, and everything offensive security" https://github.com/moohax/Talks/blob/master/slides/DerbyCon19.pdf
- 2020-10-23 GitHub: Proof Pudding (CVE-2019-20634) Implementation on GitHub https://github.com/moohax/Proof-Pudding
- 2020-10-23 youtube.com: 2019 DerbyCon video presentation "42: The answer to life, the universe, and everything offensive security" https://youtube.com/watch?ab-channel=AdrianCrenshaw&v=CsvkYoxtexQ
- 2023-03-31 AVID: AVID-2023-V009: Proof Pudding (CVE-2019-20634) is a code repository that describes how ML researchers evaded ProofPoint's email protection system by first building a copy-cat email protection ML model, and using the insights to bypass the live system. https://github.com/avidml/avid-db/blob/main/vulnerabilities/2023/AVID-2023-V009.json

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.

closed
Edge AI Product Microsoft Edge AI Product Evasion Red Team Exercise
Research · since Feb 2020 (approx.)
AI
closed
Azure Microsoft Internal Azure AI Service Red Team Exercise
Research · since Jan 2020 (approx.)
AI
closed
Commercial Face Identification Service Face Identification System Evasion via Physical Countermeasures
Research · since 1 Jan 2020
AI
closed
DGA detector Botnet DGA Detection Evasion Using Domain Mutation
Research · since 1 Jan 2020
AI
closed
Palo Alto Networks malware detection system Evasion of Deep Learning Detector for Malware C&C Traffic
Research · since 1 Jan 2020
AI
closed
CylancePROTECT Bypassing Cylance AI Malware Detection via Universal String
Research · since 7 Sep 2019
AI