This Week in Cybersecurity: The 17-Month Cloud Intrusion, Microsoft Kills SMS MFA, and Dark Web Employee Leaks

A Mystery Attacker Lived Inside City-Forum’s Salesforce & ServiceNow for 17 Months
Here’s a story that will make every IT team double-check their cloud configurations tonight. City-Forum, a prominent UK events and corporate intelligence firm, disclosed that an unknown threat actor maintained undetected persistent access to their Salesforce and ServiceNow environments for 17 months—spanning from October 2023 through February 2025.
The attacker systematically harvested corporate contact rosters, customer attendee files, and employee phone records. City-Forum stated that no financial transaction systems were impacted, but 500+ days of persistent surveillance provides threat actors with the exact organizational intelligence required to execute lethal, hyper-targeted spear phishing and executive impersonation campaigns.
What makes this incident especially critical for C-level leadership? Salesforce and ServiceNow represent gold-standard enterprise SaaS platforms. Yet, platform reputation does not equal baseline tenant security. The shared responsibility model dictates that vendors protect their cloud infrastructure, while organizations remain strictly accountable for identity lifecycle management, API token hygiene, and access logs.
“A platform is only as safe as how you configure and monitor it. Brand trust never substitutes for active security telemetry.”
Key architectural safeguards every enterprise must enact immediately:
Conduct quarterly audits of all connected integrations, OAuth grants, and administrator tiers. Deprecate stale test instances and enforce strict least-privilege boundaries.
LEAST PRIVILEGE
Leverage SSPM and CASB tools to detect abnormal session tokens, unauthorized bulk exports, and rogue API connections in real-time.
SSPM TELEMETRY
Eliminate local SaaS passwords. Enforce conditional access policies via Entra ID or Okta with hardware-bound credentials.
ENTERPRISE SSO
Microsoft Kills SMS & Voice MFA—Passkeys Are Taking Over
This is the paradigm shift cybersecurity engineers have championed for years. Microsoft is officially deprecating SMS and voice-call multi-factor authentication, systematically migrating hundreds of millions of corporate and personal accounts toward phishing-resistant FIDO2 passkeys and hardware tokens.
Why is Microsoft aggressively eliminating SMS one-time passwords? Because telecom-based MFA has been completely demolished by modern automated adversary tactics:
Attackers bribe or deceive cellular customer service reps into porting executive numbers to attacker-controlled SIM cards, routing OTPs directly to adversary devices.
+400% YoY GROWTH
Adversaries exploit archaic Signaling System 7 protocols designed in 1975 to intercept unencrypted SMS traffic mid-flight without ever alerting the subscriber.
TELECOM EXPLOIT
Automated proxy reverse-phishing frameworks (such as Evilginx) relay OTP codes and session cookies instantaneously, bypassing SMS MFA in less than 45 seconds.
SESSION HIJACK
Passkeys resolve these attack vectors cryptographically. Based on asymmetric public-key cryptography and WebAuthn standards, passkeys remain bound to the client device. Because the private key never leaves the secure enclave (TPM or YubiKey), there is no shared secret to intercept, phish, or social-engineer.
“Passkeys are phishing-resistant by mathematical design. There is no password to leak, no code to type, and no telecom carrier to compromise.”
Microsoft’s Enterprise Deprecation Timeline:
An AI Agent Just Hacked a Gym Booking System to Skip the Queue
If you assumed emerging AI threats were restricted to boardroom deepfakes or synthetic spear-phishing, this incident is a wakeup call. An Australian engineer deployed an autonomous AI agent that systematically exploited a commercial gym reservation system, securing every priority class slot before human members could load the app.
Here is the crucial reality: The engineer never wrote a custom exploit, injected SQL queries, or reverse-engineered private APIs. He simply prompted an autonomous browser agent: “Monitor class availability at 6:00 AM and reserve my Tuesday session the millisecond registration opens.” The agent autonomously handled authentication, solved UI navigation, and triggered the reservation at machine speed.
“You no longer need to be an elite hacker to break application logic. Anyone can command an AI agent to execute human tasks at machine velocity.”
The enterprise implications of agentic automation: What begins on gym booking apps scales instantly into critical business vulnerabilities:
Autonomous agents scraping and locking high-demand commercial hardware, medical equipment, and travel inventory before enterprise procurement teams can click approve.
AI agents intelligently mimicking human typing cadences, dynamically rotating residential proxies, and solving behavioral challenges to bypass legacy WAF rules.
Agents autonomously mapping checkout funnels to exploit discount stacks, race conditions, and refund loopholes that static vulnerability scanners miss entirely.
McDonald’s, Vodafone & TCS Employee Data Leaked on the Dark Web
The enterprise attack perimeter is increasingly defined by third-party vendor compromise. A major illicit marketplace post is currently offering verified internal records from McDonald’s, telecommunications giant Vodafone, and IT services titan Tata Consultancy Services (TCS) packaged together in a single consolidated archive.
The exposed dataset contains exact corporate email addresses, departmental roles, internal employee ID codes, and managerial reporting hierarchies. While credit cards or banking hashes make sensational headlines, internal workforce directory data is exponentially more hazardous for corporate security teams.
Armed with accurate organizational blueprints, cybercriminals construct weaponized spear-phishing campaigns that effortlessly evade conventional email security filters. When an email originates with exact peer names, valid contractor ID formats, and accurate department references, human employees have virtually zero suspicion of malicious intent.
“Employee identity data is not low-tier information. It is the tactical skeleton key used to engineer high-dollar wire fraud and enterprise network takeovers.”
Enforce strict zero-trust network access (ZTNA) for external staffing agencies and cloud HR vendors. Never grant direct directory synchronization to unvetted third parties.
Actively monitor breach forums and paste sites for your domain names. Proactively revoke tokens and force password resets the moment staff credentials surface online.
Establish mandatory dual-authorization protocols and secondary communication channels for any wire transfers, invoice adjustments, or payroll alterations.
Authored by our certified security architects and engineering leadership. Looking to audit your cloud SaaS posture, deploy hardware passkeys, or implement dark web credential monitoring?
