Back to Blog

Threat Bulletin: Iranian APT Groups Targeting Cloud Identity Infrastructure

The February 28 strikes changed the calculus. Iranian state-sponsored hackers will escalate — the only question is when. We've deployed 11 detection rules tuned specifically for how these groups attack cloud identity infrastructure.

The Current Threat Landscape

On February 28, 2026, the United States and Israel launched coordinated strikes against Iran under Operation Epic Fury and Operation Roaring Lion. Within hours, over 150 hacktivist incidents were claimed in open channels, and Iran's "Electronic Operations Room" was established the same day to coordinate cyber operations across multiple state-aligned groups.

While Iran's domestic internet connectivity dropped to 1-4% following the strikes, threat intelligence firms warn this is temporary. Tehran-linked hackers are stepping up digital reconnaissance and preparing for potentially disruptive cyber activity. Unit 42 assesses that while connectivity loss has degraded near-term coordination, Iranian threat actors continue to operate through proxies and VPNs outside the country.

"We expect Iran to target the U.S., Israel, and Gulf Cooperation Council countries with disruptive cyberattacks, focusing on targets of opportunity and critical infrastructure." — John Hultquist, Google Threat Intelligence Group
Key Numbers
  • 150+ hacktivist incidents claimed since February 28
  • 11 new detection rules deployed by Auth Sentry for Iranian APT tradecraft
  • 4 major Iranian APT groups tracked in this bulletin
Why Identity Infrastructure Is the Target

Iranian APT groups have a well-documented preference for cloud identity attacks. Unlike endpoint compromises that require malware, identity-based intrusions use valid credentials and legitimate tools — making them harder to detect and attribute. CISA specifically warns about credential access and persistence techniques targeting Okta, Microsoft 365, and Google Workspace.

The Actors We're Tracking

Four major Iranian APT groups have demonstrated consistent capabilities against cloud identity infrastructure. These groups operate under different Iranian intelligence agencies but share similar tradecraft when targeting Western organizations. A September 2025 leak of APT35/Charming Kitten internal documents provided detailed insight into their operational methods, target selection, and exploitation techniques — intelligence we've incorporated into our detection rules.

Group Aliases Affiliation Primary Targets
APT33 Elfin, Refined Kitten IRGC Energy, aerospace, defense
APT34 OilRig, Helix Kitten MOIS Government, finance, telecoms
APT35 Charming Kitten, Mint Sandstorm IRGC Journalists, academics, dissidents, government officials
MuddyWater MERCURY, Mango Sandstorm MOIS Government, defense, telecoms

How Iranian APT Attacks Unfold

Iranian APT cloud identity intrusions follow a consistent, phased pattern with characteristically patient dwell times between phases — a key differentiator from financially-motivated actors like Scattered Spider, which operate in minutes to hours. Understanding this kill chain is essential for detection.

[Phase 0 — Reconnaissance] Behavioral anomaly signals before any attack technique fires ↓ [Phase 1 — Initial Access] Low-and-slow password spray via distributed infrastructure ↓ 1–2 attempts per account over hours to evade lockout thresholds [Phase 2 — Credential Access] Spray-to-success: valid credential obtained ↓ Actors proceed within 1–6 hours [Phase 3 — Persistence] MFA device enrollment + OAuth app authorization ↓ Creates access that survives password resets [Phase 4 — Escalation] Privilege escalation via admin group membership ↓ [Phase 5 — Collection] Data exfiltration via cloud APIs (email, files, calendar)

The Tradecraft We Detect

Distributed Password Spraying

All tracked Iranian groups use a distributed low-and-slow spray methodology. Rather than high-volume single-source spraying that triggers lockouts, they operate via rotating VPS pools, residential proxies, and Tor exit nodes. Spray passwords follow observable patterns: seasonal formats (Spring2026!), organization-name variants ([Company]2025!), and known common passwords.

Target lists are pre-enumerated from LinkedIn, company directories, and prior intelligence — the spray is directed, not random. This is reconnaissance-informed credential testing.

MFA Persistence Enrollment

APT35 and APT33 specifically enroll new authenticator applications or hardware tokens within hours of initial access. This creates an authentication pathway that survives password resets — a common defensive response that inadvertently locks the legitimate user out while the attacker retains access via their registered MFA device. (For more on how attackers abuse MFA, see our guide on MFA fatigue attacks.)

OAuth Application Persistence

APT33 and APT34 immediately register OAuth application grants on compromised accounts. This provides a persistent access token that survives credential rotation and operates outside session-based controls, enabling long-term silent access to email, files, and calendar data.

Our Detection Approach

Auth Sentry's Iranian APT detection suite operates on two complementary layers: real-time pattern detection for known attack techniques, and behavioral analytics that can identify campaigns before specific attack patterns emerge. (For a deeper look at our detection methodology, see how Auth Sentry works.)

Real-Time Pattern Detection (IR-001 to IR-007)

Seven detection rules trigger immediately when we observe the specific tactics Iranian actors use — password sprays, persistence mechanisms, privilege escalation. When a pattern matches, your security team gets a high-confidence investigation with full context, not a low-signal alert to triage.

Predictive Behavioral Analytics (BA-IR-001 to BA-IR-004)

A second layer of detections uses graph-based behavioral analytics to identify campaigns before any specific attack technique fires. Our data science approach analyzes identity behavior patterns across three dimensions:

  • Community Analysis — Identifies spray campaigns by detecting authentication failures distributed across multiple organizational communities simultaneously — the signature of a directory-wide attack.
  • High-Value Target Profiling — Scores each identity by access breadth and organizational influence. High-value identities are disproportionately targeted by APT35/Charming Kitten for their intelligence value.
  • Behavioral Drift Detection — Measures how each identity's behavior changes against their established baseline. Iranian APT produces a characteristic moderate, sustained drift pattern — different from the rapid spikes seen in ransomware attacks.
Reconnaissance-Phase Detection

Our behavioral analytics can fire during Phase 0 — detecting the credential testing and reconnaissance phase of a campaign before traditional pattern-based signatures trigger. This gives security teams early warning while the attack is still developing, not after persistence is already established.

Detection Rules Deployed

Pattern-Based Rules

Rule Name Kill Chain Phase Severity
IR-001 Low-and-Slow Password Spray Initial Access HIGH
IR-002 Spray-to-Success (Confirmed Compromise) Credential Access CRITICAL
IR-003 MFA Enrollment After Spray Persistence HIGH
IR-004 OAuth App Persistence Persistence HIGH
IR-005 Privilege Escalation After Compromise Privilege Escalation CRITICAL
IR-006 Multi-Source Targeted Attack Initial Access HIGH
IR-007 Multi-Stage Campaign Correlation Full Campaign CRITICAL

Behavioral Analytics Rules

Rule Name Detection Method
BA-IR-001 Community-Diversity Spray Signature Detects auth failures spanning multiple organizational communities from a single source
BA-IR-002 Patient Behavioral Drift Identifies the moderate, sustained behavioral drift characteristic of Iranian APT credential testing
BA-IR-003 High-Value Targeted Identity Flags when high-influence identities show drift or multi-source auth failures
BA-IR-004 Predictive Kill Chain Estimator Estimates current attack phase from behavioral signals — fires at Phase 0 before pattern rules trigger

Rule Detail: How Our Detection Logic Works

To illustrate our approach, here's the specific logic behind two representative rules — one pattern-based and one behavioral:

IR-001: Low-and-Slow Password Spray (Pattern-Based)

Trigger: Single source IP generates authentication failures against 5+ distinct user accounts within a 4-hour window.

Why this works: Iranian APT spray campaigns deliberately stay below typical lockout thresholds (usually 3-5 failures per account). By correlating failures across accounts from the same source, we detect the campaign pattern that per-account monitoring misses.

Tuning note: Threshold can be adjusted based on organization size. Larger directories may warrant higher thresholds to reduce noise from legitimate service accounts or misconfigured applications.
BA-IR-001: Community-Diversity Spray Signature (Behavioral Analytics)

Approach: We continuously analyze how identities cluster based on shared access patterns, applications, and organizational relationships. When authentication failures from a single source span 3+ distinct identity communities simultaneously, it indicates a directory-wide spray — not targeted account guessing.

Why this matters: Traditional detection looks at individual accounts or simple IP reputation. Our approach sees the shape of an attack across your identity graph. A spray hitting Finance, Engineering, and Executive communities in the same hour has a distinct signature that's invisible to event-by-event analysis.

Early warning advantage: This rule often fires before IR-001, because it detects the structural pattern of a spray campaign even when individual account failure counts remain low.

A Note on Detection Tuning

No detection rule is perfect out of the box. IR-003 (MFA Enrollment After Spray) may trigger when a user legitimately enrolls a new device after experiencing unrelated auth failures — for example, after a password reset due to a forgotten credential. Auth Sentry correlates MFA enrollment with preceding spray activity from known-suspicious infrastructure to reduce these false positives, but some tuning is expected based on your organization's enrollment patterns.

All detection rules include suppression controls, threshold adjustments, and the ability to allowlist known service accounts or IP ranges. Our approach: high-confidence investigations by default, with knobs available when your environment requires them.

Why Iranian APT Evades Traditional Defenses

Iranian APT has a distinct operational fingerprint that separates it from other threat actors — and explains why traditional security tools often miss them:

  • Patient spray pace — Hours per cycle, not minutes. Rate-based detections calibrated for fast-moving actors like Scattered Spider will miss Iranian APT without time-window tuning.
  • Distributed infrastructure — VPS pools and residential proxies mean single-IP lockout rules are insufficient. They attack from many sources simultaneously.
  • Persistence before collection — MFA and OAuth persistence are established before any data collection occurs. By the time you see exfiltration, they've already entrenched.
  • Curated targeting — APT35 in particular maintains curated target lists. They're not spraying randomly — they know who they want.

MITRE ATT&CK Coverage

These detection rules provide coverage across the following MITRE ATT&CK techniques:

  • T1110.003 — Password Spraying
  • T1078 — Valid Accounts
  • T1098 — Account Manipulation
  • T1098.003 — Additional Cloud Roles
  • T1098.005 — Device Registration
  • T1087 — Account Discovery
  • T1530 — Data from Cloud Storage

Protect Your Organization

Auth Sentry connects to your identity providers — Okta, Microsoft Entra ID, Google Workspace — and applies graph-based behavioral analytics to detect identity-based attacks that traditional tools miss.

These Iranian APT detection rules are now live for all Auth Sentry customers. If you're not yet protected, now is the time to act. Have questions? Check our FAQ or contact us directly.

Detect Attacks Before They Escalate

Connect your identity sources and see what Auth Sentry finds in your environment.

Start Free with Monitor

Auth Sentry by Hummingbird Security

Auth Sentry is an identity threat detection and response platform built on graph intelligence. We help security teams detect identity-based attacks before they become breaches. Learn more at gethumming.io.

Ready to Transform Your Identity Security?

Reduce alert fatigue. Get complete investigations. Stop attacks before they progress.

See how Auth Sentry delivers real outcomes for your SOC.

Request Free Trial