Threat reportPhishingTL-2026-2430
Blob URL Phishing: Fake Login Pages Rendered Entirely Inside the Browser via Microsoft OAuth and Teams
Barracuda researchers identified a phishing campaign that routes victims through genuine Microsoft OAuth and Teams infrastructure before assembling a fake login page inside the victim's browser using a blob URL rather than hosting the phishing content on a server.
Key facts
| Identifiers | TL-2026-2430 |
|---|---|
| Severity | MEDIUM |
| Affected products | Microsoft Microsoft 365 / Entra ID OAuth authorization flow and Microsoft Teams redirect handling (cloud service; abused as a trust-transfer mechanism, not a versioned vulnerability) |
| Tools | Unnamed managed phishing-as-a-service (PhaaS) backend |
| MITRE ATT&CK techniques | T1027 Obfuscated Files or Information, T1036 Masquerading, T1056 Input Capture, T1059 Command and Scripting Interpreter, T1071 Application Layer Protocol, T1204 User Execution, T1566 Phishing, T1583 Acquire Infrastructure, T1608 Stage Capabilities, T1684.001 Impersonation |
| Motivation | UNKNOWN |
| Detection rules | 9 (1 shown free) |
| Sources | 8 cited sources |
Detection coverage
As of 2026-09-10, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-2430 across Splunk SPL, Microsoft KQL and Sigma, covering 8 indicator(s) of compromise. The whole corpus is readable without an account. Every threat comes with one complete detection in Splunk SPL, Microsoft KQL and Sigma, free for everyone; a free account adds daily TL Creds to unlock the rest. Paid plans add a larger daily allowance, correlation and the MCP server. Threadlinqs MCP server · View plans.
Free detection: DocuSign-Themed Calendar Invite With Crafted Microsoft OAuth Redirect (SPL)
Detects an email impersonating DocuSign that carries a .ics calendar-invite attachment whose embedded link abuses the legitimate Microsoft OAuth authorization endpoint (login.microsoftonline.com) via a crafted redirect_uri parameter pointing outside sanctioned Teams/SharePoint destinations. This is the initial-access lure step of the Barracuda-documented blob URL phishing chain, where a genuine OAuth endpoint is used to transfer trust before the victim is routed to an attacker-influenced resource.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1583.001, T1608.005, T1566.001, T1204.001
- Data sources: o365_message_trace, email_gateway_logs, m365_defender_emailevents, m365_defender_urlclickevents, m365_defender_emailattachmentinfo
- Severity: high
- Confidence: medium
Splunk SPL
index=email_security (sourcetype="o365:management:activity" OR sourcetype="proofpoint:messages" OR sourcetype="mimecast:message")
attachment_extension=".ics"
| eval sender_impersonates_docusign=if(match(lower(from_display_name), "docusign") AND NOT match(lower(from_domain), "docusign\.(com|net)$"), 1, 0)
| where sender_impersonates_docusign=1
| rex field=body_urls "redirect_uri=(?<redirect_target>[^&\s\"]+)"
| where like(url, "%login.microsoftonline.com%") AND isnotnull(redirect_target)
| eval redirect_target_decoded=urldecode(redirect_target)
| where NOT match(redirect_target_decoded, "^https://(teams\.microsoft\.com|.*\.sharepoint\.com)/")
| table _time, from_address, from_display_name, to_address, subject, attachment_name, url, redirect_target_decoded, message_id
| sort - _timeMicrosoft KQL
EmailAttachmentInfo
| where FileType =~ "ics"
| join kind=inner (
EmailEvents
| where SenderDisplayName has "DocuSign" and SenderFromDomain !endswith "docusign.com" and SenderFromDomain !endswith "docusign.net"
| where Timestamp > ago(24h)
) on NetworkMessageId
| join kind=inner (
UrlClickEvents
| where Url has "login.microsoftonline.com" and Url has "redirect_uri="
| extend RedirectTarget = extract(@"redirect_uri=([^&]+)", 1, Url)
| where isnotempty(RedirectTarget) and RedirectTarget !has "teams.microsoft.com" and RedirectTarget !has "sharepoint.com"
| where Timestamp > ago(24h)
) on NetworkMessageId
| project Timestamp, SenderFromAddress, SenderDisplayName, RecipientEmailAddress, Subject, FileName, Url, RedirectTarget, NetworkMessageId
| order by Timestamp descSigma
title: DocuSign-Themed Calendar Invite With Crafted Microsoft OAuth Redirect Parameter
id: 7c4e2b1a-3f9d-4e6a-8b2c-5a1d9e3f6c72
status: experimental
description: Detects an email carrying a DocuSign-impersonating sender with an .ics calendar-invite attachment whose embedded link abuses the legitimate Microsoft OAuth authorization endpoint (login.microsoftonline.com) via a crafted redirect_uri parameter pointing outside sanctioned Teams/SharePoint destinations, consistent with the Barracuda-documented blob URL phishing chain.
references:
- https://blog.barracuda.com/2026/09/09/browser-based-phishing-blob-urls-microsoft-redirects
author: Threadlinqs Team
date: 2026/09/10
tags:
- attack.resource-development
- attack.t1583.001
- attack.t1608.005
- attack.initial-access
- attack.t1566.001
- attack.execution
- attack.t1204.001
logsource:
category: mail_delivery
product: m365
detection:
selection_attachment:
FileExtension: '.ics'
selection_display_name:
SenderDisplayName|contains: 'DocuSign'
selection_redirect:
Url|contains:
- 'login.microsoftonline.com'
- 'redirect_uri='
filter_legit_domain:
SenderFromDomain|endswith:
- 'docusign.com'
- 'docusign.net'
condition: selection_attachment and selection_display_name and selection_redirect and not filter_legit_domain
falsepositives:
- Legitimate DocuSign calendar invites sent from correctly authenticated docusign.com/docusign.net domains
- Internal meeting-scheduling automation that legitimately chains through Microsoft OAuth for calendar provisioning
level: highKnown false positives:
- Legitimate DocuSign calendar invites sent from correctly authenticated docusign.com/docusign.net domains
- Internal scheduling automation that legitimately routes through Microsoft OAuth for calendar/meeting provisioning
Tuning notes: Maintain an allowlist of sanctioned redirect_uri destinations (internal apps, verified partner SSO) to reduce noise; extend the DocuSign-impersonation check to other frequently spoofed e-signature/collab brands (Adobe Sign, Dropbox Sign) since this campaign class is not brand-exclusive.
- DocuSign-Themed Calendar Invite With Crafted Microsoft OAuth Redirect (KQL) · unlock with TL Creds
- External CDN Resource Fetch and CSP Blob-URI Violation Consistent With In-Browser Phishing Page Construction (KQL) · unlock with TL Creds
- Low-Jitter postMessage-Style Polling to [indicator] Correlated With OAuth Credential Submission (KQL) · unlock with TL Creds
- DocuSign-Themed Calendar Invite With Crafted Microsoft OAuth Redirect (Sigma) · unlock with TL Creds
- External CDN Resource Fetch and CSP Blob-URI Violation Consistent With In-Browser Phishing Page Construction (Sigma) · unlock with TL Creds
- Low-Jitter postMessage-Style Polling to [indicator] Correlated With OAuth Credential Submission (Sigma) · unlock with TL Creds
- External CDN Resource Fetch and CSP Blob-URI Violation Consistent With In-Browser Phishing Page Construction (SPL) · unlock with TL Creds
- Low-Jitter postMessage-Style Polling to [indicator] Correlated With OAuth Credential Submission (SPL) · unlock with TL Creds
Indicators of compromise
As of 2026-09-10, Threadlinqs Intelligence tracks 8 indicators of compromise for TL-2026-2430.
Sample indicators (2 of 8)
- Domain:
cdn[.]bloom[.]io - Domain:
login[.]microsoftonline[.]com
Defanged, TLP:CLEAR. Full set of 8 indicators: unlock with TL Creds.
Technical analysis
The page registers a service worker and runs logic in a sandboxed iframe, letting the attacker's backend send live instructions via browser messaging so the phishing content can change without a hardcoded redirect, leaving no persistent URL for security tools to detect or blocklist.
On 2026-09-09, Barracuda Networks Associate Threat Analyst Ashitosh Deshnur published research describing a phishing campaign that begins with a DocuSign-themed email carrying a calendar invite (.ics) attachment. The invite itself is not malicious; it embeds a crafted redirect parameter that points to a genuine Microsoft OAuth endpoint (login.microsoftonline.com), giving the message the appearance of a routine business meeting request rather than an obvious phishing lure.
From the legitimate OAuth endpoint, the crafted redirect routes the victim through Microsoft Teams, which is made to load an external resource hosted on cdn.bloom[.]io. The victim's browser converts this externally sourced content into a blob URL -- a temporary, browser-generated address that points to data held only in local memory rather than on a remote server. The fake Microsoft login page is assembled and rendered entirely from this in-memory blob object inside the victim's own browser session, so no static phishing webpage ever exists for a security tool to crawl, fingerprint, or add to a blocklist.
Once rendered, the page registers a service worker and runs its logic inside a sandboxed iframe; together these manage the phishing workflow and user navigation. Barracuda's analysis states the phishing page 'receives instructions from backend infrastructure using browser messaging mechanisms,' which lets the attacker push live updates to the page's content and behavior after it has already loaded, instead of relying on a hardcoded redirect chain that a defender could map and block in advance.
Because every step -- the OAuth endpoint, the Teams content loader, and the final rendering surface -- takes place inside domains and browser behaviors that are either genuinely Microsoft-owned or exist only transiently in the victim's memory, the technique removes most of the static indicators (a persistent phishing domain, a fixed URL, a stable HTML artifact) that automated detection and blocklisting traditionally depend on. Barracuda characterizes this as phishing evolving 'beyond fake websites and suspicious domains,' reframing the credential-theft page as an ephemeral, dynamically-controlled browser construct instead of a hosted asset. Coverage was syndicated within 24 hours by SecurityWeek, Help Net Security, and several trade/MSP-focused outlets, extending visibility of the technique to a broad enterprise-security and SMB audience. No CVE applies -- this is an abuse of legitimate OAuth/Teams redirect and browser platform behavior (blob URLs, service workers, iframes, postMessage), not a software vulnerability. No malware hash, second-stage payload, actor name, or targeted-sector/region breakdown was disclosed in any of the reviewed sources; a BeaconBeagle lookup on cdn.bloom.io returned zero indexed C2 correlations.
Barracuda's writeup also discloses that the page carries a 'hidden command-and-control configuration' and states the operation is 'not a simple stand-alone [page], but part of a managed platform' that can be centrally operated, updated, and steered across many victims at once -- i.e., the campaign is consistent with a phishing-as-a-service (PhaaS) delivery model rather than a one-off, hand-built kit, though no branded kit name was disclosed and follow-up searches confirm this is a distinct campaign from named kits such as LogoKit or Sneaky 2FA. Every observable behavior in the chain -- constructing the Blob object from the fetched cdn.bloom[.]io resource, registering the service worker, executing logic inside the sandboxed iframe, and handling the browser-messaging instructions from the backend -- is implemented and driven by JavaScript running client-side in the victim's browser session; Barracuda's own framing of the core defensive gap is blunt: 'There is no phishing page to block.' Analyst Ashitosh Deshnur is quoted across the syndicated coverage: 'This campaign demonstrates how phishing is evolving beyond fake websites and suspicious domains, removing many of the indicators that security teams have traditionally relied on for detection.'
MITRE ATT&CK techniques by tactic:
Defense Evasion
T1027 Obfuscated Files or Information; T1036 Masquerading; T1684.001 Impersonation
Credential Access
Execution
T1059 Command and Scripting Interpreter; T1204 User Execution
Command and Control
T1071 Application Layer Protocol
Initial Access
Resource Development
Remediation
Immediate actions
- Monitor OAuth authorization flows and redirect chains for unexpected destinations, including redirects that traverse trusted Microsoft endpoints en route to a third-party resource
- Inspect blob URL creation and usage occurring in login or authentication contexts within the browser
- Flag service worker registrations tied to externally loaded, unrecognized content rather than the parent application's own origin
Workarounds
- Train staff to question unexpected document-signing or calendar-invite requests even when the click path runs entirely through legitimate Microsoft infrastructure
Longer-term hardening
- Deploy phishing-resistant MFA such as FIDO2 security keys or passkeys to reduce the value of a harvested credential
- Adopt email security tooling that analyzes a link's entire click path (including OAuth/Teams redirect hops) rather than evaluating only the first-hop URL
- Shift detection posture toward identity-based and behavioral controls rather than static URL/domain blocklisting, since this technique is designed to leave no persistent artifact to blocklist
Affected and fixed versions
- Microsoft — Microsoft 365 / Entra ID OAuth authorization flow and Microsoft Teams redirect handling (cloud service; abused as a trust-transfer mechanism, not a versioned vulnerability)
Timeline
- Barracuda Networks publishes the original research 'Email attacks evade detection by generating phishing pages directly inside the victim's browser,' authored by Associate Threat Analyst Ashitosh Deshnur, including the technical breakdown and defensive recommendations.
- Barracuda threat analysts identify a phishing campaign chaining a DocuSign-themed email lure through genuine Microsoft OAuth and Teams infrastructure to a blob-URL-rendered, browser-only fake login page.
- Threadlinqs HUNT phase ingests the Help Net Security article into the TL-2026-2430 threat skeleton for RESEARCH triage.
- Trade and MSP-focused outlets (iTWire, SMBtech, ITBrief, CloudLinkTech, vmBlog) syndicate the Barracuda research, extending visibility of the technique to enterprise IT and managed-service-provider audiences.
- SecurityWeek publishes independent coverage including the Barracuda quote that 'this campaign demonstrates how phishing is evolving beyond fake websites and suspicious domains.'
- Help Net Security publishes coverage summarizing the Barracuda blob URL phishing findings for a broader security audience.
Sources cited
- Email attacks evade detection by generating phishing pages directly inside the victim's browser blog.barracuda.com · Vendor report · 2026-09-09
- Barracuda Releases New Research: Email attacks evade detection by generating phishing pages directly inside the victim's browser vmblog.com · Vendor report · 2026-09-10
- Cybercriminals are building phishing pages that exist only inside victims' browsers Help Net Security · News · 2026-09-10
- New Phishing Attack Creates Malicious Pages Inside the Victim's Browser SecurityWeek · News · 2026-09-10
- Barracuda Research Finds Phishing Attack That Builds Fake Login Pages Inside The Victim's Browser, Leaving No Trail For Security Tools smbtech.au · News · 2026-09-10
- Email attacks evade detection by generating phishing pages directly inside the victim's browser, Barracuda research finds itwire.com · News · 2026-09-10
- Phishing Hides Pages in Browser via Blob URLs, Teams Lure cloudlinktech.com · News · 2026-09-10
- Barracuda spots blob URL phishing in Microsoft browser itbrief.asia · News · 2026-09-10
Previous and next in phishing
Drafted by the Threadlinqs research pipeline from 8 sources; published ; last updated —.