← Breach analysis by the Dual Vulnerabilities framework

Telecoms and BPO companies SIM-swapping intrusion campaign (UNC3944/Scattered Spider, 2022)

2022 · J - Information and communication · Social-engineering intrusion campaign enabling SIM swapping and data theft

Analysed 2026-11-07 by Hermes

Open this case in the tool →

What happened

Social-engineering intrusion campaign enabling SIM swapping and data theft

Dual vulnerability type

Expertise to Execution

Evidence coverage: 3/6 stages, weight 12/60

Decisive stage: entry (clearest evidence) · Alternative: Strategy to Execution

Why this classification

(1) Telecoms companies and their BPO support providers are a critical-infrastructure sector whose business depends on identity verification: the standing expectation, stated at regulator level in the joint CISA/FBI/NCSC advisory on this actor, is phishing-resistant MFA and independently verified password/MFA resets, and the technical remedies (FIDO/WebAuthn, number matching, removal of SMS as an MFA factor, locked-down self-enrollment, video verification of reset requests) were known and prescribed as effective. (2) In practice, across the 2022 to early-2023 wave at these companies, smishing text messages and calls impersonating IT staff harvested valid credentials, help desks reset passwords and MFA factors for callers who supplied PII gathered about the employees, and permissive MFA self-enrollment policies let the actors register their own trusted devices; the actors then reached mobile carrier systems to carry out SIM swaps, ran constant, typically daily activity that EDR did not surface because the tooling was legitimate, and took sensitive data out through cloud channels, while the slow implementation of containment measures let the actor add persistence and even re-enable accounts the victims had disabled. (3) The gap sits within the material column: specialists and the regulator knew the remedies, the sector's own obligation to verify identities was explicit, and the chain broke where these known fixes were not realised in authentication configuration and reset-process design - no source documents a victim policy, programme or budget decision that failed to translate into operations (which would make it Strategy to Execution) or an expert warning that leadership ignored (Expertise to Leadership); the classification rests on the entry stage, the best-documented and load-bearing stage, with the detection and egress gaps pointing to the same type and the entry vector itself well documented rather than unknown. RM22 (Expertise to Execution lens): the authentication controls meant to stop exactly this tradecraft stayed ineffective in operations - SMS- and OTP-based MFA that yielded to impersonation or push fatigue and self-enrollment policies that let attackers add trusted MFA devices, so the actors logged in as legitimate users (causal role: Enabling). RM45 (Expertise to Execution lens): the account-recovery and SIM-swap/number-porting processes lacked independent verification, dual authorisation or segregation of duties, so a caller with gathered PII could walk them and the campaign's defining impact - SIM swaps against subscribers - was possible (causal role: Enabling). RM34 (Expertise to Execution lens): the monitoring the companies could deploy and tune was insufficiently configured to flag legitimate remote-administration tooling and residential-proxy access, so daily hands-on-keyboard activity persisted until external threat hunting and incident response engagements surfaced it (causal role: Obscuring).

Evidence by stage

Risk markers

Sources

Notes

The case file's attribution/blog URL (mandiant.com/resources/blog/unc3944-sms-phishing-sim-swapping-ransomware) no longer serves the article - it redirects to a generic Mandiant consulting page - so the Wayback snapshot of 15 September 2023, opened and read in this run, is cited instead; the same report is also mirrored on the Google Cloud blog (opened; not cited because the extracted mirror text was partial). Victims are unnamed organisations in a multi-company campaign, so Country is null; Sector follows the dataset's own critical-infrastructure subcategory (receiver_subcategory Telecommunications, SBI J) since the output schema allows only one sector entry, while the BPO co-victims would fall under SBI N. Year 2022 follows the dataset start date of 01.01.2022 and the method rule to take the first year (activity ran to early 2023). The mechanism evidence rests on vendor and agency reporting of the group's activity across multiple unnamed victim investigations (group-level pattern), not on an incident-specific victim, regulator or court disclosure, which caps confidence at 3. The EuRepoC intelligence_impact value is internally contradictory (data corruption without exfiltration OR major exfiltration without corruption); it is read as data theft/exfiltration without corruption, consistent with incident_type Data theft and the CrowdStrike account. Row 59 (case file opened) describes the same actor's separate mid-2023 wave with ransomware against a wider target set; this row covers the 2022 telecom/BPO SIM-swapping wave, so no M21-style recurring-incident marker was coded and the two rows are not merged. No consent-gated actions were needed in this unattended run.